Hung Quoc To

---
name: conversation-format-selector
description: Choose the right conversation format — casual chat, scheduled meeting, or phone/video call — for a customer discovery interaction. Use this skill whenever the user is deciding between a formal meeting and a casual conversation, wondering how long customer interviews should be, unsure whether to meet in person or do a phone or video call, preparing to talk to potential customers at a conference or event or meetup, asking how to approach someone at a networking event or industry gathering, spending too much time scheduling formal meetings and not getting enough conversations done (the meeting anti-pattern), defaulting to hour-long Zoom calls for every customer interaction, planning the logistics or setting or duration of a customer conversation, or wondering whether a conversation should be structured or informal — even if they don't explicitly mention "format" or "meeting type." This skill is about HOW to have the conversation (format, duration, setting, formality level), not about finding people to talk to (use conversation-sourcing-planner) or what questions to ask (use conversation-question-designer).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-mom-test/skills/conversation-format-selector
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: verified
source-books:
  - id: the-mom-test
    title: "The Mom Test"
    authors: ["Rob Fitzpatrick"]
    chapters: [4, 6]
tags: [customer-discovery, conversation-format, meeting-planning, casual-conversations, customer-interviews]
depends-on:
  - conversation-sourcing-planner
  - conversation-question-designer
execution:
  tier: 1
  mode: plan-only
  inputs:
    - type: document
      description: "Description of the conversation context — who the user wants to talk to, what they want to learn, and their current relationship with the target person"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment with file read/write access."
---

# Conversation Format Selector

## When to Use

You need to decide how to structure an upcoming customer discovery interaction — the format, duration, setting, and approach. Typical situations:

- The user is about to reach out to a potential customer and is deciding whether to request a formal meeting or keep it casual
- The user has an opportunity to talk to someone at an event, conference, or social setting and wants to know how to approach it
- The user is defaulting to scheduling calendar meetings for every conversation and may be falling into the Meeting Anti-Pattern
- The user is considering a phone or video call instead of meeting in person
- The user wants guidance on how long a conversation should be given their learning goals

Before starting, verify:
- Does the user know what they want to learn? (If not, suggest they use the `conversation-question-designer` skill first)
- Does the user have a specific person or type of person in mind? (If not, they need to define their target first)

**Mode: Plan-only** — The agent recommends a conversation format with timing, setting, and approach guidance. The human executes the actual conversation.

## Context & Input Gathering

### Required Context (must have — ask if missing)

- **Who they want to talk to:** Role, seniority, relationship to the user (stranger, acquaintance, warm intro, existing contact). This determines formality requirements and which formats are available.
  - Check prompt for: person descriptions, titles, "I want to talk to...", event attendee descriptions
  - Check environment for: `customer-segments.md`, `conversation-notes/`, contact lists
  - If still missing, ask: "Who do you want to talk to? What is their role and how do you know them (or don't)?"

- **What they want to learn:** The user's current learning goals. This determines how much time the conversation actually needs.
  - Check prompt for: assumptions to validate, questions to answer, "I want to find out..."
  - Check environment for: `question-script.md`, `learning-log.md`, `product-idea.md`
  - If still missing, ask: "What are the 1-3 most important things you want to learn from this conversation?"

### Observable Context (gather from environment)

- **Product stage:** Pre-product (learning only) vs post-product (learning + demo). This changes duration needs and format constraints.
  - Look for: product descriptions, prototypes mentioned, demo references
  - If unavailable: assume pre-product

- **Conversation setting:** Where the interaction might happen — industry event, office, coffee shop, online, or unplanned encounter.
  - Look for: event mentions, location references, "I'll be at..."
  - If unavailable: assume the user is choosing and needs a recommendation

- **Previous conversations:** Whether the user has talked to this type of person before.
  - Look for: `conversation-notes/`, learning logs, "I've already talked to..."
  - If unavailable: assume first conversations with this segment

### Default Assumptions

- If no product stage specified -> assume pre-product (learning phase)
- If no relationship specified -> assume cold or weak-tie contact
- If no setting specified -> recommend the most effective format for their learning goals

### Sufficiency Threshold

```
SUFFICIENT when ALL of these are true:
- Target person type is known
- At least 1 learning goal is identified

PROCEED WITH DEFAULTS when:
- Learning goals are approximate ("I want to understand their workflow")
- Person type is general ("marketing managers" without specific individual)

MUST ASK when:
- No target person type at all
- No learning goals or topic area
```

## Process

### Step 1: Assess Learning Depth Required

**ACTION:** Categorize the user's learning goals into one of four depth levels. Each level has a natural duration range that should drive format selection.

**WHY:** The biggest mistake in conversation planning is letting the calendar dictate duration instead of letting the learning goal dictate it. A 1-hour formal meeting costs roughly 4 hours when you factor in scheduling, commuting, and review time. If your learning goal only requires 5 minutes of actual conversation, that overhead is a massive waste of the most precious startup resource: founder time and attention.

| Depth Level | Duration | What You Are Learning | Example Goals |
|-------------|----------|----------------------|---------------|
| **Validate existence** | 5 minutes | Whether a problem exists and matters to this person | "Is hiring actually painful for them?", "Do they even care about analytics?" |
| **Understand workflow** | 10-15 minutes | How they currently achieve a goal or handle a problem, what they have tried | "How do they onboard new employees?", "What does their reporting process look like?" |
| **Deep industry dive** | 30-60 minutes | Industry dynamics, complex purchasing processes, organizational politics | "How does budget approval work in their company?", "What is the competitive landscape for their tools?" |
| **Product feedback** | 30 minutes (structured) | Reactions to a prototype or product, purchasing intent, next steps | "Will they try the beta?", "What would they change about the interface?" |

**IF** the user's learning goals span multiple depth levels -> recommend the shallowest level that covers the most critical goal. Additional goals can be addressed in follow-up conversations.

**OUTPUT:** The identified depth level with estimated actual conversation time needed.

### Step 2: Evaluate Format Options

**ACTION:** Score each of three conversation formats against the user's specific context using the trade-off criteria below. Recommend the highest-scoring option.

**WHY:** Each format has structural advantages and disadvantages that affect the quality of learning. Casual conversations produce less biased data because the other person does not feel like they are "doing you a favor." Formal meetings are sometimes necessary but carry overhead and expectation costs. Phone and video calls sacrifice body language, weaken the power dynamic, and eliminate the possibility of warm introductions afterward. Choosing the wrong format means you either waste time on unnecessary formality or miss learning that required deeper engagement.

**Three format options:**

#### Option A: Casual Conversation
- **Best for:** Validate-existence and understand-workflow depth levels. Pre-product stage. Situations where you can encounter the target person organically (events, shared communities, mutual connections).
- **Advantages:** No scheduling overhead, no expectation you will show a product, produces less biased responses because the person does not know they are in a "meeting," allows serendipitous follow-up topics, easy to do many in a single outing.
- **Constraints:** Not appropriate for senior executives you have no relationship with. Hard to take detailed notes in real time. Requires knowing your 3 key questions so you can navigate the conversation without a script.
- **Duration:** 5-15 minutes of actual conversation.
- **Settings:** Industry events, conferences, meetups, coffee shops, coworking spaces, social gatherings, online communities (DMs or forum threads).

#### Option B: Scheduled Meeting
- **Best for:** Deep-industry-dive and product-feedback depth levels. Post-product stage. Senior people outside your network who require a formal ask. Situations where you need a full hour of focused time.
- **Advantages:** Dedicated attention, ability to show a product or prototype, appropriate for decision-makers and enterprise contacts, easier to take structured notes.
- **Constraints:** High overhead (1-hour meeting typically costs 4 hours total with scheduling, travel, and review). Sets expectation you will show something. Creates formal dynamic that can bias responses. Harder to get the meeting in the first place.
- **Duration:** 30-60 minutes of scheduled time.
- **Settings:** Office meetings, video calls, coffee meetings by appointment.

#### Option C: Phone or Video Call
- **Best for:** Geographic constraints where in-person is impossible. Quick follow-ups with people you have already met. Situations where the learning goal is narrow and specific.
- **Advantages:** No travel time, easy to schedule, works across distances.
- **Constraints:** Loses body language and environmental observation. Creates an interview dynamic that feels formal even when the content is casual. Kills the possibility of in-person warm introductions to colleagues. The person feels less engaged and more like they are doing you a favor. Tends to default to 30-minute calendar blocks regardless of actual learning needs.
- **Duration:** Match to actual learning need (often 10-15 minutes is sufficient, not 30).

**Scoring criteria:**

| Criterion | Question to Ask |
|-----------|----------------|
| **Time efficiency** | Does the format match the actual time needed, or does it force unnecessary overhead? |
| **Data quality** | Does the format minimize bias and maximize honest responses? |
| **Relationship building** | Does the format create opportunities for follow-up and warm introductions? |
| **Accessibility** | Can the user actually achieve this format given their relationship with the target person? |

**IF** the user is defaulting to scheduled meetings for every conversation -> flag the Meeting Anti-Pattern (the tendency to relegate every customer conversation opportunity into a formal calendar block). Explain that over-reliance on formal meetings wastes time, sets expectations you will show a product, and causes you to overlook perfectly good chances for serendipitous learning.

**IF** the user is considering a phone call primarily to avoid awkwardness -> flag this as an avoidance pattern, not a strategic choice. Phone calls sacrifice too much signal quality to be used as a comfort mechanism.

### Step 3: Apply the Advisor Evaluation Mindset

**ACTION:** Reframe the user's internal approach to the conversation, regardless of format.

**WHY:** The default mindset for customer conversations is needy: you are asking someone for their time, hoping they will validate your idea, grateful for their attention. This creates a power imbalance where you feel subordinate, which leads to pitching, compliment-seeking, and failure to press for real answers. Reframing the conversation as an advisor evaluation flips the dynamic: instead of seeking customers, you are evaluating whether this person would make a good advisor for your space. The conversation topics stay the same, but you are now in the evaluator's seat, which makes you calmer, more confident, and more likely to ask hard questions.

**Guidance to include in the recommendation:**
- Go into the conversation looking for helpful, knowledgeable people who understand your problem space, not looking for customers or validation
- Evaluate whether this person has genuine insight and experience, which naturally puts you in control of the dynamic
- You do not need to tell them you are evaluating them as an advisor; this is about your internal narrative, not an explicit request
- This mindset works for every format (casual, meeting, or call) but has the biggest impact on formal meetings where the needy dynamic is strongest

**IF** the recommended format is a formal meeting -> emphasize the advisor evaluation mindset especially strongly, because formal meetings are where the needy dynamic is most damaging.

### Step 4: Produce the Format Recommendation

**ACTION:** Write a concrete recommendation document covering format choice, timing, setting, approach, and preparation checklist.

**WHY:** The recommendation must be actionable — not a theoretical comparison. The user should be able to read it once and know exactly what to do, how long it will take, and what mindset to carry into the conversation.

**Output format:**

```markdown
# Conversation Format Recommendation

## Context
- **Target:** [who they are talking to]
- **Learning Goals:** [what they want to learn]
- **Product Stage:** [pre-product / post-product]
- **Date Prepared:** [today]

## Recommended Format
**[Casual Conversation / Scheduled Meeting / Phone Call]**

### Why This Format
[2-3 sentences explaining why this format fits their specific context,
referencing the depth level and trade-off analysis]

### Timing
- **Actual conversation time needed:** [X minutes]
- **Total time investment:** [Y minutes, including overhead]
- **Duration guidance:** [specific advice on when to wrap up or extend]

### Setting
- **Where:** [specific setting recommendation]
- **Alternative:** [backup option if primary is not available]

### Approach
- **Mindset:** [advisor evaluation framing specific to their situation]
- **Opening:** [how to start the conversation naturally given the format]
- **Formality check:** [how to tell if the conversation is too formal]

### Preparation Checklist
- [ ] Know your 3 key questions (use `conversation-question-designer` if needed)
- [ ] [Format-specific prep item]
- [ ] [Format-specific prep item]

### Formality Warning Signs
Watch for these signals that the conversation has become too formal:
- You open with "Thanks for agreeing to this interview..."
- You are using rating scales ("On a scale of 1 to 5...")
- It feels like they are doing you a favor by talking to you
- You are following a rigid script instead of having a natural conversation

If you notice these → loosen up, drop the script, and ask about
something genuinely interesting about their work.
```

**IF** the user provided a file path or working directory -> write the output to `format-recommendation.md`
**ELSE** -> present the output directly in the conversation

## Examples

### Example 1: Founder Planning Customer Conversations at an Industry Conference

**Scenario:** A founder building a tool for restaurant inventory management is attending a food industry trade show next week. They want to talk to restaurant owners about how they manage ingredient ordering.

**Trigger:** "I'm going to a food industry conference next week. I want to learn how restaurant owners handle inventory and ordering. Should I try to set up meetings beforehand, or just wing it at the event?"

**Process:**
1. Depth assessment: Understand-workflow level (10-15 minutes needed per conversation)
2. Format scoring: Casual conversation scores highest — organic access to target customers at an event, no scheduling overhead, can do 6-10 conversations in a single day
3. Advisor evaluation: Reframe from "trying to get meetings with busy restaurant owners" to "evaluating which restaurant operators really understand the inventory challenge"

**Output (abbreviated):**
```
## Recommended Format
**Casual Conversation**

### Why This Format
You are attending an event where your target customers are already present
and open to talking. Scheduling formal meetings in advance would cost you
4 hours of overhead per conversation for learning that takes 10-15 minutes.
At an event, you can have 6-10 of these conversations in a single day by
simply being curious and interested.

### Timing
- Actual conversation time needed: 10-15 minutes
- Total time investment: 10-15 minutes (no overhead)
- Duration guidance: You can usually get what you need in 10 minutes.
  If the conversation is going well and they are engaged, let it run.
  People love talking about the annoying parts of their work.

### Approach
- Mindset: You are evaluating whether these restaurant operators have
  real insight into the inventory problem. Not everyone will. Some will
  have it solved. That is useful data too.
- Opening: "How do you handle ordering for a place this size? I imagine
  that gets complicated."
- Formality check: If you catch yourself saying "Can I ask you a few
  questions about your business?" you have made it too formal. Just be
  interested.
```

---

### Example 2: Product Manager Deciding Between Phone Call and In-Person Meeting

**Scenario:** A product manager at a B2B analytics company wants to understand how marketing directors at mid-sized companies create executive reports. They have a warm introduction to a specific marketing director at a company across town.

**Trigger:** "I have a warm intro to a marketing director. Should I suggest a phone call or go meet them in person? I don't want to waste their time with a full meeting."

**Process:**
1. Depth assessment: Understand-workflow level (10-15 minutes), but could tip into deep-industry-dive if the reporting process is complex (30 minutes)
2. Format scoring: Scheduled meeting slightly edges out casual because this is a warm intro to a specific senior person. Phone call scores lowest — loses body language, creates interview feel, and eliminates the warm-intro pipeline (the marketing director might introduce you to their team or peers if you are there in person).
3. Advisor evaluation: Reframe from "I don't want to waste their time" (needy framing) to "I want to find out if this person really understands the reporting challenge well enough to be a useful advisor"

**Output (abbreviated):**
```
## Recommended Format
**Scheduled Meeting (30 minutes)**

### Why This Format
You have a warm introduction to a senior person, which makes a meeting
request natural and appropriate. A phone call would work logistically
but costs you two things: body language observation (you will see their
reaction when they describe report-building pain) and the opportunity
for them to introduce you to peers on the spot. Those in-person warm
intros are extremely valuable and impossible over the phone.

### Timing
- Actual conversation time needed: 15-20 minutes of questions
- Total time investment: ~2 hours (travel + meeting + review)
- Duration guidance: Request 30 minutes. Spend the first 5 on greetings,
  15-20 on your questions, and the last 5 asking who else you should
  talk to.

### Approach
- Mindset: You are evaluating whether this marketing director
  has deep enough reporting pain to be a useful advisor for your space.
  You are not asking for a favor — you are seeing if they qualify.
- Formality check: If you open with "Thanks for taking the time..."
  you are signaling that they are doing you a favor. Instead, open
  with genuine curiosity about their work.
```

---

### Example 3: Early-Stage Founder Defaulting to Formal Meetings

**Scenario:** A solo founder is exploring a productivity app idea and has been scheduling 1-hour Zoom calls with anyone who will talk to them. After 3 weeks, they have completed only 4 conversations and are feeling overwhelmed by the time investment.

**Trigger:** "I've been doing hour-long Zoom interviews but I've only managed 4 in three weeks. This is taking forever. Am I doing something wrong?"

**Process:**
1. Depth assessment: Validate-existence level (5 minutes needed) — the founder is still in early exploration and does not yet know if the problem is real
2. Format scoring: Flag the Meeting Anti-Pattern immediately — the founder is relegating every conversation to a 1-hour calendar block when they need 5-minute casual chats. Four conversations in three weeks is a symptom of format mismatch, not a scheduling problem.
3. Advisor evaluation: The founder's frustration comes from the needy dynamic — each call feels like a big ask, which makes scheduling emotionally draining

**Output (abbreviated):**
```
## Recommended Format
**Casual Conversation**

### Why This Format
You are falling into the Meeting Anti-Pattern — relegating every
conversation to a formal calendar block. At your current stage (early
exploration, pre-product), you need to learn whether the problem you
are exploring actually exists and matters to people. That takes about
5 minutes of conversation, not 60.

Your current setup costs roughly 4 hours per conversation (scheduling,
prep, the call itself, review). At that rate, 4 conversations in 3
weeks is actually the expected throughput — the problem is the format,
not your hustle.

Switch to casual conversations. Go where your target users already are
(coworking spaces, meetups, online communities). Strike up conversations
about their work. You should be able to have 5-10 of these in a single
outing, each taking 5-10 minutes.

### Formality Warning Signs
You mentioned "hour-long Zoom interviews" — the word "interview" itself
is a red flag. Learning from customers does not require wearing a suit
and drinking boardroom coffee. If you are using rating scales, structured
interview guides, or opening with "Thanks for agreeing to this," you
have made it too formal.

### Preparation Checklist
- [ ] Know your 3 key questions (the 3 most important things you want
      to learn — use `conversation-question-designer` if needed)
- [ ] Identify 2-3 places where your target users gather (online or
      in person) — use `conversation-sourcing-planner` for ideas
- [ ] Practice your opening: a casual question about their work, not
      a request for an interview
```

## Key Principles

- **Format should match learning depth, not social convention** — A 1-hour formal meeting is appropriate when you need a deep industry dive or product feedback session. For validating whether a problem exists, 5 minutes of casual conversation produces the same learning with a fraction of the time cost. Let what you need to learn determine how you have the conversation, not the arbitrary conventions of calendar software.

- **Formality biases responses** — When a conversation feels like an interview, people give interview answers: polished, diplomatic, and unhelpful. When it feels like a chat between peers, they tell you what actually happens. If it feels like they are doing you a favor by talking to you, the conversation is probably too formal. The best customer conversations happen when the other person does not even realize it was a "meeting."

- **The Meeting Anti-Pattern is the most common format mistake** — First-time founders default to scheduling formal meetings for every customer interaction because it feels structured and professional. But the overhead (scheduling, commuting, prep, review) means a 1-hour meeting actually costs 4 hours. At that rate, you will never talk to enough people. Casual conversations at events and in communities let you compress a dozen learnings into a single outing.

- **Evaluate advisors, do not seek validation** — The default mindset going into customer conversations is needy: you want their approval, their time, their enthusiasm. Flip this by going in to evaluate whether they would make a good advisor for your space. You are looking for helpful, knowledgeable people. This reframe changes the power dynamic without changing the conversation topics, and it works regardless of format.

- **Phone calls are a considered trade-off, not a default** — Phone and video calls are convenient but sacrifice body language observation, weaken the power dynamic, and eliminate the possibility of in-person warm introductions. Use them when geography makes in-person impossible or when following up with someone you have already met, not as a way to avoid the awkwardness of face-to-face conversation.

## References

- For sourcing conversation opportunities and outreach framing, use the `conversation-sourcing-planner` skill
- For designing specific questions to ask during the conversation, use the `conversation-question-designer` skill
- For evaluating commitment signals after the conversation, use the `commitment-signal-evaluator` skill

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Mom Test by Rob Fitzpatrick.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-conversation-sourcing-planner`
- `clawhub install bookforge-conversation-question-designer`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

---
name: conversation-data-quality-analyzer
description: Analyze customer conversation notes or transcripts after a meeting to classify every statement as fact, compliment, fluff, or idea — separating real signal from noise. Use this skill whenever the user wants to review interview notes, check whether a customer call produced reliable data, figure out if enthusiastic feedback was genuine interest or polite lies, identify bad data patterns in a transcript, audit whether a conversation that "went great" actually produced usable facts, or suspects they are collecting compliments instead of validated evidence — even if they don't mention "data quality" or "bad data." Do NOT use this skill to write or improve questions before a conversation (use conversation-question-designer) or to evaluate whether a meeting produced real commitment signals like time, reputation, or money (use commitment-signal-evaluator).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-mom-test/skills/conversation-data-quality-analyzer
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: verified
source-books:
  - id: the-mom-test
    title: "The Mom Test"
    authors: ["Rob Fitzpatrick"]
    chapters: [2]
tags: [customer-discovery, conversation-analysis, bad-data, data-quality, compliments, fluff, false-positives]
depends-on:
  - conversation-question-designer
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "Conversation notes, transcript, or meeting summary from a customer conversation"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment with file read/write access."
---

# Conversation Data Quality Analyzer

## When to Use

You have notes or a transcript from a customer conversation and need to know how much of the data is actually reliable. Typical situations:

- The user just finished a customer conversation and wants the notes reviewed before acting on them
- The user has meeting notes and is about to update their product roadmap or assumptions based on them
- The user says a meeting "went really well" and wants to verify whether it actually produced signal
- The user has a transcript and wants to extract only the validated facts
- The user suspects they are getting false positives from conversations but cannot identify why
- The user has multiple conversation notes and wants a pattern analysis across them

Before starting, verify:
- Does the user have conversation notes, a transcript, or a meeting summary to analyze? (If not, this skill does not apply)
- Does the user know what product or problem area the conversation was about? (Needed to distinguish relevant facts from noise)

**Mode: Hybrid** -- The agent analyzes conversation notes or transcripts after the fact. The human conducted the actual conversation.

## Context & Input Gathering

### Required Context (must have -- ask if missing)

- **Conversation notes or transcript:** The raw material to analyze. This is the core input.
  - Check prompt for: pasted notes, quotes, transcript text, meeting summary
  - Check environment for: `conversation-notes/`, `meeting-*.md`, `transcript-*.md`, `notes-*.md`
  - If still missing, ask: "Please provide the conversation notes or transcript you want analyzed. You can paste them directly or point me to a file."

- **Product idea or problem area:** What the conversation was about. This is needed to distinguish relevant behavioral facts from unrelated statements.
  - Check prompt for: product descriptions, problem statements, startup ideas
  - Check environment for: `product-idea.md`, `README.md`
  - If still missing, ask: "What product or problem area was this conversation about? A sentence is enough."

### Observable Context (gather from environment)

- **Learning goals for the conversation:** What the user wanted to learn. This lets the analysis prioritize which facts matter.
  - Look for: `question-script.md`, `learning-log.md`, learning goals mentioned in the prompt
  - If unavailable: analyze all statements without goal-based prioritization

- **Previous conversation notes:** Past learnings for cross-referencing patterns
  - Look for: `conversation-notes/`, `learning-log.md`
  - If unavailable: analyze standalone

### Default Assumptions

- If no learning goals specified, analyze all customer statements without filtering by relevance
- If no product idea specified, classify data types but skip relevance assessment
- If notes are sparse (fewer than 5 statements), produce a shorter assessment with a note about limited data

### Sufficiency Threshold

```
SUFFICIENT when ALL of these are true:
- Conversation notes or transcript is available
- Product idea or problem area is known (or defaulted)

PROCEED WITH DEFAULTS when:
- Notes are available but no learning goals
- Product idea is approximate

MUST ASK when:
- No conversation notes, transcript, or meeting summary at all
```

## Process

### Step 1: Parse the Conversation into Individual Statements

**ACTION:** Read through the conversation notes or transcript and extract every distinct customer statement. Number each statement for reference. Separate the interviewer's statements from the customer's.

**WHY:** Bad data hides inside longer passages. A single paragraph of notes might contain one genuine fact, two compliments, and a piece of fluff. By atomizing the conversation into individual statements, each one can be classified independently. Without this step, compliments contaminate nearby facts -- a genuine workflow description followed by "this sounds really useful!" makes the whole passage feel validated when only the workflow part was real data.

**IF** the input is a raw transcript with speaker labels, extract only the customer's statements
**IF** the input is unstructured notes without clear attribution, flag any ambiguous statements where it is unclear whether the customer or the interviewer said it
**IF** the input contains the interviewer's questions, preserve them alongside the customer responses (the question quality affects the response quality)

### Step 2: Classify Each Statement

**ACTION:** Classify every customer statement into one of four categories using the classification rules below.

**WHY:** The four categories represent fundamentally different levels of reliability. Facts are usable data. Compliments, fluff, and ideas feel like data but carry zero validated signal. Misclassifying even one compliment as a fact can send a product team in the wrong direction for months. The classification system catches the specific ways customer conversations produce false positives.

**Classification rules:**

| Category | Definition | Detection Markers | Reliability |
|----------|-----------|-------------------|-------------|
| **FACT** | A concrete, specific statement about past behavior, current workflow, measurable outcome, or observable reality | Past tense, specific details (names, dates, numbers, tools), describes what actually happened, references current process | HIGH -- usable data |
| **COMPLIMENT** | Positive feedback about your idea, product, or pitch that costs the speaker nothing to say | "That's cool," "I love it," "Sounds great," "That would be awesome," any praise directed at your concept rather than describing their situation | ZERO -- entirely worthless |
| **FLUFF** | Vague, non-specific feedback that sounds like data but lacks concrete grounding | Three shapes: (1) Generic claims: "I usually," "I always," "I never"; (2) Future-tense promises: "I would," "I will," "I would definitely buy that"; (3) Hypothetical maybes: "I might," "I could," "I could see myself" | ZERO -- unreliable |
| **IDEA** | Feature requests, suggestions, or solutions proposed by the customer | "You should add...," "It would be great if...," "Can you make it do...," "What about syncing to...," "Have you thought about..." | LOW -- understand motivation, do not obey the request |

**FOR EACH** customer statement:
1. Assign a category: FACT, COMPLIMENT, FLUFF, or IDEA
2. Note which specific markers triggered the classification
3. If FLUFF, identify which of the 3 shapes it matches (generic claim, future-tense promise, hypothetical maybe)
4. If a statement contains mixed types (e.g., a fact followed by a compliment), split and classify each part separately
5. Assess confidence: CLEAR (unambiguous classification) or BORDERLINE (could go either way -- explain why)

### Step 3: Detect Anti-Patterns in the Conversation

**ACTION:** Scan the full conversation for the 6 named anti-patterns that systematically produce bad data. Check both the interviewer's behavior (from questions asked) and the customer's responses.

**WHY:** Individual bad statements are symptoms. Anti-patterns are the disease. A conversation might contain 10 compliments because the interviewer was fishing for them, or 8 pieces of fluff because the interviewer asked fluff-inducing questions. Identifying the root-cause anti-pattern tells the user not just what went wrong in this conversation, but what to fix for the next one.

**Anti-patterns to detect:**

| Anti-Pattern | What It Is | Detection Signals |
|-------------|-----------|-------------------|
| **Compliment Acceptance** | Treating compliments as validation data | Notes say "meeting went really well," "getting positive feedback," "everybody loves it." Compliments recorded as key takeaways. No facts to back up positive assessment. |
| **Fishing for Compliments** | Intentionally seeking approval instead of truth | Interviewer asked opinion-seeking questions: "Do you think it will work?", "Do you like it?" Questions that start with "Do you think my..." or "What do you think of our..." |
| **Ego Exposure Bias** (The Pathos Problem) | Accidentally triggering protective compliments by revealing how much you care | Interviewer shared personal stakes ("I quit my job for this," "This is my passion project"), asked for honest feedback while clearly invested ("I can take it -- tell me what you really think") |
| **Accepting Fluff** | Treating vague generics as concrete evidence | Fluff statements recorded without follow-up anchoring questions. No "when was the last time" or "can you walk me through" questions after generic claims. Fluffy responses accepted at face value. |
| **Being Pitchy** | Talking about the product more than listening | Interviewer talked for extended stretches without asking questions. Multiple product descriptions or feature explanations. Customer responses get shorter as conversation progresses. Symptoms: "No no, I don't think you get it," "Yes, but it also does this!" |
| **Obeying Feature Requests** | Adding feature requests to the roadmap without digging into the motivation behind them | Ideas recorded as action items without "why do you want that?" follow-up. Feature requests listed without underlying motivation. No "how are you coping without it?" questions. |

**IF** the input includes interviewer questions, check each question for fluff-inducing patterns: "Do you ever...", "Would you ever...", "What do you usually...", "Do you think you...", "Might you...", "Could you see yourself..."
**IF** the conversation ended with a compliment and no commitment, flag as potential zombie lead pattern

### Step 4: Assess Overall Conversation Quality

**ACTION:** Calculate a data quality summary and assess whether the conversation produced actionable signal.

**WHY:** A conversation can feel productive while producing zero usable data. By quantifying the ratio of facts to noise, the user gets an objective measure of conversation quality instead of relying on their gut feeling (which is exactly what compliments exploit). This assessment also reveals whether the conversation is worth following up on or whether it needs to be re-run with better questions.

**Calculate:**
- Total customer statements analyzed
- Count and percentage by category (FACT / COMPLIMENT / FLUFF / IDEA)
- Number of anti-patterns detected
- Overall quality rating:
  - **STRONG:** More than 60% facts, zero or one anti-patterns detected, facts include specific past behaviors
  - **MIXED:** 30-60% facts, or facts present but contaminated by significant fluff/compliments
  - **WEAK:** Less than 30% facts, or multiple anti-patterns detected, or no concrete past behaviors
  - **EMPTY:** Mostly compliments and fluff with no usable facts

**IF** quality is WEAK or EMPTY, recommend which questions to ask in a follow-up conversation to recover the missing data (reference the `conversation-question-designer` skill for designing those questions)

### Step 5: Extract the Cleaned Insights Summary

**ACTION:** Produce a cleaned insights document containing only validated facts, with compliments and fluff stripped out. For each idea (feature request), include the underlying motivation if it was uncovered.

**WHY:** The whole point of this analysis is to produce a document the user can trust and act on. The cleaned summary separates what was actually learned from what felt learned. This is the artifact that should inform product decisions -- not the raw notes, which contain a mix of signal and noise.

**Cleaned summary structure:**

```markdown
# Conversation Data Quality Assessment

## Metadata
- **Date:** [conversation date if known]
- **Customer/Interviewee:** [name or identifier]
- **Product/Problem Area:** [from context]
- **Conversation Quality:** [STRONG / MIXED / WEAK / EMPTY]

## Quality Breakdown
| Category | Count | Percentage |
|----------|-------|------------|
| Facts | X | X% |
| Compliments | X | X% |
| Fluff | X | X% |
| Ideas | X | X% |

## Validated Facts (usable data)
1. [Fact statement] — *[why this is reliable: specific past behavior, named tool, concrete number]*
2. [Fact statement] — *[why this is reliable]*
...

## Ideas with Motivations (understand, do not obey)
1. **Request:** [what they asked for]
   **Underlying motivation:** [why they want it, if uncovered]
   **Currently coping by:** [their workaround, if mentioned]
   *If motivation was NOT explored:* Flag for follow-up — ask "Why do you want that?" and "How are you coping without it?"

## Discarded Data (compliments and fluff)
| # | Statement | Category | Why Discarded |
|---|-----------|----------|---------------|
| 1 | [statement] | COMPLIMENT | [explanation] |
| 2 | [statement] | FLUFF (generic claim) | [explanation] |
...

## Anti-Patterns Detected
- **[Pattern name]:** [evidence from conversation] → [what to fix next time]

## Recommendations
- [Specific actions: follow-up questions to ask, questions to redesign, anti-patterns to watch for]
```

**IF** the user provided a file path or working directory, write the output to `data-quality-assessment.md`
**ELSE** present the output directly in the conversation

## Examples

### Example 1: Founder reviews notes from a "great" meeting about a project management tool

**Scenario:** A founder building a project management tool for freelancers says: "I had a great conversation with a freelance designer yesterday. Here are my notes."

**Trigger:** User provides meeting notes containing a mix of facts, compliments, and fluff.

**Process:**
1. Parse notes into individual statements
2. Classify each statement:
   - "I usually manage projects with sticky notes and email" -> FLUFF (generic claim: "I usually")
   - "Last month I lost a client because I sent the wrong version of a file" -> FACT (specific past event with consequence)
   - "Your tool sounds amazing, I would totally use it" -> COMPLIMENT + FLUFF ("sounds amazing" = compliment, "I would totally use it" = future-tense promise)
   - "I have 6 active clients right now and track them in a spreadsheet" -> FACT (specific current state with number)
   - "You should add a Figma integration" -> IDEA (feature request)
   - "It would save me so much time" -> FLUFF (hypothetical future)
3. Detect anti-patterns: Ego Exposure Bias (founder mentioned "my tool" triggering compliments), Obeying Feature Requests (Figma integration noted without asking why)
4. Quality: MIXED -- 2 facts (33%), 1 compliment, 2 fluff, 1 idea

**Output (abbreviated):**
```
Conversation Quality: MIXED

Validated Facts:
1. Lost a client last month due to sending wrong file version — specific past event with real consequence
2. Has 6 active clients tracked in a spreadsheet — concrete current workflow with number

Ideas with Motivations:
1. Request: Figma integration
   Underlying motivation: NOT EXPLORED — follow up with "Why do you want that?" and "What would that let you do?"

Discarded:
- "I usually manage projects with sticky notes" — FLUFF (generic claim, ask "walk me through your last project")
- "Your tool sounds amazing" — COMPLIMENT (zero data)
- "I would totally use it" — FLUFF (future-tense promise, the world's most deadly fluff)

Anti-Patterns:
- Ego Exposure Bias: Mentioned "my tool" which triggered protective compliments
- Obeying Feature Requests: Figma integration recorded without motivation digging

Recommendations:
- Follow up to anchor the "sticky notes and email" claim: "Walk me through how you managed your most recent client project from start to finish"
- Dig into Figma integration request: "Why do you want that? How are you coping without it?"
- Use conversation-question-designer to redesign questions that avoid mentioning the product
```

---

### Example 2: Product manager analyzes a transcript full of enthusiastic but empty responses

**Scenario:** A product manager shares a transcript from a call about an automated reporting feature. The customer was very enthusiastic throughout.

**Trigger:** User provides a transcript where the customer gave many positive responses but few concrete details.

**Process:**
1. Parse transcript into 15 customer statements
2. Classify: 2 FACTS, 7 COMPLIMENTS, 4 FLUFF, 2 IDEAS
3. Detect anti-patterns: Being Pitchy (PM talked for 60% of the conversation), Compliment Acceptance (PM noted "great call" at the end), Fishing for Compliments ("Don't you think automated reports would save time?")
4. Quality: WEAK -- 13% facts, 3 anti-patterns

**Output (abbreviated):**
```
Conversation Quality: WEAK

Validated Facts:
1. Team spends "nearly a week at the end of each month" pulling reports together — specific time cost
2. Currently using Excel and email to coordinate — specific current tools

Discarded (13 statements):
- "That's exactly what we need!" — COMPLIMENT
- "I would definitely buy that" — FLUFF (future-tense promise; the deadliest fluff)
- "We'd probably use it every day" — FLUFF (hypothetical maybe + future-tense)
- "You should integrate with Salesforce" — IDEA (motivation not explored)
...

Anti-Patterns:
- Being Pitchy: PM described the product for 4 consecutive exchanges without asking a question
- Fishing for Compliments: Asked "Don't you think automated reports would save time?" (opinion + hypothetical)
- Compliment Acceptance: PM wrote "great call" in notes — the 2 facts do not justify this assessment

Recommendations:
- This conversation needs a complete redo with the conversation-question-designer skill
- The 2 facts are valuable — dig deeper with: "Walk me through the last time you pulled the monthly report together"
- Do NOT put "Salesforce integration" on the roadmap until you ask: "Why Salesforce specifically? What would that let you do that you can't do now?"
```

---

### Example 3: Analyzing sparse notes from a casual conversation

**Scenario:** A founder jotted down quick notes from a casual coffee chat: "Met Sarah, runs a bakery. Says inventory is a nightmare. Would love an app for it. Gave me her card."

**Trigger:** User provides minimal notes from an informal conversation.

**Process:**
1. Parse into 3 statements: "inventory is a nightmare" (borderline -- emotional but vague), "would love an app for it" (FLUFF), "gave me her card" (FACT -- concrete action)
2. Quality: WEAK -- too few data points, one fact is a social gesture not a commitment

**Output (abbreviated):**
```
Conversation Quality: WEAK (insufficient data)

Validated Facts:
1. Gave business card — concrete action, but this is a social gesture, not a commitment signal. A card costs nothing.

Borderline (needs anchoring):
- "Inventory is a nightmare" — emotional signal worth digging into, but currently vague. Follow up: "Tell me about the last time inventory caused a problem. What happened?"

Discarded:
- "Would love an app for it" — FLUFF (future-tense promise)

Note: This conversation produced too little data to draw conclusions. The emotional signal about inventory is promising but unvalidated. Recommended follow-up questions:
- "When was the last time inventory went wrong? Walk me through what happened."
- "What are you using to track inventory right now?"
- "How much time do you spend on inventory management each week?"
- "What have you tried to fix this?"
```

## Key Principles

- **Compliments are the fool's gold of customer learning: shiny, distracting, and entirely worthless** -- Even genuine compliments contain zero usable data. A venture capitalist (a professional judge of the future) is wrong far more often than right. A random person's opinion has even less weight. The only exception is industry experts who have built very similar businesses. Everyone else's approval is noise.

- **The deadliest fluff sounds the most concrete** -- "I would definitely buy that" feels like a commitment because it uses decisive language. But it is a future-tense promise about a hypothetical, which means it is pure fluff. The first startup the author worked at fell for this trap and lost ten million dollars. Always check: did they describe something that already happened, or something they imagine happening?

- **Ideas and feature requests should be understood, but not obeyed** -- When a customer says "you should add Excel sync," the wrong response is to write "Excel sync" on the roadmap. The right response is "Why do you want that? What would that let you do?" The request is a surface symptom. The underlying motivation is the real data. The author built three months of unnecessary analytics features for MTV because he accepted the feature request at face value instead of asking why.

- **Anti-patterns are the disease; bad data is the symptom** -- A conversation full of compliments usually means the interviewer was fishing for them or exposed their ego. A conversation full of fluff usually means the interviewer asked fluff-inducing questions. Fixing individual data points is pointless if the underlying conversational behavior remains. This skill diagnoses both.

- **A conversation that "went well" is a warning sign, not a positive signal** -- When you leave a meeting feeling great about it, that feeling is almost always driven by compliments. Genuine learning conversations often feel uncomfortable because you are hearing things you did not want to hear. The best conversations produce facts and commitments, not warm feelings.

## References

- For the complete bad data taxonomy with all fluff shapes, anti-pattern symptoms, and detection markers, see [bad-data-classification-guide.md](references/bad-data-classification-guide.md)
- For designing better questions before the next conversation, use the `conversation-question-designer` skill
- For evaluating whether a meeting produced real commitment signals (time, reputation, money), use the `commitment-signal-evaluator` skill

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Mom Test by Rob Fitzpatrick.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-conversation-question-designer`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/bad-data-classification-guide.md
# Bad Data Classification Guide

> Deep reference for the conversation-data-quality-analyzer skill.
> Contains the full taxonomy, detection markers, and recovery techniques.

## The Three Types of Bad Data

Bad data gives false negatives (thinking the idea is dead when it is not) and -- more dangerously -- false positives (convincing yourself you are right when you are not).

### 1. Compliments

**Definition:** Vacuous positive feedback that tricks you into thinking the meeting went well.

**Why they are worthless:** Even if they genuinely like it, that data is still worthless. Venture capitalists (professional judges of the future) are wrong far more often than right. If even a VC's opinion is probably wrong, a random person's opinion has even less weight. The only exception is industry experts who have built very similar businesses.

**Detection markers in statements:**
- "That's cool"
- "I love it"
- "Sounds great"
- "Sounds terrific"
- "Keep me in the loop"
- "That would be awesome"
- Any praise directed at your concept rather than describing their situation

**Detection markers in your own behavior (symptoms):**
- In the meeting: "Thanks!" or "I'm glad you like it"
- Back at the office: "That meeting went really well"
- Back at the office: "We're getting a lot of positive feedback"
- Back at the office: "Everybody I've talked to loves the idea"

**Recovery technique -- DEFLECT:**
- Ignore the compliment entirely
- Redirect to facts: "How are you dealing with this stuff at the moment?"
- Prevention: do not mention your idea at all

### 2. Fluff

**Definition:** Vague, non-specific feedback that sounds like data but lacks concrete grounding.

**Three shapes of fluff:**

| Shape | Trigger Phrases | Example |
|-------|----------------|---------|
| Generic claims | "I usually," "I always," "I never" | "I usually manage my projects pretty well" |
| Future-tense promises | "I would," "I will" | "I would definitely buy that" |
| Hypothetical maybes | "I might," "I could" | "I could see myself using something like that" |

**The world's most deadly fluff:** "I would definitely buy that." It sounds concrete. As a founder, you desperately want to believe it is money in the bank. But people are wildly optimistic about what they would do in the future. They are always more positive, excited, and willing to pay in the imagined future than they are once it arrives.

**Fluff-inducing questions (the interviewer's fault):**
- "Do you ever..."
- "Would you ever..."
- "What do you usually..."
- "Do you think you..."
- "Might you..."
- "Could you see yourself..."

These questions are not toxic in themselves -- they can transition into concrete questioning. The mistake is in valuing the answers, not asking the questions.

**Recovery technique -- ANCHOR:**
- Ask: "When's the last time that happened?"
- Ask: "Can you talk me through that?"
- Ask: "Can you walk me through a specific example?"
- Transition sequence: fluffy question -> fluffy answer -> anchor to past -> concrete details

**Anchoring example:**
- Them: "I'm an Inbox Zero zealot. It's totally changed my life." (generic claim)
- You: "What's your inbox at right now?" (anchor to present)
- Them: "Looks like about ten that have come in since this morning." (fact)
- You: "When's the last time it totally fell apart for you?" (anchor to past)
- Them: "Three weeks ago, I was travelling and the internet at the hotel totally didn't work. It took me like ten days to get back on track." (concrete fact)

### 3. Ideas (Feature Requests)

**Definition:** Feature requests, suggestions, or solutions proposed by the customer once they get excited about your concept.

**Why they are dangerous:** Startups are about focusing and executing on a single, scalable idea rather than jumping on every good one which crosses your desk. When you accept feature requests at face value, you get feature creep. The author built three months of unnecessary analytics features for MTV because he accepted "we need analytics" without asking why they wanted it. The real need was automated weekly emails to clients, not a self-serve analytics dashboard.

**Detection markers:**
- "You should add..."
- "Can you make it do..."
- "Are you going to be able to sync to..."
- "What about adding..."
- "It would be great if it could..."
- "Have you thought about..."

**Recovery technique -- DIG:**

Questions to dig into feature requests:
1. "Why do you want that?"
2. "What would that let you do?"
3. "How are you coping without it?"
4. "Do you think we should push back the launch to add that feature, or is it something we could add later?"
5. "How would that fit into your day?"

Questions to dig into emotional signals:
1. "Tell me more about that."
2. "That seems to really bug you -- I bet there's a story here."
3. "What makes it so awful?"
4. "Why haven't you been able to fix this already?"
5. "You seem pretty excited about that -- it's a big deal?"
6. "Why so happy?"
7. "Go on."

## The Six Anti-Patterns

### 1. Compliment Acceptance

**What it is:** Treating compliments as validation data.

**Symptoms:**
- Your meeting notes list compliments as positive signals
- You tell your team the meeting "went well" without citing specific facts
- Your pipeline is full of people who "loved it" but have not committed anything

**Root cause:** Compliments feel like data because we desperately want them to be.

**Fix:** Strip all compliments from your notes. If the remaining facts do not support your conclusion, the meeting did not actually go well.

### 2. Fishing for Compliments

**What it is:** Intentionally seeking approval instead of truth. You have already made up your mind and need someone's blessing to take the leap.

**Symptom phrases from the interviewer:**
- "I'm thinking of starting a business... so, do you think it will work?"
- "I had an awesome idea for an app -- do you like it?"

**Fix:** Do not mention your idea. Ask about their life instead.

### 3. Ego Exposure Bias (The Pathos Problem)

**What it is:** Accidentally triggering protective compliments by exposing how much you care about the idea. Once someone detects that your ego is on the line, they will give you fluffy half-truths and extra compliments. Even asking for honest criticism does not fix it -- people still pull their punches.

**Symptom phrases from the interviewer:**
- "So here's that top-secret project I quit my job for... what do you think?"
- "I can take it -- be honest and tell me what you really think!"

**Fix:** Keep the conversation focused on the other person. Ask about specific, concrete cases and examples. People rarely lie about specific stuff that has already happened, regardless of your ego.

### 4. Accepting Fluff

**What it is:** Treating generic claims, hypothetical future promises, and "I would definitely buy that" statements as meaningful signal.

**Detection:** Fluff statements in notes without corresponding anchoring follow-ups. No "when was the last time" or "can you walk me through" questions visible.

**Fix:** Anchor every generic claim to a specific past instance. If they cannot provide one, the claim is not data.

### 5. Being Pitchy

**What it is:** Holding someone hostage until they say they like your idea. The dark side of approval-seeking -- instead of being vulnerable, you are being annoying.

**Symptom phrases from the interviewer:**
- "No no, I don't think you get it..."
- "Yes, but it also does this!"

**Fix:** Apologize and redirect. "Whoops -- I just slipped into pitch mode. I'm really sorry about that. Can we jump back to what you were saying?"

### 6. Obeying Feature Requests

**What it is:** Adding feature requests to the roadmap without digging into the motivation behind them.

**Detection:** Ideas listed as action items or roadmap entries. No "why do you want that?" follow-up visible in the notes.

**Fix:** For every feature request, ask the digging questions (see Ideas section above). The request is a surface symptom. The underlying motivation is the real data.

## The Complainer vs Customer Distinction

When anchoring fluff, you may discover the person is a complainer, not a customer. The pattern looks like this:

- "Someone should definitely make an X!" -> Have you looked for an X? -> "No, why?" -> They are not your customer.
- "I would definitely use that!" -> Have you tried any existing solutions? -> "I downloaded a couple but they were annoying" -> "Why didn't you keep using them?" -> "I'll do it next time. Not a real problem." -> They are a complainer, not a customer.

Long story short, if someone is not willing to try solving the problem themselves, they are not going to care about your solution either. Complainers live in the la-la-land of imagining they are the sort of person who finds clever ways to solve the petty annoyances of their day.

## Quick Reference: The Deflect-Anchor-Dig Workflow

```
HEAR a compliment?     → DEFLECT: Ignore it. Ask about their current process.
HEAR fluff?            → ANCHOR: "When was the last time that happened?"
HEAR a feature request → DIG: "Why do you want that? What would that let you do?"
HEAR an emotion?       → DIG: "Tell me more about that. What makes it so [awful/exciting]?"
```

---
name: commitment-signal-evaluator
description: Evaluate whether a customer meeting produced real interest or just polite enthusiasm by classifying commitment signals into time, reputation, and money currencies. Use this skill after any customer conversation, product demo, sales call, or pitch where the user wants to know if the meeting actually advanced the deal, asks "was that a good meeting" or "how did it go" or "did that go well," says a meeting "went great" but nothing happened afterward, received enthusiastic feedback and wants to know if it is real, wonders whether someone is actually interested or just being polite, heard "I would definitely buy that" and wants to know if it means anything, wants to distinguish real leads from false-positive prospects (zombie leads), needs to score a pipeline of prospects for conversion likelihood, wants to detect polite rejections disguised as enthusiasm (compliment-stall pattern), or wants to identify early evangelists in their prospect pool — even if they don't explicitly mention "commitment signals" or "meeting evaluation." This skill evaluates meeting OUTCOMES and prospect INTEREST, not conversation data quality (use conversation-data-quality-analyzer) or conversation logistics (use conversation-format-selector).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-mom-test/skills/commitment-signal-evaluator
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: verified
source-books:
  - id: the-mom-test
    title: "The Mom Test"
    authors: ["Rob Fitzpatrick"]
    chapters: [5]
tags: [customer-discovery, commitment, sales-signals, meeting-evaluation, zombie-leads, earlyvangelists]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "Meeting notes, conversation transcript, or summary of a customer meeting outcome"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment with access to meeting notes or transcript files"
---

# Commitment Signal Evaluator

## When to Use

You have just finished (or are reviewing) a customer meeting, product demo, sales call, or investor conversation and need to assess whether the meeting produced real signals of interest or just polite enthusiasm. Typical triggers:

- The user shares meeting notes and asks "how did it go?" or "was that a good meeting?"
- The user has a pipeline of leads and wants to identify which are real vs false-positive prospects (zombie leads)
- The user received enthusiastic feedback but is unsure whether it translates to actual demand
- The user wants to know what commitment to push for at the next meeting
- The user is reviewing a batch of conversations and needs to score signal strength across prospects

**Preconditions to verify:**
- Does the user have meeting notes, a transcript, or at least a summary of what happened?
- Does the user know what product stage they are at? (Pre-product learning vs product demo vs active selling)

**This skill does NOT cover:**
- Evaluating whether the questions asked were good (use `conversation-data-quality-analyzer`)
- Structuring conversation learnings into validated/invalidated assumptions (use `conversation-learning-process`)

## Context & Input Gathering

### Required Context (must have -- ask if missing)

- **Meeting notes or transcript:** The raw material to evaluate. Without this, there is nothing to score.
  -> Check prompt for: pasted text, file path, quotes from the meeting, summary of what happened
  -> Check environment for: files in a `conversation-notes/` or `meetings/` directory
  -> If still missing, ask: "Can you share the meeting notes or paste the key moments from the conversation? I need the actual words and outcomes to evaluate signal strength."

- **Product stage:** Determines which commitment currencies are appropriate to expect. Asking a pre-product prospect for money is premature; not asking a post-demo prospect for a next step is a missed opportunity.
  -> Check prompt for: mentions of "prototype," "MVP," "beta," "launched," "idea stage," "we showed them the product"
  -> If still missing, ask: "What stage is your product at? (a) Still exploring the problem -- no product yet, (b) Have wireframes or prototype, (c) Have working product or beta, (d) Product is live and selling"

### Observable Context (gather from environment)

- **Previous meeting notes:** Past conversations with the same person or segment reveal whether commitment is escalating or stalling.
  -> Look for: files mentioning the same contact name, company, or segment
  -> If unavailable: evaluate this meeting in isolation

- **Commitment tracker:** An existing log of commitments received from various prospects.
  -> Look for: `commitment-tracker.md`, `pipeline.md`, `leads.csv`
  -> If unavailable: create one as part of the output

### Default Assumptions

- If product stage is unclear: assume early-stage (pre-product) and calibrate expectations accordingly. Early-stage meetings should produce time and reputation commitments, not financial ones.
- If only a brief summary is provided (not full notes): evaluate based on available signals but flag that the assessment is partial.

### Sufficiency Threshold

```
SUFFICIENT when ALL of these are true:
- Meeting notes or transcript with identifiable outcomes/statements are available
- Product stage is known or can be inferred

PROCEED WITH DEFAULTS when:
- Meeting notes are brief but contain at least one closing statement or outcome
- Product stage can be inferred from context

MUST ASK when:
- No meeting content is available at all
- The user describes a meeting but provides zero specific quotes or outcomes
```

## Process

### Step 1: Extract Meeting Outcomes and Closing Statements

**ACTION:** Read the meeting notes or transcript. Identify all statements, actions, and promises made by the prospect at or near the end of the meeting. Also flag any strong emotional reactions throughout.

**WHY:** The closing moments of a meeting reveal what the prospect is actually willing to do, not just what they said they felt. Enthusiasm during the meeting is cheap -- what they commit to afterward is the real signal. Emotional intensity mid-conversation (pain, frustration, excitement about their problem) also matters because it predicts earlyvangelist potential.

**Extract and list:**
- Direct quotes from the prospect (especially closing statements)
- Any promises made (introductions, follow-ups, purchases, trials)
- Any actions the prospect took during or after the meeting (signed up, scheduled, introduced)
- Emotional intensity markers (frustration with current solutions, excitement about the problem space)
- Anything the prospect gave up (time for next meeting, shared internal docs, made an intro)

### Step 2: Classify Each Signal Using Commitment Currencies

**ACTION:** For each extracted signal, classify it into one of four categories based on what the prospect gave up:

**Commitment Currencies (ranked by signal strength):**

| Currency | What They Give Up | Examples | Signal Strength |
|----------|------------------|----------|----------------|
| **No currency (compliment)** | Nothing | "That's so cool!", "Love it!", "Looks great" | ZERO -- costs them nothing, worth nothing |
| **Time** | Their calendar, attention, effort | Clear next meeting with specific goals, sitting down to review wireframes, using a trial for a non-trivial period | LOW-MEDIUM |
| **Reputation** | Their social capital and credibility | Intro to peers or team, intro to decision maker (boss, spouse, lawyer), public testimonial or case study | MEDIUM-HIGH |
| **Money** | Cash, budget, purchasing authority | Letter of intent, pre-order, deposit, pulling out a credit card | HIGHEST |

**WHY:** A compliment costs the prospect nothing, so it carries zero informational value. The more they give up, the more seriously you can take their words. This is the core insight: words are free, actions cost something. Classifying by currency type immediately separates real signals from noise. Strong commitments often combine multiple currencies -- for example, agreeing to a paid trial with their whole team risks time, money, AND reputation.

**IF** a signal combines multiple currencies -> note it as a compound commitment (strongest type)
**IF** a signal is purely verbal with no cost to the prospect -> classify as compliment (zero signal)

### Step 3: Score Each Signal as Good Meeting / Bad Meeting

**ACTION:** Evaluate each closing statement against the meeting outcome scoring rubric:

| What They Said | Verdict | Why | Fix |
|----------------|---------|-----|-----|
| "That's so cool. I love it!" | BAD | Pure compliment. Zero data. May feel good but contains no actionable information. | Deflect the compliment, ask about their current process or next steps. |
| "Looks great. Let me know when it launches." | BAD | Polite rejection disguised as enthusiasm (compliment-stall). A polite way to say "Don't call me, I'll call you." The big error is mistaking this for a pre-order. | Ask for a commitment available today: intro to their boss, agreement to be a beta tester, or a letter of intent. |
| "There are a couple people I can intro you to when you're ready." | BAD (currently) | Some signal exists but the promise is too vague to act on. "When you're ready" is a deferral. Who specifically? Ready for what? | Convert fuzzy promise to concrete specifics: "Who did you have in mind? Could we set that up this week?" |
| "What are the next steps?" | GOOD | They are actively advancing the deal. But you must know YOUR next steps to benefit -- if you say "let me think about it and get back to you," you have ruined a good meeting. | Have your next steps prepared before the meeting. Always know what commitment you will ask for. |
| "I would definitely buy that." | BAD (dangerous) | This is the most dangerous false positive. Future-tense promises are unreliable. This exact phrase has sunk startups that treated it as validation. | Shift to concrete current commitment: letter of intent, pre-order, deposit, or intros to other decision makers. |
| "When can we start the trial?" | MAYBE GOOD | Depends on what the trial costs them. A free trial of consumer software costs nothing (low signal). A trial that requires training staff and integrating systems costs real time and reputation (high signal). | Evaluate the trial in terms of currency: what are they actually giving up to try it? If too cheap, increase the cost -- ask for a case study commitment after 2 weeks. |
| "Can I buy the prototype?" | GREAT | Strongest possible signal. They want it so badly they will pay for an unfinished version. This is rare and precious. | Take their money. Keep them close. This person is a potential earlyvangelist. |
| "When can you come back to talk to the rest of the team?" | GOOD | Reputation commitment -- they are willing to stake their credibility by bringing you to their colleagues. In enterprise sales, this is a strong advancement signal. If they will not introduce you, it is a dead end. | Schedule the meeting immediately. Prepare for a different audience (technical team vs decision maker). |

**WHY:** Every meeting either succeeds (produces commitment and advancement) or fails (produces only compliments or stalling). There is no such thing as a meeting that "went well" -- that phrase is itself a warning sign. Scoring each outcome forces an honest assessment and prevents the natural human tendency to interpret politeness as progress.

### Step 4: Detect Anti-Patterns

**ACTION:** Scan the meeting outcomes for two critical anti-patterns:

**Anti-Pattern 1: False-Positive Prospects (Zombie Leads)**
Symptoms:
- A pipeline of prospects who keep taking meetings but never convert
- Meetings that end with compliments but no clear next steps
- Meetings described as having "gone well"
- The prospect has not given up anything of value across multiple meetings

**WHY:** Zombie leads are caused by avoiding the scary question -- the one that could get you rejected. By failing to push for commitment, you stay in a comfortable no-man's-land where the prospect keeps being friendly but never moves forward. This is like being friend-zoned by your startup. The fix is to give them a concrete chance to reject you, because rejection is data and no-man's-land is not.

**Anti-Pattern 2: Polite Rejection Disguised as Enthusiasm (Compliment-Stall)**
Symptoms:
- Enthusiastic words with no concrete follow-through
- Future-tense promises: "I would...", "When you launch...", "Eventually..."
- Vague offers: "I know some people...", "We should definitely..."
- The prospect defers all action to an unspecified future date

**WHY:** The compliment-stall is the most common way meetings fail. The prospect is being polite, not committed. Every flavor of "let me know when it's ready" is a polite way to say "Don't call me, I'll call you." The fix is to look for a commitment available today, not in the future.

**IF** zombie lead pattern detected -> flag the entire relationship, not just this meeting
**IF** compliment-stall detected -> recommend specific commitment asks for the next interaction

### Step 5: Evaluate Earlyvangelist Potential

**ACTION:** Check whether the prospect meets the four criteria for an earlyvangelist (early evangelist, a term from Steve Blank's Customer Development methodology) -- the rare person who will become your first real customer:

1. **Have the problem** -- They experience the pain your product addresses
2. **Know they have the problem** -- They are consciously aware of it, not just theoretically affected
3. **Have the budget to solve it** -- They can actually spend money on a solution
4. **Have already cobbled together their own makeshift solution** -- They care enough that they built a workaround

**Emotional intensity signal:** There is a significant difference between "Yeah, that's a problem" and someone who is deeply frustrated, emotional, or excited about the problem space. When you see deep emotion, keep that person close -- they are a rare potential first customer.

**WHY:** First customers are irrational in a good way. They want what you are making so badly they will be the first to try it, knowing the risks. Identifying them early is critical because they will front money on pre-orders, tell friends, give feedback, and become your case studies. Not every interested prospect is an earlyvangelist -- most are just being polite.

**IF** prospect meets all 4 criteria -> flag as HIGH PRIORITY earlyvangelist candidate
**IF** prospect meets 3 of 4 -> flag as POTENTIAL, note which criterion is missing
**IF** prospect shows deep emotion about the problem -> flag regardless of other criteria, investigate further

### Step 6: Determine Overall Meeting Score and Recommend Next Steps

**ACTION:** Produce the final meeting assessment with:

1. **Overall verdict:** GOOD MEETING, BAD MEETING, or MIXED SIGNALS
2. **Signal strength:** Based on highest commitment currency obtained
3. **Commitment level:** What they actually gave up (or did not)
4. **Anti-patterns detected:** Zombie lead risk, compliment-stall instances
5. **Earlyvangelist score:** How many of the 4 criteria are met
6. **Recommended next step:** What specific commitment to push for in the next interaction, calibrated to your product stage

**Stage-appropriate commitment targets:**

| Product Stage | Appropriate Commitment Ask |
|---------------|---------------------------|
| Problem exploration (no product) | Time: next meeting with specific learning goals. Reputation: intro to others with the same problem. |
| Wireframes or prototype | Time: sit down to review wireframes and give feedback. Reputation: intro to decision maker or technical team. |
| Working product or beta | Time: non-trivial trial period. Reputation: case study or testimonial. Money: letter of intent or pre-order. |
| Live product | Money: purchase, deposit, or pre-order. Reputation: public testimonial. Time: team-wide rollout. |

**WHY:** A meeting without a next step was a pointless meeting. Knowing what to ask for -- calibrated to your stage -- prevents two failure modes: asking for too much too early (scares them off) and asking for too little when you should be closing (creates zombie leads). The goal is always advancement: moving the relationship to the next concrete step in your funnel.

**HANDOFF TO HUMAN:** The agent produces the scored assessment and recommended next steps. The human executes the follow-up in the next conversation.

## Outputs

### Meeting Commitment Assessment Report

Write the report to a file the user specifies, or to `meeting-assessment-{date}.md`:

```markdown
# Meeting Commitment Assessment

## Meeting Details
- **Date:** {date}
- **Prospect:** {name/company}
- **Product stage:** {stage}
- **Meeting type:** {learning / product demo / sales / follow-up}

## Overall Verdict: {GOOD MEETING / BAD MEETING / MIXED SIGNALS}

## Signal Analysis

| Signal | Quote/Action | Currency | Strength | Verdict |
|--------|-------------|----------|----------|---------|
| {signal 1} | "{exact quote}" | {time/reputation/money/none} | {zero/low/medium/high} | {good/bad/mixed} |
| {signal 2} | ... | ... | ... | ... |

## Commitment Level
- **Highest currency obtained:** {none / time / reputation / money / compound}
- **What they gave up:** {specific description}
- **What they did NOT give up:** {what was expected but missing}

## Anti-Pattern Check
- [ ] Zombie lead risk: {yes/no — with evidence}
- [ ] Compliment-stall detected: {yes/no — with evidence}
- [ ] Future-tense promises without current action: {yes/no — with evidence}

## Earlyvangelist Assessment
- [ ] Has the problem: {yes/no/unknown}
- [ ] Knows they have it: {yes/no/unknown}
- [ ] Has budget to solve it: {yes/no/unknown}
- [ ] Already built makeshift solution: {yes/no/unknown}
- **Emotional intensity:** {low / moderate / high}
- **Earlyvangelist likelihood:** {HIGH PRIORITY / POTENTIAL / UNLIKELY}

## Recommended Next Steps
1. **Immediate action:** {what to do within 48 hours}
2. **Next meeting commitment ask:** {specific commitment to request, calibrated to stage}
3. **If they decline:** {what that tells you and how to respond}

## Key Quote
> "{the single most informative quote from the meeting}"
```

## Examples

**Scenario: SaaS founder reviews demo meeting notes**
Trigger: "I just demoed our project management tool to a mid-size agency. They said 'This looks amazing, we'd definitely use this! Let me know when you have the team plan ready.' How did it go?"
Process: Extracted closing statement. Classified "This looks amazing" as compliment (zero currency). Classified "we'd definitely use this" as future-tense promise (zero currency -- most dangerous false positive). Classified "let me know when you have the team plan ready" as compliment-stall (deferring action to unspecified future). No currency exchanged. Detected compliment-stall anti-pattern. Product stage is working product (beta). Checked earlyvangelist criteria: they have the problem (project management pain) and know it, but no evidence of budget or existing workaround.
Output: BAD MEETING. Zero commitment currencies exchanged. Compliment-stall detected. Recommended next step: "Go back and ask for a specific commitment available today. Since you have a working beta, ask them to trial it with one project team for 2 weeks, and to write a brief case study afterward. If they say yes, it's a real lead. If they hesitate, you know where you stand. Do not wait for them to contact you."

**Scenario: Founder evaluates a pipeline of 5 prospects**
Trigger: "I've had meetings with 5 potential customers over the past month. Here are my notes. Which ones are real leads?"
Process: Read all 5 meeting note files. For each prospect, extracted closing statements and classified commitment currencies. Prospect A: gave an intro to their CTO (reputation currency -- GOOD). Prospect B: said "love it, keep me posted" across 3 meetings (compliment-stall, zombie lead -- BAD). Prospect C: asked to buy early access (money currency -- GREAT, earlyvangelist candidate). Prospect D: scheduled a follow-up with specific agenda (time currency -- GOOD). Prospect E: enthusiastic but cancelled the follow-up twice (zombie lead -- BAD).
Output: Pipeline assessment ranking all 5 prospects by commitment strength. Prospect C flagged as earlyvangelist (met all 4 criteria). Prospects B and E flagged as zombie leads with recommendation to force a decision (ask for commitment or accept the rejection). Prospects A and D rated as advancing with recommended next commitment asks.

**Scenario: Pre-product founder assessing problem interviews**
Trigger: "I've been doing problem interviews for my HR analytics idea. One person said 'Oh my god, I spend 3 hours every week manually pulling data from our HRIS into spreadsheets. I even wrote a janky Python script to automate part of it.' Is this a good signal?"
Process: Extracted signal. No commitment currency exchanged yet (this is a learning conversation, not a product meeting). However, identified strong earlyvangelist indicators: has the problem (manual data pulling), knows they have it (quantified at 3 hours/week), and has already cobbled together a makeshift solution (Python script). Budget unknown. Emotional intensity is high (the "Oh my god" opener indicates genuine frustration). No compliment-stall -- this is concrete past behavior, not future promises.
Output: STRONG SIGNAL (even though no commitment was asked for -- this is pre-product stage). This person meets 3 of 4 earlyvangelist criteria. Recommended next step: "In your next conversation with this person, verify budget authority and ask for a time commitment: 'Would you be willing to spend 30 minutes reviewing wireframes when we have a prototype? And could you introduce me to others on your team who deal with this data problem?' These are stage-appropriate asks that test real interest."

## Key Principles

- **Compliments cost nothing, so they are worth nothing** -- A prospect saying "I love it" has given up zero currency. It feels validating but contains no data. The natural human desire to hear praise makes compliments actively dangerous because they feel like progress. Train yourself to hear compliments as warning flags that the prospect may be trying to end the conversation politely.

- **It is not a real lead until you have given them a concrete chance to reject you** -- Zombie leads persist because founders avoid the scary moment of asking for commitment. Rejection is a valid and useful outcome -- it tells you where you stand. No-man's-land (friendly but non-committal) tells you nothing. Always push for a next step or a clear "no."

- **The more they give up, the more seriously you can take their words** -- This is the fundamental measurement principle. Evaluate every signal by asking: what did it cost them to say or do this? A pre-order costs money. An intro to their boss costs reputation. A scheduled follow-up with a specific agenda costs time. "That's cool" costs nothing. Rank signals by cost, not by enthusiasm.

- **Calibrate your ask to your product stage** -- Asking for money when you have no product is premature and counterproductive. Not asking for money when you have a live product is wasteful. Match commitment requests to what is appropriate for your current stage, and escalate as you progress.

- **Meetings either succeed or fail -- there is no "went well"** -- The phrase "the meeting went well" is itself a red flag. Every meeting either produces advancement (a concrete next step with commitment) or it does not. If you leave with only compliments and vague promises, the meeting failed. Acknowledging this honestly is the first step to fixing it.

## References

- For cross-referencing data quality in the same conversation, see the `conversation-data-quality-analyzer` skill
- For synthesizing learnings across multiple meetings into validated/invalidated assumptions, see the `conversation-learning-process` skill
- For the commitment currency taxonomy and meeting outcome scoring rubric, see [references/commitment-currencies.md](references/commitment-currencies.md)

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Mom Test by Rob Fitzpatrick.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/commitment-currencies.md
# Commitment Currencies — Deep Reference

Read this reference when you need the full detail on commitment currency classification, the complete meeting outcome scoring rubric, or earlyvangelist identification criteria.

## The Three Currencies

Commitment signals are measured by what the prospect gives up. There are three currencies, each representing a different type of cost to the prospect.

### Currency 1: Time

Time commitments indicate the prospect values the interaction enough to invest their calendar. The strength depends on how structured and costly the time investment is.

| Time Commitment | Signal Strength | Why |
|----------------|----------------|-----|
| "Let's grab coffee sometime" | VERY LOW | No specifics, no agenda, easy to cancel |
| Clear next meeting with known goals | MEDIUM | Structured, has purpose, harder to cancel without reason |
| Sitting down to give detailed wireframe feedback | MEDIUM-HIGH | Requires effort, not just attendance |
| Using a trial for a non-trivial period | HIGH | Ongoing investment, integration into workflow |
| Deploying to their team for a multi-week pilot | VERY HIGH | Compound: time + reputation (vouching to team) |

### Currency 2: Reputation

Reputation commitments mean the prospect stakes their social capital on you. These are harder to give than time and signal deeper conviction.

| Reputation Commitment | Signal Strength | Why |
|----------------------|----------------|-----|
| "I know some people..." (vague) | VERY LOW | Costs nothing to say, no specifics to follow up on |
| Introduction to peers or team members | MEDIUM | They risk looking foolish if you waste their colleagues' time |
| Introduction to a decision maker (boss, spouse, lawyer) | HIGH | They are putting their credibility on the line with someone who matters |
| Giving a public testimonial or case study | VERY HIGH | Permanent, public, their name is attached |
| Advocating for your product inside their organization | VERY HIGH | Career risk, compound with time |

### Currency 3: Money (Financial)

Financial commitments are the strongest signal because they require actual sacrifice. Even small amounts carry more data than enthusiastic words.

| Financial Commitment | Signal Strength | Why |
|---------------------|----------------|-----|
| "I would pay for that" | ZERO | Future tense = no cost = no data |
| Letter of intent (non-binding) | MEDIUM | Requires deliberation and written commitment, even if non-binding |
| Pre-order | HIGH | Real money, before the product is ready |
| Deposit | VERY HIGH | Money given up with risk of not getting value back |
| Purchasing the prototype | HIGHEST | They want it so badly they will pay for an unfinished version |

### Compound Commitments (Strongest Signals)

The strongest signals combine multiple currencies. Examples:
- **Paid trial with whole team** = Time + Money + Reputation
- **Public case study with pre-order** = Reputation + Money
- **Intro to boss plus wireframe feedback session** = Reputation + Time

When you see compound commitments, the prospect is almost certainly genuine.

## Complete Meeting Outcome Scoring Rubric

These 8 scenarios cover the most common meeting endings. Use them as pattern-matching templates when evaluating your own meetings.

### 1. "That's so cool. I love it!"
- **Verdict:** BAD MEETING
- **Currency:** None (pure compliment)
- **Danger level:** Low (obviously empty, unlikely to mislead experienced founders)
- **Fix:** Deflect the compliment. Ask "How are you handling this currently?" or "What would the next step look like for you?"

### 2. "Looks great. Let me know when it launches."
- **Verdict:** BAD MEETING
- **Currency:** None (compliment + stalling tactic)
- **Danger level:** HIGH (frequently mistaken for a pre-order)
- **Pattern name:** Compliment-stall
- **Fix:** Look for a commitment available today. "Before we wrap up -- would you be willing to [stage-appropriate ask]? We're looking for early design partners who can help shape the product."

### 3. "There are a couple people I can introduce you to when you're ready."
- **Verdict:** BAD (currently), POTENTIALLY GOOD
- **Currency:** Reputation (attempted, but too vague to count)
- **Danger level:** MEDIUM (feels like progress but the promise is unactionable)
- **Fix:** Make it concrete. "That would be incredibly helpful. Who did you have in mind? Could we set that up this week?" If they cannot name names or commit to a timeline, the offer was empty.

### 4. "What are the next steps?"
- **Verdict:** GOOD MEETING
- **Currency:** Time (they are volunteering to continue)
- **Danger level:** Low (genuine advancement signal)
- **Critical requirement:** YOU must know your next steps. If you say "Let me think about it and get back to you," you have ruined a perfectly good meeting. Always enter meetings knowing what commitment you will ask for.

### 5. "I would definitely buy that."
- **Verdict:** BAD MEETING (dangerous false positive)
- **Currency:** None (future-tense promise, costs nothing to say)
- **Danger level:** MAXIMUM (this exact phrase has directly caused startups to waste millions building products nobody actually purchased)
- **Fix:** Shift from future promise to current commitment. "That's great to hear. We're offering early access to our first 10 design partners. Would you like to put down a deposit to secure a spot?" Or: "Could you introduce me to whoever controls the budget for this?"

### 6. "When can we start the trial?"
- **Verdict:** MAYBE GOOD
- **Currency:** Depends on trial cost
- **Evaluation:** Think in terms of currency -- what are they actually giving up to try it? Setting up new software, training staff, integrating with their systems = HIGH signal (time + reputation). Clicking a sign-up button for a free consumer app = LOW signal (costs nothing).
- **Fix if trial is too cheap:** Increase the cost. Ask them to write a case study after 2 weeks of use. Ask them to commit their whole team. Charge a token fee refundable if they cancel within 30 days.

### 7. "Can I buy the prototype?"
- **Verdict:** GREAT MEETING
- **Currency:** Money (strongest possible)
- **Danger level:** None (this is the gold standard)
- **Action:** Take their money. This person is almost certainly an earlyvangelist. Keep them close. They will become your first case study, your first referral source, and your most honest feedback channel.

### 8. "When can you come back to talk to the rest of the team?"
- **Verdict:** GOOD MEETING
- **Currency:** Reputation (willing to bring you into their organization)
- **Danger level:** Low (enterprise advancement signal)
- **Note:** In enterprise sales, you will always need to talk to multiple people. If your contact will not introduce you to the team, it is almost certainly a dead end. This invitation is one of the earliest strong signals in B2B.

## Earlyvangelist Identification (Steve Blank's Customer Development)

Earlyvangelists are the rare prospects who will become your first customers. They commit before it makes rational sense to do so. In the enterprise world, they are the ones who will fight against their boss and lawyers when the tech is unproven. In the consumer world, they are the fans who will front you money as a pre-order when all you have is a duct-tape prototype.

### The 4 Criteria

All four should be present for a true earlyvangelist:

1. **Have the problem** -- They personally experience the pain your product addresses. Not theoretically, not "someone on my team does" -- they feel it themselves.

2. **Know they have the problem** -- They are consciously aware of the pain. Many people have problems they have normalized. An earlyvangelist has not normalized it -- they are actively frustrated.

3. **Have the budget to solve it** -- They can actually spend money. Enthusiasm without purchasing power is just entertainment. In B2B, this means they either hold budget or have direct access to someone who does.

4. **Have already cobbled together their own makeshift solution** -- This is the strongest single indicator. If they care enough to have built a workaround (spreadsheets, scripts, manual processes, competing products duct-taped together), the problem is real and urgent. If they have not tried to solve it, they will not buy yours either.

### Emotional Intensity as a Signal

There is a significant difference between:
- "Yeah, that's a problem" (mild acknowledgment)
- "THAT IS THE WORST PART OF MY LIFE AND I WILL PAY YOU RIGHT NOW TO FIX IT" (earlyvangelist intensity)

When you see deep emotion about the problem space, keep that person close regardless of where they fall on the 4-criteria checklist. They are rare and precious.

## Zombie Lead Detection Checklist

Review your pipeline for these symptoms:

- [ ] Prospect has taken 3+ meetings but made zero commitments of any currency
- [ ] Every meeting ends with compliments ("That was great!", "Really interesting!")
- [ ] No clear next steps have been established in any meeting
- [ ] Prospect uses future-tense language exclusively ("When you launch...", "Eventually...")
- [ ] You describe the relationship as "warm" but cannot point to a single thing they have given up
- [ ] The prospect has not introduced you to anyone else in their organization
- [ ] You feel reluctant to push for commitment because you do not want to "ruin the relationship"

If 3+ of these are true, this is a zombie lead. The fix is not more meetings -- it is giving them a concrete chance to commit or reject. Rejection is data. Zombie-land is not.

---
name: warrant-tester
description: Test the warrants in a research argument — the general principles that connect reasons to claims. Use this skill when a reader might accept a reason as true but still deny it is relevant to the claim, the argument contains a logical leap between a reason and a claim that no stated principle explains, the user needs to decide whether to state a warrant explicitly or leave it implicit, the argument needs to identify which type of warrant is being used (experience-based, authority-based, system/definitional, cultural, methodological) so it can be challenged or defended appropriately, the user is writing for an audience outside their field who will ask "but why does that reason matter to your claim?", a warrant appears to be too broad and needs qualification before it can survive challenge, competing warrants exist and the argument must show why its warrant should prevail, or the user suspects their argument is flawed but cannot identify where — surfacing the implicit warrant often reveals the problem. This skill depends on research-argument-builder (which assembles the full argument structure) and is typically used after reasons and claims have been identified but before final drafting.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/warrant-tester
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [11]
tags: [research-methodology, academic-writing, argumentation, warrants, logical-reasoning, critical-thinking]
depends-on: [research-argument-builder]
execution:
  tier: 1
  mode: full
  inputs:
    - type: text
      description: "A research argument — at minimum a claim and one or more reasons. The full five-element argument (claim, reasons, evidence, acknowledgment/response, warrant) is ideal but not required."
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user should have a claim and at least one reason before using this skill"
discovery:
  goal: "Verify that every reason in a research argument is genuinely connected to its claim by an explicit or defensible general principle, and identify where that principle needs to be stated, qualified, or defended"
  tasks:
    - "Surface the implicit warrant connecting each reason to its claim"
    - "Run the five-test validation battery against each warrant"
    - "Classify each warrant by type and identify the appropriate challenge strategy"
    - "Decide whether each warrant must be stated, qualified, or can remain implicit"
    - "Diagnose flawed arguments by revealing the warrant that makes the flaw visible"
  audience: "Researchers, graduate students, analysts, and professionals who have an assembled argument and need to verify that the logical connections between their reasons and claims will survive a careful reader's scrutiny"
  triggers:
    - "User's argument contains a claim and reasons but the connection between them is not obvious"
    - "Reviewer says the reasoning does not follow or the logic is unclear"
    - "User is writing for a general or cross-disciplinary audience who will not share field-specific assumptions"
    - "User needs to challenge someone else's warrant and needs a type-specific strategy"
    - "User suspects their argument is broken but cannot locate the problem"
    - "Argument sounds compelling but a skeptical reader says 'so what?' after hearing the reason"
---

# Warrant Tester

## When to Use

You have a claim and at least one reason. Your reason may be true, well-evidenced, and clearly stated — and a careful reader might still ask: *But how does that follow? Why does that reason matter to your claim?*

That question is asking for a warrant: a general principle that connects the reason to the claim. Warrants are the logical bridges in an argument. They are almost always present, but rarely stated. This skill surfaces them, tests them, and determines what to do with them.

**The core logic of a warrant:**

A warrant has two parts in the form *When X (general circumstance), then Y (general consequence)*. The specific reason must be a valid instance of the general circumstance. The specific claim must be a valid instance of the general consequence. If both fits hold, the argument is logically coherent. If either fit fails, the argument has a structural problem.

Example:

- Claim: "The Russian Federation faces a falling standard of living."
- Reason: "Its birthrate is only 13.2 per 1,000 and life expectancy for men is only about 63 years."
- Implicit warrant: "When a nation's labor force shrinks, its economic future is grim."

A reader who does not share that warrant will accept the reason but deny it is relevant to the claim. Supplying the warrant makes the logic explicit and gives the reader something specific to engage with.

**Preconditions to verify:**

- Does the user have a claim and at least one reason? If not, invoke `research-argument-builder` first to assemble the argument structure.
- Can the user state the claim and each reason in a single sentence? If not, ask them to do so before proceeding — warrant testing requires precise formulations.

**This skill does NOT cover:**

- Assembling the full argument (claim, reasons, evidence, acknowledgment): use `research-argument-builder`
- Generating detailed strategies for each reader objection or counterargument: use `counterargument-handler`

## Context and Input Gathering

### Required (ask if missing)

- **The main claim:** What the argument asserts. A single sentence.
  - Check prompt for: thesis statement, main conclusion, or any disputable assertion
  - If missing, ask: "What is the main claim your argument is trying to establish — the conclusion you want readers to accept?"

- **The reason(s):** Each statement offered in support of the claim.
  - Check prompt for: because-clauses, support statements, or any assertion offered as grounds for the claim
  - If missing, ask: "What are the reasons you are giving for your claim — the statements you use to justify it?"

### Useful (gather from environment if available)

- **Field or discipline:** Determines which warrants readers will take for granted and which need explicit defense
  - If not stated, ask: "What field or research community are you writing for?"

- **Audience expertise level:** Experts share field warrants; non-experts do not
  - Look for audience description in context; if unclear, ask: "Are your readers specialists in this field, or a more general academic audience?"

- **The full argument (if available):** Evidence, acknowledgments, and existing warrants
  - Check for files like `argument.md`, `outline.md`, `draft.md`, or any document containing the assembled argument

## Process

### Step 1 — Surface the Implicit Warrant for Each Reason

**WHY:** Most arguments rely on unstated warrants. Until the warrant is made explicit, it cannot be examined, tested, or defended. Making it explicit also helps the writer see whether the reason and claim are actually connected — which sometimes reveals that they are not, despite appearing to be.

For each reason→claim pair, surface the implicit warrant by completing this template:

> **When** [general circumstance that the reason is an instance of], **then** [general consequence that the claim is an instance of].

Work backward from the specific argument to find the general principle that would make it valid:

1. Identify what *class of thing* the specific reason belongs to (the general circumstance)
2. Identify what *class of outcome* the specific claim belongs to (the general consequence)
3. State the general principle connecting those two classes

**Example — Gun ownership argument:**

- Reason: "Guns were rarely mentioned in wills in the period 1750–1850."
- Claim: "Gun ownership was probably not widespread in early America."
- Implicit warrant: "When a household object considered valuable is not listed in a will, the household likely did not own one."

**Example — Video game argument (flawed):**

- Reason: "Children aged 12–16 today are significantly more violent than their counterparts from a generation ago."
- Claim: "Violent video games are exerting a destructive influence on today's youth."
- Implicit warrant: "When children are constantly exposed to images of sadistic violence, they are influenced for the worse."
- **Problem:** The reason (rising violence) is not a valid instance of the warrant's general circumstance (constant exposure to sadistic violent imagery). The reason does not cover exposure — it only covers an outcome. The warrant cannot connect this reason to this claim.

Surfacing the warrant reveals that the argument needs a new reason: one about *exposure* to violent imagery, not just *outcomes* in terms of behavior.

### Step 2 — Run the Five-Test Validation Battery

**WHY:** Readers challenge warrants in predictable ways. Running these five tests before readers do lets the writer address weaknesses proactively rather than defensively. A warrant that fails any of the five tests requires either qualification, supporting argument, or replacement.

Apply all five tests to each warrant:

---

**Test 1 — Is the warrant reasonable?**

A warrant is reasonable when readers can accept that its general consequence predictably follows from its general circumstance.

- Ask: "Would my readers, if they considered this general principle on its own, find it acceptable?"
- If no: The warrant needs to be argued for — treat it as a claim in its own sub-argument, supported by its own reasons and evidence.

Example: "When a household object considered valuable is not listed in a will, the household likely did not own one" — readers familiar with wills as legal documents will find this reasonable.

---

**Test 2 — Is the warrant sufficiently limited?**

Most warrants are reasonable only within certain limits. An unlimited version invites easy refutation.

- Ask: "Is this warrant stated so broadly that readers can easily find exceptions that would exclude my own case?"
- If too broad: Add qualifying language (*most*, *usually*, *in conditions where X*), then verify the exceptions do not exclude the specific reason and claim at hand.

Example — too broad: "Valuable objects were listed in wills."
Better: "**Most** household objects **considered valuable by their owners** were **usually** listed in wills."

Caution: Once you add qualifiers, you take on an obligation to show the exceptions do not apply to your case. Be precise about what *most* and *usually* mean, and whether guns counted as valuable.

---

**Test 3 — Is the warrant superior to any competing warrants?**

Other researchers may hold warrants that support the opposite conclusion from the same evidence. A warrant that faces no competition is rare; identifying competition and showing why your warrant prevails is part of building a defensible argument.

- Ask: "Is there a competing general principle that would lead a reasonable person to a different conclusion from the same reason?"
- If yes: Either reconcile the competing warrants by limiting both, or argue explicitly for why your warrant prevails.

Example — vaccination mandate debate:
- Warrant A: "When parents believe a medical procedure may harm their children, they have a right to refuse it."
- Warrant B: "When medical decisions concern matters of public health, the state has a right to regulate them."
- Both are reasonable. Reconciliation requires limiting both: Warrant A holds *so long as that does not jeopardize the health of others*; Warrant B holds *so long as the state encroaches as little as possible on parental medical decisions*.

---

**Test 4 — Is the warrant appropriate to this field?**

A warrant that seems reasonable in general may be inadmissible in a specific research community. Field-specific reasoning standards can override common-sense warrants.

- Ask: "Would researchers and reviewers in my field accept this principle as a valid basis for inference, or does my field require a different kind of reasoning?"
- If no: Either find the field-appropriate warrant, or explicitly argue that the field's standard warrant should be updated or supplemented.

Example: Common sense says "When a person is wronged, the law should correct it." In legal reasoning, this warrant has no standing — legal warrants based on precedent and procedure may override it entirely. Law students learn this at their cost.

---

**Test 5 — Does the warrant cover this specific reason and claim?**

Even a reasonable, limited, superior, and field-appropriate warrant fails if the specific reason is not a valid instance of the warrant's general circumstance, or the specific claim is not a valid instance of the warrant's general consequence.

- Ask (circumstance side): "Is my specific reason genuinely an instance of what the warrant's general circumstance describes?"
- Ask (consequence side): "Is my specific claim genuinely an instance of what the warrant's general consequence predicts?"

Example:
- Warrant: "When you are not safe, you should protect yourself."
- Reason: "You live alone."
- Claim: "You should buy a gun."
- Circumstance check: "Living alone" is not obviously an instance of "being unsafe" — this is the gap.
- Consequence check: Even if unsafe, "buying a gun" may not be the only or best instance of "protecting yourself."

If the reason does not fit the circumstance: Revise the reason (or find new evidence) so it becomes a valid instance of the warrant's general circumstance.
If the claim does not fit the consequence: Revise the claim, or find a different warrant whose consequence does cover the claim.

### Step 3 — Classify Each Warrant by Type

**WHY:** The type of warrant determines how it can be challenged and how it should be defended. Applying a challenge strategy designed for one type to a different type wastes effort and often backfires. Knowing the type also helps identify whether the challenge is even winnable — some warrant types are unchallengeable in principle.

| Warrant type | Basis | Examples | Challenge strategy |
|---|---|---|---|
| **Experience-based** | Individual or reported observations | "When people habitually lie, we don't trust them." "When insecticides leach into ecosystems, bird populations fall." | (1) Challenge reliability of the experience or report — difficult; (2) Find counterexamples that cannot be dismissed as special cases |
| **Authority-based** | Expertise, position, or credibility of a source | "When authority X says Y, Y must be so." | Friendly: show the authority lacked full information or exceeds their expertise in this case. Aggressive: argue the source is not a genuine authority at all. |
| **System/definitional** | Formal definitions, principles, or theories (mathematics, biology, law) | "When we add two odd numbers, we get an even one." "When we drive without a license, we commit a misdemeanor." | Facts are largely irrelevant. Either challenge the system itself (very difficult) or show the case does not fall under the warrant's definition. |
| **Cultural** | Common experience of an entire community | "Out of sight, out of mind." "An insult justifies retaliation." | Possible but high resistance — readers experience this as a challenge to their heritage. These warrants change slowly. |
| **Methodological** | Abstract patterns of reasoning | Generalization, analogy, sign | Challenge the application or identify limiting conditions: "Yes, we can analogize X to Y, but not in this respect because..." |

**Methodological warrants in detail:**

- *Generalization:* When every known case of X has quality Y, then all Xs probably have quality Y. Challenge: find a known case that does not have quality Y, or show the sample was not representative.
- *Analogy:* When X is like Y in most respects, then X will be like Y in other respects. Challenge: identify a relevant difference that breaks the analogy.
- *Sign:* When Y regularly occurs before, during, or after X, then Y is a sign of X. Challenge: show that Y and X are correlated but one does not reliably indicate the other.

**Note on warrants based on articles of faith:** Some warrants are beyond rational challenge — they are backed by the certainty of those who espouse them, not by evidence or argument. Do not build a research argument that depends on others accepting such a warrant; and when you encounter them while gathering data, do not treat them as warrants to test. They operate outside the logic of academic argument.

### Step 4 — Decide Whether to State Each Warrant Explicitly

**WHY:** Stating a warrant your readers already accept wastes their time and can appear condescending. Failing to state a warrant your readers need to understand your reasoning — or are likely to challenge — leaves a logical gap that careful readers will exploit. The decision is an audience judgment, not a structural one.

State a warrant explicitly when any of the following apply:

1. **Readers are outside your field.** If you are writing as a specialist for non-specialists, explain how experts in your field draw conclusions — especially if that reasoning is unusual by general standards.

2. **The principle is new or controversial in your field.** If you rely on an unconventional reasoning principle, state and defend it proactively. Cite others in your field who use it. If you cannot, build the argument for it yourself.

3. **Readers will resist the claim because they do not want it to be true.** Lead with a warrant readers can accept *before* laying out the reason and claim they may resist. This does not guarantee acceptance, but it obliges readers to engage rationally rather than emotionally.

Do not state a warrant when:
- Readers are specialists in your field who already take it for granted (stating it appears condescending)
- The warrant is so general and uncontested that every adult accepts it without prompting
- Stating it would slow the argument without adding logical protection

**Ethos implication:** Stating a warrant when your readers need it signals intellectual care and respect. Keeping a warrant implicit when you know readers will ask for it signals overconfidence or evasion. Both choices communicate something about your credibility as a researcher.

### Step 5 — Diagnose and Repair Flawed Arguments Using Warrants

**WHY:** The most common cause of an argument that "feels wrong" but is hard to locate is a reason that does not genuinely connect to the claim. Surfacing the implicit warrant makes the problem visible: the reason is not a valid instance of the warrant's general circumstance, or the claim is not a valid instance of the warrant's general consequence.

**Diagnosis procedure:**

1. State the claim and the reason in one sentence each.
2. Surface the implicit warrant (Step 1).
3. Check whether the reason fits the warrant's circumstance (Step 2, Test 5, circumstance side).
4. Check whether the claim fits the warrant's consequence (Step 2, Test 5, consequence side).
5. Identify which fit fails.

**Repair options:**

| Failure | Repair |
|---|---|
| Reason does not fit warrant's circumstance | Revise the reason to be a valid instance of the circumstance; gather new evidence to support the revised reason |
| Claim does not fit warrant's consequence | Narrow or restate the claim so it is a valid instance of the consequence; or find a different warrant whose consequence does cover the claim |
| Warrant does not survive the five-test battery | Qualify the warrant, argue for it as a sub-claim, replace it, or abandon the reason it was meant to support |
| No warrant can connect this reason to this claim | Discard the reason; find a reason that a defensible warrant can cover |

**Worked example — Video game argument:**

Flawed:
- Reason: "Children 12–16 are significantly more violent than a generation ago."
- Claim: "Violent video games are exerting a destructive influence on today's youth."
- Implicit warrant: "When children are constantly exposed to images of sadistic violence, they are influenced for the worse."
- Failure: The reason (observed rising violence) is not an instance of the warrant's circumstance (exposure to violent imagery). Rising violence is an outcome, not an exposure.

Repaired:
- New reason: "Over the past decade, video games have become a major source of children's exposure to violent imagery."
- Same warrant applies: the new reason is now a valid instance of the warrant's circumstance (exposure to sadistic violent imagery).
- New evidence needed: data showing the proportion of children's screen time spent on violent video games and the nature of that violence.

## Examples

### Example 1 — Historical argument

**Argument:**
- Claim: "Gun ownership was probably not widespread in early America before 1850."
- Reason: "Guns were rarely mentioned in wills filed in seven states between 1750 and 1850."

**Step 1 — Surfaced warrant:** "When valuable household objects were not listed in a will, the household did not own them."

**Five-test battery:**
1. Reasonable? Plausible — wills were legal inventories of valuables. But readers who believe widespread gun ownership was common will resist. Needs supporting argument: Watson (1989) confirmed this practice in Cumberland County wills. [sub-argument needed]
2. Sufficiently limited? The original version allows no exceptions. Qualified: "Most household objects *considered valuable by their owners* were *usually* listed in wills." Now must address: Were guns always considered valuable? What is "usually"?
3. Superior to competing warrants? A competing warrant: "When war was common, most households owned weapons as a matter of survival." This competes. Must argue why the will-listing warrant prevails for this time and place.
4. Field-appropriate? For historical demography — yes. Document-based inference from legal records is standard in this field.
5. Covers this reason and claim? Check: "Guns rarely mentioned in wills" is an instance of "valuable object not listed" only if guns were valuable. Check: "Gun ownership not widespread" is an instance of "household did not own them" at the aggregate level. Both fits hold, but the "valuable" condition is the vulnerability.

**Decision on stating the warrant:** State explicitly — the argument's expected readers (general readers who believe widespread gun ownership was historically common) will not share this principle and will challenge relevance.

---

### Example 2 — Interdisciplinary argument with competing warrants

**Argument:**
- Claim: "The state can require vaccination of children against measles."
- Reason: "When most children are vaccinated, everyone is safer."

**Step 1 — Surfaced warrant:** "When medical decisions concern matters of public health, the state has a right to regulate them."

**Competing warrant:** "When parents believe a medical procedure may harm their children, they have a right to refuse it."

**Test 3 — Competing warrant analysis:** Both are reasonable and backed by defensible principles. Neither can simply be dismissed. Reconciliation: limit both.

- Warrant A limited: "Parents may refuse a medical procedure *so long as that refusal does not jeopardize the health of others.*"
- Warrant B limited: "The state may regulate medical decisions concerning public health *so long as it encroaches as little as possible on parental medical decisions.*"

**Decision on stating:** State the warrant and explicitly acknowledge the competing one — this argument is inherently about competing principles. Stating the reconciliation shows intellectual honesty and invites rational engagement.

---

### Example 3 — Self-diagnosis of a broken argument

**Input from user:** "My argument feels wrong but I cannot find the problem. I'm arguing that social media should be regulated, because teenagers are anxious today."

**Step 1 — Surfaced warrant:** "When teenagers are anxious, social media is harming them."

**Test 5 — Coverage check:**
- Circumstance side: "Teenagers are anxious today" — is this an instance of "social media is harming them"? No. Anxiety is an outcome, not an instance of harm *by social media*. The reason does not fit the warrant's circumstance.
- This is the flaw: rising anxiety does not logically implicate social media without a reason that specifically connects the two.

**Repair:** Revise the reason to: "Teenagers who spend more than three hours daily on social media report significantly higher anxiety rates than those who spend less than one hour." This reason is now an instance of the warrant's circumstance (harm from social media use). Gather evidence: longitudinal survey data showing the correlation with exposure to social media specifically, not screens in general.

## Output

For each reason→claim pair, produce:

1. **Warrant statement:** The general principle in *When X, then Y* form
2. **Five-test results:** Pass/fail for each test, with the specific issue named if it fails
3. **Warrant type:** Classified from the taxonomy in Step 3
4. **Challenge strategy:** The type-specific approach a skeptical reader would use
5. **State or leave implicit?** Decision based on Step 4, with the reasoning stated
6. **Repair instructions (if needed):** Specific changes to the reason, claim, or warrant, with what new evidence is required

If the argument has multiple reasons: process each separately, then note whether the warranted reasons together constitute a sufficient case for the claim.

## Anti-Pattern Quick Reference

| Anti-pattern | Signal | Fix |
|---|---|---|
| Reason stated but not connected | Reader says "so what?" or "why does that matter?" | Surface and state the implicit warrant |
| Warrant too broad | A single obvious exception invalidates the entire argument | Add qualifiers (*most*, *usually*, *in cases where X*) and address whether exceptions exclude your case |
| Reason does not fit warrant's circumstance | Reason describes an outcome but warrant requires an exposure or condition | Revise the reason to be a valid instance of the warrant's circumstance |
| Competing warrant ignored | Opponent uses same evidence to reach opposite conclusion | Limit both warrants explicitly; argue for why yours prevails in this context |
| Field-inappropriate warrant | Reviewer rejects the reasoning principle, not the data | Identify the field's accepted warrant and use it; or explicitly argue for updating the field's reasoning standard |
| Faith-based warrant in a research argument | Warrant cannot be challenged because it rests on certainty, not evidence | Do not build a research argument on such a warrant; find an evidence-based principle or revise the argument |
| Unstated warrant for a skeptical audience | Argument sounds valid to the writer but not to the reader | Identify the audience's knowledge level and state the warrant that bridges the gap |

## References

- `references/warrant-types-and-challenge-strategies.md` — Full taxonomy of warrant types with worked examples and type-specific challenge strategies
- `references/five-test-battery-guide.md` — Detailed guidance on each of the five validation tests with cross-field examples and common failure modes

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-argument-builder`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/five-test-battery-guide.md
# Five-Test Warrant Validation Battery

Before stating a warrant or relying on one implicitly, run all five tests. A warrant that fails any test requires intervention — either qualification, a supporting sub-argument, or replacement.

---

## Test 1: Is the Warrant Reasonable?

**What this tests:** Whether readers can accept, in principle, that the warrant's general consequence predictably follows from its general circumstance.

**How to apply:** State the warrant on its own, without the specific reason and claim. Ask: "Would my target readers, reading this general principle, find it acceptable without further argument?"

**Pass:** Readers would accept the warrant as a reasonable general principle.

**Fail:** Readers would question whether the consequence reliably follows from the circumstance.

**What to do when it fails:** Treat the warrant as a claim in its own sub-argument. Give it its own reasons and evidence. This is a supporting argument for the warrant, not for the main claim.

**Example — Fails:**
- Warrant: "When a valuable object is not listed in a will, the household did not own it."
- Problem: Readers may know that wills were sometimes incomplete, hastily prepared, or that certain objects were transferred informally.
- Fix: Build a sub-argument with historical evidence that wills in this period and region reliably recorded valuable household possessions.

**Example — Passes:**
- Warrant: "When an overwhelming majority of competent experts in a field arrive at the same conclusion, we can probably trust it."
- Most readers accept this as a reasonable epistemic principle without further argument.

---

## Test 2: Is the Warrant Sufficiently Limited?

**What this tests:** Whether the warrant is stated so broadly that it admits exceptions that would undermine the argument.

**How to apply:** Look for absolute phrasing — "always," "every," "all," or the absence of any qualifier. Ask: "Can I easily imagine a legitimate exception that would exclude my specific reason or claim?"

**Pass:** The warrant is qualified appropriately so that exceptions are acknowledged and the specific case still falls within the warrant's scope.

**Fail:** The warrant is stated as a universal that admits no exceptions, making it easily refuted by a single counterexample.

**What to do when it fails:**
1. Add qualifying language: *most*, *usually*, *in conditions where X*, *except when Y*.
2. After qualifying, verify that the exceptions identified by the qualifiers do not exclude your specific reason or claim.
3. Answer new questions raised by qualifiers: If "most" or "usually" — what is the frequency? Does your case fall within the typical range?

**Example — Before and after:**
- Too broad: "Valuable objects were listed in wills."
- Better: "Most household objects **considered valuable by their owners** were **usually** listed in wills in the eighteenth and early nineteenth centuries."
- Follow-up obligation: Show that guns were considered valuable in this period, and that the "usually" does not cover enough exceptions to exclude the gun case.

**Tension to manage:** The more you limit a warrant, the more specific you must be about the conditions — which creates new questions to answer. The goal is the minimum qualification needed to survive challenge, not the maximum qualification that avoids all possible objections.

---

## Test 3: Is the Warrant Superior to Any Competing Warrants?

**What this tests:** Whether a reasonable person, using a different but also defensible general principle, would reach the opposite conclusion from the same reason.

**How to apply:** Ask: "Is there another general principle my readers might hold that would lead them, accepting the same reason, to a different or opposite conclusion?"

**Pass:** No reasonable competing warrant exists, or the competing warrant is clearly weaker or more limited.

**Fail:** A competing warrant exists that is also reasonable and that leads to an incompatible conclusion.

**What to do when it fails:**

Option A — Reconcile by limiting both:
- Add limiting conditions to your warrant that carve out the domain where yours applies.
- Add limiting conditions to the competing warrant that acknowledge where it applies but show it does not govern your case.

Option B — Argue for superiority:
- Show that your warrant is more precisely applicable to this specific context.
- Show that the competing warrant was developed for a different domain or rests on weaker evidence.
- Show that accepting the competing warrant leads to results that the audience would find unacceptable.

**Example — Vaccination mandate:**
- Your warrant: "When medical decisions concern public health, the state has a right to regulate them."
- Competing: "When parents believe a medical procedure may harm their children, they have a right to refuse it."
- Reconciled: Each warrant holds *within the limits* that do not override the other's core domain. The argument then turns on where the line falls in the specific case.

**Note:** When competing warrants exist and cannot be definitively resolved, the dispute is not about evidence but about values. Acknowledge this explicitly — trying to win a values dispute with data looks intellectually dishonest.

---

## Test 4: Is the Warrant Appropriate to This Field?

**What this tests:** Whether the research community you are writing for accepts this general reasoning principle as a valid basis for inference.

**How to apply:** Ask: "Would peer reviewers or other specialists in my field accept this principle as a legitimate way to reason in this domain, or does my field use a different or stricter reasoning standard?"

**Pass:** The warrant uses reasoning principles that are standard in the field.

**Fail:** The warrant uses reasoning principles from common sense or another field that the target research community does not recognize as valid.

**What to do when it fails:**

Option A — Replace with a field-appropriate warrant:
- Identify what general principle your field uses to make the same kind of inference.
- Reframe the argument around that principle.

Option B — Argue for updating the field's standard:
- Acknowledge that your reasoning principle is unconventional in the field.
- Cite researchers in the field who have used it and defend it.
- If none exist, build your own argument for why the field should accept this principle.

**Example — Legal reasoning:**
- Common-sense warrant: "When a person is wronged, the law should correct it."
- Field-appropriate warrant: "When one ignores legal obligations, even inadvertently, one must suffer the consequences."
- Lesson: Legal argument runs on precedent, procedural rules, and statutory definition — not on general moral intuitions. Writers crossing from ethics into law must translate their warrants.

**Cross-disciplinary writing:** When writing for audiences across fields, state and briefly defend the field-specific warrant so readers from other traditions can follow the reasoning.

---

## Test 5: Does the Warrant Cover This Specific Reason and Claim?

**What this tests:** Whether the specific reason is a valid instance of the warrant's general circumstance, and whether the specific claim is a valid instance of the warrant's general consequence.

**How to apply:**

Circumstance check: "Is my specific reason genuinely an instance of what the warrant's general circumstance describes? Or does the reason describe something different — an outcome, a correlation, a related but distinct condition?"

Consequence check: "Is my specific claim genuinely an instance of what the warrant's general consequence predicts? Or does the claim go further, or in a different direction, than the consequence allows?"

**Pass:** Both fits hold. The reason is a clear instance of the circumstance; the claim is a clear instance of the consequence.

**Fail:** One or both fits do not hold. The reason or claim is outside the scope of what the warrant covers.

**What to do when the circumstance fit fails:**
- Revise the reason so it becomes a valid instance of the warrant's general circumstance.
- This often requires new evidence to support the revised reason.

**What to do when the consequence fit fails:**
- Narrow or restate the claim so it falls within the warrant's general consequence.
- Or find a different warrant whose consequence does cover the specific claim.

**Example — Gun ownership (circumstance check):**
- Warrant: "When valuable objects are not listed in a will, households did not own them."
- Reason: "Guns were rarely mentioned in wills."
- Circumstance fit: Holds — *if* guns were valuable. The condition "considered valuable" must be established separately.
- Consequence fit: "Gun ownership not widespread" is an instance of "household did not own them" at an aggregate level. Holds.
- Vulnerability: The circumstance fit is conditional on the "valuable" criterion being met. If guns were not considered valuable, the reason falls outside the warrant's circumstance.

**Example — Video games (circumstance failure):**
- Warrant: "When children are constantly exposed to images of sadistic violence, they are influenced for the worse."
- Reason: "Children aged 12–16 today are significantly more violent than a generation ago."
- Circumstance fit: Fails. Rising violence is an *outcome*, not an *exposure*. The warrant's circumstance requires evidence about exposure to violent imagery, not about resulting behavior.
- Fix: Revise reason to: "Over the past decade, video games have become a major source of children's exposure to violent imagery." This is now a valid instance of the warrant's circumstance.

---

## Summary: Five-Test Checklist

| Test | Core question | Action if failed |
|---|---|---|
| 1. Reasonable? | Would readers accept this general principle without further argument? | Build a sub-argument supporting the warrant as a claim |
| 2. Sufficiently limited? | Is the warrant qualified enough to survive counterexamples? | Add qualifiers; verify exceptions don't exclude your case |
| 3. Superior? | Does a competing warrant lead to the opposite conclusion? | Limit both warrants; argue for why yours governs this case |
| 4. Field-appropriate? | Does your research community accept this reasoning principle? | Replace with field-standard warrant or argue for updating the standard |
| 5. Covers your case? | Does the reason fit the warrant's circumstance; does the claim fit its consequence? | Revise reason or claim so fits hold; gather supporting evidence for revised reason |

FILE:references/warrant-types-and-challenge-strategies.md
# Warrant Types and Challenge Strategies

A warrant is a general principle connecting a reason to a claim, in the form: *When X (general circumstance), then Y (general consequence)*. Different types of warrants rest on different kinds of supporting arguments — and must therefore be challenged in different ways.

## The Six Warrant Types

### 1. Experience-Based Warrants

**What backs them:** Individual experience or reports by others of repeated observations.

**Form:** When [observable condition], then [observable consequence].

**Examples:**
- "When people habitually lie, we don't trust them."
- "When insecticides leach into the ecosystem, eggshells of wild birds become so weak that fewer chicks hatch and the bird population falls."

**Challenge strategies:**
1. **Challenge the reliability of the experience or report.** Show that the observations were limited in scope, poorly recorded, or subject to bias. This is rarely easy — firsthand experience has strong rhetorical weight.
2. **Find counterexamples that cannot be dismissed as special cases.** A single counterexample can be dismissed as exceptional; several that span different conditions undermine the generalization.

**Defender's response:** Show that the experience is well-documented, replicated across contexts, and that the counterexamples offered are genuine special cases.

---

### 2. Authority-Based Warrants

**What backs them:** The expertise, position, or credibility of a source or institution.

**Form:** When authority X says Y, then Y must be so.

**Examples:**
- "The Surgeon General has determined that cigarette smoking is dangerous to your health."
- "The court ruled that the contract was void."

**Challenge strategies:**
1. **Friendly challenge:** Show that on the specific matter in question, the authority lacks full information, is working outside their expertise, or reached their conclusion under conditions that no longer apply.
2. **Aggressive challenge:** Argue that the source is not a genuine authority at all — their credentials are overstated, their methodology is flawed, or their position creates a conflict of interest.

**Defender's response:** Specify the scope of the authority's expertise and show the matter falls squarely within it; demonstrate independence and methodological rigor.

---

### 3. System/Definitional Warrants

**What backs them:** Formal systems of definitions, principles, or theories — mathematics, formal logic, law, linguistics, or domain-specific theoretical frameworks.

**Form:** When [condition defined by the system], then [consequence the system specifies].

**Examples:**
- *From mathematics:* "When we add two odd numbers, we get an even one."
- *From biology:* "When an organism reproduces sexually, its individual offspring differ."
- *From law:* "When we drive without a license, we commit a misdemeanor."

**Challenge strategies:**
1. **Challenge the system itself.** Argue that the system's definitions or principles are wrong, incomplete, or should not govern this case. This is difficult and requires building a counter-system or appealing to a higher-level principle.
2. **Show the case does not fall under the warrant.** Argue that the specific reason is not a valid instance of the system's defined condition — for instance, that the object in question does not meet the legal definition of the term being applied.

**Note:** In system-based argumentation, empirical facts are largely irrelevant to challenging the warrant directly. The dispute is about definitions and system membership, not observations.

---

### 4. Cultural Warrants

**What backs them:** The common experience of an entire community — shared history, values, and intuitions that function as "common sense" within that culture.

**Form:** When [culturally recognized situation], then [culturally expected consequence].

**Examples:**
- "Out of sight, out of mind."
- "An insult justifies retaliation."
- "Handling toads causes warts."

**Challenge strategies:**
- **Direct challenge:** Argue that the cultural principle is false, outdated, or inapplicable to the modern context. This is possible but provokes strong resistance — readers experience the challenge as an attack on their shared heritage, not just on an argument.
- **Scope limitation:** Accept the cultural principle in general but show that this specific case falls outside its scope.

**Practical advice:** Cultural warrants change slowly and are rarely worth challenging head-on in a research argument. It is more effective to show the specific case is an exception than to argue the general principle is wrong.

---

### 5. Methodological Warrants (Meta-Warrants)

**What backs them:** Abstract patterns of reasoning with no content of their own — they apply to any subject matter.

**Three primary types:**

#### Generalization
**Form:** When every known case of X has quality Y, then all Xs probably have quality Y. (*Seen one, seen them all.*)

**Challenge:** Find a known case of X that does not have quality Y. Alternatively, show the sample of known cases was not representative — it was limited to one time period, one geography, or one sub-population.

#### Analogy
**Form:** When X is like Y in most respects, then X will be like Y in other respects. (*Like father, like son.*)

**Challenge:** Identify a relevant difference between X and Y that breaks the analogy in precisely the respect the argument requires. Philosophers distinguish strong from weak analogies; for practical purposes, the challenge is to show the analogy does not hold in the specific domain the argument needs.

#### Sign
**Form:** When Y regularly occurs before, during, or after X, then Y is a sign of X. (*Cold hands, warm heart.*)

**Challenge:** Show that Y and X are correlated but that the correlation is explained by a third factor, is coincidental, or does not hold consistently enough to function as a reliable sign.

---

### 6. Warrants Based on Articles of Faith

**What backs them:** The certainty of those who espouse them, not evidence or argument.

**Examples:**
- "When a claim is experienced as revealed truth, it must be true."
- "When a claim is based on divine teaching, it must be true."

**Challenge strategies:** None viable in the context of academic research argument. These warrants are backed not by evidence but by certainty that is by definition impervious to evidence or argument. Challenging them is pointless.

**Practical guidance:** If you encounter such warrants in your research data, do not try to refute them — treat them as a subject for inquiry (Why do people hold this belief?) rather than an argumentative premise to defeat. Do not build a research argument that requires others to accept a warrant of this type.

---

## Quick Reference: Type → Challenge Strategy

| Warrant type | Primary challenge approach |
|---|---|
| Experience-based | Reliability of observation + counterexamples |
| Authority-based | Scope/expertise limits (friendly) or credential challenge (aggressive) |
| System/definitional | Challenge the system or show case does not fall under the definition |
| Cultural | Show the specific case is an exception (more viable than challenging the principle) |
| Methodological — Generalization | Counterexample or unrepresentative sample |
| Methodological — Analogy | Identify the relevant difference that breaks the analogy |
| Methodological — Sign | Show correlation without reliable indication |
| Articles of faith | Do not challenge; treat as data about belief, not as an argument to refute |

---
name: source-incorporator
description: Incorporate quoted, paraphrased, and summarized sources into research writing by applying a 3-branch selection decision tree, 3 integration methods for quotations, and a 5-mechanism inadvertent plagiarism prevention checklist. Use this skill when drafting or revising a paper that uses sources and you need to decide whether to quote, paraphrase, or summarize a passage; when you need to weave quotations into your prose grammatically and meaningfully; when you need to make explicit to readers why evidence is relevant; when you must choose a citation style; or when you want to audit a draft for the five most common forms of inadvertent plagiarism before submitting.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/source-incorporator
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [14]
tags: [research-methodology, academic-writing, citation, source-incorporation, plagiarism-prevention]
depends-on: []
execution:
  tier: 1
  mode: full
  inputs:
    - type: document
      description: "Draft text containing source material to incorporate, or a set of source notes (quotations, paraphrases, summaries) to be woven into a paper."
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Works from a draft document or a list of source passages. Output: revised draft with properly integrated sources and a plagiarism audit report."
discovery:
  goal: "Produce writing that integrates sources in the form best suited to each passage, integrates quotations grammatically, makes evidence relevant through framing sentences, and is clean of all five forms of inadvertent plagiarism."
  tasks:
    - "Apply the quote/paraphrase/summarize decision tree to each source passage"
    - "Select one of three integration methods appropriate to each quotation"
    - "Add a framing sentence before complex evidence to establish its relevance"
    - "Run the five-mechanism plagiarism audit on every incorporated passage"
    - "Apply or verify the correct citation style for every in-text reference"
  audience: "researchers, students, academics, professionals writing any evidence-based document"
  when_to_use: "When drafting or revising research writing that draws on external sources and must show credibility, accuracy, and respect for intellectual ownership"
  environment: "Draft document or source notes. Research question and target citation style should be known."
  quality: placeholder
---

# Source Incorporator

## When to Use

You are writing a research paper, report, or any evidence-based document and must bring in material from external sources. This skill applies when:

- You have collected source passages and must choose whether to quote, paraphrase, or summarize each one
- You have quoted a source but the quotation is not integrating smoothly into your prose
- You have placed evidence in your paper but readers may not see why it supports your claim
- You want to audit a near-final draft for inadvertent plagiarism before submitting
- You are unsure which citation style to use or how to format in-text references

**The core pattern:** source incorporation is a three-stage decision — (1) choose the right form (quote, paraphrase, or summarize), (2) integrate that form grammatically and meaningfully into your sentence, (3) make the relevance of the evidence explicit with a framing sentence. Then audit every incorporated passage for the five inadvertent plagiarism mechanisms.

Before starting, confirm you have:
- A draft or source notes you want to incorporate
- A working research question or thesis
- Knowledge of your required citation style, or permission to choose one

---

## Context and Input Gathering

### Required Context

- **Draft or source passages:** The text to be incorporated — direct quotations, paraphrases already drafted, or summaries.
- **Target citation style:** Chicago author-title, MLA, Chicago author-date, or APA. If unknown, ask.

### Observable Context

If a draft document is provided, scan for:
- Passages surrounded by quotation marks — check for integration method and framing
- Passages attributed to a source name without quotation marks — check if paraphrase or summary
- Parenthetical references or footnotes — check format against the declared citation style
- Long quoted passages (5+ lines) — check if set as block quotations
- Claims supported by evidence with no explanatory sentence — candidates for framing

### Default Assumptions

- If citation style is unspecified → ask before proceeding; applying the wrong style wastes revision effort
- If source passage length is borderline (4 vs. 5 lines) → use the block format when in doubt; it signals care
- If a passage combines the writer's words with source words invisibly → treat as plagiarism risk regardless of intent
- If source type is legal or statutory text → quotation marks may be omitted per field convention; confirm before flagging

### Sufficiency Check

You have enough to proceed when:
1. You know which passages are the writer's own words and which come from sources
2. You know the required citation style
3. You have at least one source passage to work with

If you cannot distinguish source text from writer text, stop and ask the writer to mark source passages before continuing.

---

## Execution

### Step 1 — Apply the Quote / Paraphrase / Summarize Decision Tree

For each source passage, choose the right form using these three branches. Use only one form per passage; mixing forms on the same passage is a common error.

**Branch 1 — Summarize** when:
- The details in the passage are not essential to your argument
- The source is a background reference, not a central exhibit
- The source is not important enough to warrant close analysis

*Why:* Readers only need what serves your argument. Unnecessary detail weakens focus and suggests you are padding rather than arguing.

**Branch 2 — Paraphrase** when:
- Your argument depends on the details in the passage, but not on the specific words
- You can restate the idea more clearly or concisely than the original
- The original language is technical and your audience needs a plain-language equivalent

*Why:* Paraphrase shows you have genuinely understood the source; quotation can sometimes hide that you have not. Reserve the source's exact words for when they do work a paraphrase cannot.

**Branch 3 — Quote** when one or more of these conditions holds:
- The words themselves are evidence (e.g., a historical document, a legal statute, a primary source statement)
- The words come from an authority whose exact phrasing endorses your claim
- The phrasing is strikingly original or conceptually precise in a way that would be diluted by paraphrase
- You are quoting to disagree and fairness requires stating the view exactly

*Why:* Direct quotation carries the highest evidentiary weight but also the highest reader cost. Every quoted passage asks readers to pause and read carefully. Use that cost purposefully.

**Example application:**

| Passage type | Decision | Reasoning |
|---|---|---|
| Broad historical background on a topic you mention once | Summarize | Details not central to your argument |
| A researcher's four-step method you will explain and apply | Paraphrase | Details matter, but your own clearer phrasing serves readers better |
| An economist's coinage of "creative destruction" as the precise term your argument turns on | Quote | Strikingly original phrase; paraphrase would lose precision |
| A critic's statement of a position you are refuting | Quote | Fairness requires exact language |

---

### Step 2 — Choose an Integration Method for Each Quotation

Once you decide to quote, you must integrate the quotation into your prose. There are two format decisions and three integration methods.

**Format decision — run-in or block:**
- **Run-in quotation:** 4 or fewer quoted lines — integrate directly into your paragraph, surrounded by quotation marks.
- **Block quotation:** 5 or more quoted lines — indent as a separate block, no quotation marks around the block.

*Why the threshold matters:* Long run-in quotations fragment your prose rhythm and are harder to read. Block format signals to readers that a substantial passage follows and they should read it as a unit. Using block format for short passages wastes space and looks evasive.

**Integration method — choose one of three:**

**Method A — Drop in:** Introduce the quotation with a brief identifying phrase (*Author says*, *According to Author*, *As Author puts it*). Use when the quotation is self-explanatory in context.

> Diamond says, "The histories of the Fertile Crescent and China . . . hold a salutary lesson for the modern world: circumstances change, and past primacy is no guarantee of future primacy" (417).

**Method B — Introduce:** Precede the quotation with a sentence that interprets or characterizes what the quotation will show. Use when the quotation needs framing to be understood at all.

> Diamond suggests what we can learn from the past: "The histories of the Fertile Crescent and China . . . hold a salutary lesson for the modern world . . ." (417).

**Method C — Weave:** Grammatically merge the quotation into your own sentence structure. Use when only part of the source sentence is needed or when seamless flow is important.

> Diamond suggests that the chief "lesson for the modern world" in the history of the Fertile Crescent and China is that "circumstances change, and past primacy is no guarantee of future primacy" (417).

*Why three methods:* Drop-in is fastest but can leave readers wondering why the quotation matters. Introduce provides the most interpretive control. Weave demonstrates mastery but risks distorting grammar if done carelessly. Match the method to the complexity of the evidence and the density of argument around it.

**Modifying quotations (permitted):**
- Use ellipsis `. . .` to mark deleted words within a quotation
- Use square brackets `[word]` to change a word so the quotation fits your grammar or to add a clarifying word
- Never change the meaning of the quotation while modifying it — this is a form of dishonesty

---

### Step 3 — Frame Complex Evidence Before Presenting It

Evidence never speaks for itself, especially long quotations or data tables. Readers who encounter unframed evidence must guess why it is relevant.

**The framing pattern:**
1. State your claim
2. Add a framing sentence that explains what the evidence will show and why it is relevant (the reason)
3. Report the evidence

*Without framing (ineffective):*
> When Hamlet comes upon his stepfather, Claudius, at prayer, he demonstrates cool rationality: [long quotation from Hamlet 3.3]

The connection between the claim ("cool rationality") and the quotation is not visible in the text.

*With framing (effective):*
> When Hamlet comes upon his stepfather at prayer, he demonstrates cool rationality. He impulsively wants to kill Claudius but pauses to reflect: if he kills Claudius while praying, Claudius will go to heaven, but Hamlet wants him damned to hell, so he coolly decides to kill him later: [quotation]

Now the evidence and claim are linked. Readers see exactly which aspect of the quotation matters.

*Rule:* Introduce complex evidence with a sentence explaining what you want readers to get out of it. This applies equally to long quotations, data tables, and figures.

---

### Step 4 — Run the Five-Mechanism Inadvertent Plagiarism Audit

For every source passage you have incorporated — quotation, paraphrase, or summary — check each of the five mechanisms. A single failure on any mechanism is a plagiarism risk regardless of intent.

**Mechanism 1 — Uncited paraphrase or summary**
*Check:* Does every paraphrase and summary have a citation? Even if you never quoted a single word, using someone's ideas without attribution is plagiarism.
*Action:* Add the citation immediately when you write the paraphrase; do not wait until a later revision pass — you may not remember the source originated with someone else.

**Mechanism 2 — Unclosed quotation**
*Check:* If you cite a source and include its words, are those words surrounded by quotation marks (or set as a block quotation)? Even a single borrowed line without quotation marks is plagiarism even if you cited the source.
*Action:* For strikingly original or technically precise phrases, always enclose in quotation marks the first time you use them, even if the phrase is brief. Once you have introduced and cited a phrase, you may use it subsequently without marks.

**Mechanism 3 — Too-close paraphrase**
*Check:* Can you run your finger along your paraphrase sentence and find synonyms for the same ideas in the same order as in the source? If yes, the paraphrase is too close.
*Action:* Read the original, look away, think about it, then write the paraphrase without looking back. Check that the resulting sentence cannot be matched word-for-word-synonym against the original. If it can, revise again.

*Too close (plagiarism):*
> Success seems to depend on a combination of talent and preparation. However, when psychologists closely examine the gifted and their careers, they discover that innate talent plays a much smaller role than preparation (Gladwell 38).

*Acceptable paraphrase:*
> As Gladwell observes, summarizing studies on the highly successful, we tend to overestimate the role of talent and underestimate that of preparation (38).

**Mechanism 4 — Uncited non-common-knowledge idea**
*Check:* For every idea that is not your own, is it either common knowledge in your field or cited? The test: is this idea (a) associated with a specific person AND (b) new enough not to be part of the field's common knowledge? If both, it must be cited.
*Action:* When in doubt, cite. Citing when unnecessary makes you look careful; not citing when necessary makes you look dishonest.

**Mechanism 5 — Free-content fallacy**
*Check:* Did you use any content from freely available online sources without citing? Free and publicly available does not mean unattributed. Website text, Wikipedia, open-access articles — all require citation.
*Action:* Apply the source-recognition test: if the person you borrowed from read your writing, would they recognize their words or ideas? If yes, cite and enclose any exact words.

---

### Step 5 — Apply the Correct Citation Style

Choose and consistently apply one style. There are two structural patterns:

**Author-title pattern** (humanities): bibliography entry begins Author → Title → publication data
**Author-date pattern** (sciences, social sciences): bibliography entry begins Author → Date → title → publication data

The date-first pattern serves fields where recency matters most — readers can immediately spot how old a source is.

**Four common styles:**

| Style | Pattern | In-text format | List name |
|---|---|---|---|
| Chicago author-title | Author-title | Superscript footnote/endnote | Bibliography |
| MLA | Author-title | (Author page) — no comma | Works Cited |
| Chicago author-date | Author-date | (Author date, page) | Bibliography |
| APA | Author-date | (Author, date, p. page) | References |

**Key in-text mechanics:**
- If you name the author in your sentence, drop the name from the parenthetical: *Smith argues that prices rise (1999, 45)* → not *(Smith 1999, 45)*
- If citing multiple works by the same author, add a short title: *(Kay, A Life, 220)*
- If a paraphrase or summary extends across multiple paragraphs, cite only once at the end
- Cite each passage that comes from a different page individually

*Why citation style details matter:* Readers use citations to locate sources. Every deviation from style convention — a missing comma, a wrong order of elements — potentially breaks the bibliographic trail and signals careless work. Experienced readers judge a writer's reliability partly by citation precision.

---

## Output

Deliver the revised text with:

1. **Integrated source passages** — each passage in the correct form (quote, paraphrase, or summary) with the integration method applied
2. **Framing sentences** — added before any complex or long evidence
3. **Plagiarism audit log** — a brief checklist confirming each of the five mechanisms was checked, with notes on any passages that required remediation
4. **Citation style consistency check** — confirmation that in-text references follow the declared style

If working on a full draft, deliver the audit log as an appendix to the revised draft. If working on individual passages, deliver the audit log inline after each passage.

---

## Quick Reference

**Decision tree (one choice per passage):**
- Source not central → **summarize**
- Details matter, not words → **paraphrase**
- Words are evidence / authority / unforgettable / fair-to-quote-to-disagree → **quote**

**Integration methods (for quotations):**
- Self-explanatory in context → **drop in** (*Author says, . . .*)
- Needs setup → **introduce** (sentence + colon + quotation)
- Partial quotation needed → **weave** (grammatically merge)

**Five plagiarism risks:**
1. Paraphrase/summary without citation
2. Source words without quotation marks (even with citation)
3. Too-close paraphrase — same synonyms, same order
4. Non-common-knowledge idea without citation
5. Free/online content treated as unattributable

**Citation style selector:**
- Humanities → Chicago author-title or MLA
- Sciences / social sciences → Chicago author-date or APA
- When unsure → ask your target venue

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/citation-styles.md
# Citation Styles: Four Common Formats

Source: *The Craft of Research, 4th ed.*, Booth et al., Chapter 14 (§14.5, Quick Tip), pp. 233–241.

---

## Two Structural Patterns

All four common citation styles share the same starting point: every citation begins with the author (or editor, or whoever is responsible for the source). What distinguishes styles is what follows the author.

**Author-title pattern:** Author → Title → publication data
Used in humanities, where the work's identity matters more than its recency.

**Author-date pattern:** Author → Date → title → publication data
Used in sciences and social sciences, where recency is critical. Readers can immediately locate the date at the start of the entry.

---

## The Four Styles

### 1. Chicago Author-Title Style

**Authority:** *The Chicago Manual of Style*, 17th ed. (University of Chicago Press). Also known as Turabian style (after Kate L. Turabian's condensed manual for student writers).

**In-text format:** Superscript number → footnote or endnote at bottom of page or end of paper.

> Some have claimed that Castro would reform Cuban politics.⁵
> 
> 5. George Smith, *Travels in Cuba* (Boston: Hasbro Press, 1999), 233.

**List name:** Bibliography

**When to use:** Literary and historical scholarship, humanities generally, any context where footnotes are conventional.

**Parenthetical variant:** Chicago author-title also permits parenthetical citation (Author, Short Title, page) when footnotes are impractical:
> Only one writer provides data on this matter (Kay, *A Life*, 220).

---

### 2. MLA Style

**Authority:** *MLA Handbook for Writers of Research Papers*, 9th ed. (Modern Language Association, 2021).

**In-text format:** Parenthetical — author and page, no comma between them.

> Only one writer provides data on this matter (Kay 220).

If the author is named in the sentence, omit the name from the parenthetical:
> Kay is the only writer who provides data on this matter (220).

**List name:** Works Cited

**When to use:** Literature, composition, languages, some humanities disciplines.

**Note:** MLA and Chicago author-title look nearly identical in parenthetical form. The difference is a comma: Chicago uses *(Kay, 220)*, MLA uses *(Kay 220)*.

---

### 3. Chicago Author-Date Style

**Authority:** Same as Chicago author-title — *The Chicago Manual of Style* covers both patterns.

**In-text format:** Parenthetical — author, year, page.

> Only one writer provides data on this matter (Kay 2006, 220).

If the author is named in the sentence:
> Kay is the only writer who provides data on this matter (2006, 220).

**List name:** Bibliography

**When to use:** Social sciences, some humanities fields when recency matters.

---

### 4. APA Style

**Authority:** *Publication Manual of the American Psychological Association*, 7th ed. (APA, 2020).

**In-text format:** Parenthetical — author, comma, year, comma, page with "p." prefix.

> Only one writer provides data on this matter (Kay, 2006, p. 220).

If the author is named in the sentence:
> Kay is the only writer who provides data on this matter (2006, p. 220).

**List name:** References

**When to use:** Psychology, social sciences, education, health sciences, and related fields.

---

## Quick-Reference Comparison Table

| Style | Pattern | In-text | Author in sentence → | List name |
|---|---|---|---|---|
| Chicago author-title | Author-title | Footnote/endnote, or (Author, page) | Drop author from parenthetical | Bibliography |
| MLA | Author-title | (Author page) — no comma | Drop author | Works Cited |
| Chicago author-date | Author-date | (Author year, page) | (year, page) | Bibliography |
| APA | Author-date | (Author, year, p. page) | (year, p. page) | References |

---

## Mechanics Shared Across All Four Styles

**Multiple works by the same author:** Add a short title to distinguish them.
- Chicago/MLA: (Kay, *A Life*, 220)
- APA/Chicago author-date: (Kay, 2006, 220) and (Kay, 1999, 45) — the different dates distinguish the works

**Multi-page paraphrase or summary:** Cite only once at the end. If you resume citing the source after a digression, start the citation again.

**Different pages from the same source:** Cite each page individually.

**Style software:** Many researchers use Zotero, Mendeley, or similar tools to auto-generate citations. These tools are accurate for standard cases but make errors with unusual source types (interviews, archival documents, legal materials). Always verify generated citations against the style guide, especially for non-standard sources.

**The golden rule:** Follow the style exactly — every comma, every space, every capitalization choice. Readers use citations to locate sources. A deviation from style convention may break the bibliographic trail and signals that you have not learned the rules.

FILE:references/plagiarism-mechanisms.md
# Inadvertent Plagiarism: Five Mechanisms

Source: *The Craft of Research, 4th ed.*, Booth et al., Chapter 14 (§14.6), pp. 236–241.

This reference expands on the five inadvertent plagiarism mechanisms summarized in the main skill. Use it when you need to explain a plagiarism risk to a writer or want a deeper treatment of edge cases.

---

## Why Inadvertent Plagiarism Happens

Intentional plagiarism (buying a paper, copying a passage) is not the subject here. Inadvertent plagiarism happens because:

- Writers lose track of which words and ideas are theirs and which are borrowed
- Writers mistakenly believe that paraphrasing or citing a source makes all other obligations disappear
- Writers misunderstand the rules about free content, common knowledge, and close paraphrase
- Writers plan to add citations "later" and never do

The consequences are the same regardless of intent: if a reader can identify your words or ideas as someone else's, the plagiarism charge sticks. Intent is not a defense in academic or professional writing. "We read words, not minds." (Booth et al., 14.6.5)

---

## Mechanism 1: Uncited Paraphrase or Summary

**What it is:** You restate or condense a source's ideas in your own words without attributing the ideas to the source.

**Why it matters:** The citation obligation follows the idea, not just the words. If you paraphrase or summarize without a citation, readers have no way to verify the information or follow the bibliographic trail.

**When it commonly happens:** Writers draft quickly and think "I'll add the citation at the end." By the end, they no longer remember which sentences originated from a source.

**Prevention rule:** Add the citation at the moment you write the paraphrase or summary — not in a later pass.

**Edge case:** A paraphrase or summary that spans multiple paragraphs needs only one citation, placed at the end of the final paragraph. But if you return to the same source after writing about something else, start the citation again.

---

## Mechanism 2: Unclosed Quotation (Source Words Without Quotation Marks)

**What it is:** You borrow exact words from a source, you cite the source, but you do not put the borrowed words in quotation marks (or a block quotation).

**Why it matters:** The citation tells readers you used this source. The quotation marks tell readers exactly which words are not yours. Without quotation marks, readers cannot tell where the source text ends and your text begins. Even one borrowed line without quotation marks is plagiarism, even with a citation present.

**When it commonly happens:** Writers paraphrase most of a passage and leave one or two exact phrases from the original embedded invisibly in their prose.

**Prevention rule:** After drafting a passage that incorporates source material, re-read it sentence by sentence. Ask: "Did I write every word in this sentence, or did some come from the source?" Enclose any borrowed phrase in quotation marks.

**Edge case — striking phrases:** Technical coinages and strikingly original phrases require quotation marks even when the rest of the sentence is your own. Example: if you borrow Diamond's phrase "technology begets more technology," you must enclose it in quotes the first time you use it. After that first cited use, you may use the phrase as if it were common vocabulary.

**Edge case — field conventions:** Legal and statutory text is sometimes quoted without quotation marks by convention in legal writing. Follow your field's norms, but confirm the convention before departing from the default.

---

## Mechanism 3: Too-Close Paraphrase

**What it is:** You restate a source's ideas in "your own words," but the sentence structure and word choices follow the original so closely that a reader can match your version against the source word by synonym.

**Why it matters:** Too-close paraphrase gives the superficial appearance of independent thought while actually copying the original's intellectual structure. Courts and plagiarism detectors treat this as infringement even when every individual word is technically different.

**When it commonly happens:** Writers read a passage, then immediately write a "paraphrase" while looking at or glancing back at the original. The original's sentence architecture dominates the result.

**The test:** Run your finger along your paraphrase. Can you find a synonym for each idea in the same order as the original? If yes, the paraphrase is too close.

**Prevention rule:** Read the original. Put it aside. Think about the idea for a moment. Then write the paraphrase without looking back. This forces you to reconstruct the idea from memory rather than translate the text.

**Example (from Booth et al., Gladwell's *Outliers*):**

*Original:* "Achievement is talent plus preparation. The problem with this view is that the closer psychologists look at the careers of the gifted, the smaller the role innate talent seems to play and the bigger the role preparation seems to play."

*Too-close (plagiarism):* "Success seems to depend on a combination of talent and preparation. However, when psychologists closely examine the gifted and their careers, they discover that innate talent plays a much smaller role than preparation."

*Acceptable paraphrase:* "As Gladwell observes, summarizing studies on the highly successful, we tend to overestimate the role of talent and underestimate that of preparation."

The acceptable version is not a synonym substitution; it is a genuinely different sentence that conveys the idea.

---

## Mechanism 4: Uncited Non-Common-Knowledge Idea

**What it is:** You use an idea that originated with a specific person and is new enough not to be common knowledge in your field, but you do not cite it.

**Why it matters:** Ideas that meet both conditions below require citation, just as exact quotations do:
1. The idea is associated with a specific named person or group of researchers
2. The idea is not yet common knowledge in the relevant field

**When it commonly happens:** Writers assume that if an idea is widely discussed, it must be common knowledge. But widely discussed is not the same as established common knowledge. An idea that is still associated with named researchers — still credited to them in conversation and literature — is not common knowledge.

**How to distinguish common knowledge from citable ideas:**

*Common knowledge (no citation needed):* The world is round. Psychologists believe humans think and feel in different parts of the brain. (So familiar no reader would think you are claiming it as your own.)

*Citable idea:* The claim that emotions are crucial to rational decision-making. (Still associated with specific researchers, still new enough to need attribution.)

**Heuristic:** If someone could reasonably credit the idea to a named researcher — if the idea "belongs" to someone — cite it.

---

## Mechanism 5: The Free-Content Fallacy

**What it is:** The belief that content freely available online or in the public domain does not require citation or quotation marks.

**Why it matters:** Accessibility is not the same as unattributed use. A text is free to read; it is not free to use without attribution. The legal concept of fair use affects reproduction, not attribution. You must cite text whether you paid for it, found it freely online, or found it in a public-domain archive.

**The recognition test:** Would the person you borrowed from, reading your writing, recognize their words or ideas as their own — including paraphrases, summaries, or general ideas or methods? If yes, you must:
- Cite the source
- Enclose any exact words in quotation marks or set them as a block quotation

There are no exceptions for online content, Wikipedia, open-access journals, or government publications. There are no excuses based on intent.

---

## The Deeper Principle

Plagiarism harms more than your credibility. When you fail to cite a source's intellectual contribution, you deprive the original author of credit that may affect their reputation, funding, promotions, and career. You also hide the bibliographic trail that allows future researchers to discover the foundational work. Accurate citation is both an ethical obligation and a practical contribution to cumulative knowledge.

---
name: source-evaluator
description: Evaluate, triage, and actively read a set of research sources — books, articles, and online materials — by applying a dual-axis relevance-and-reliability screen, source-type skim protocols, and a two-pass active reading method that extracts data, arguments to respond to, and generative agreements and disagreements. Use this skill when you have a candidate source list and need to cut it to a workable set, when you need to verify that a source is credible before citing it, when you are reading sources to find a research problem or refine a hypothesis, when you need to take notes that accurately capture what a source argues without misrepresenting it, or when you must identify where sources agree and disagree so you can position your own argument within a field's conversation.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/source-evaluator
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [5, 6]
tags: [research-methodology, source-evaluation, critical-reading, academic-writing, note-taking, literature-review]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "Candidate source list or a single source to evaluate. Can be a bibliography file, a list of titles/URLs, or a document set containing source files."
  tools-required: [Read, Write]
  tools-optional: [WebFetch]
  mcps-required: []
  environment: "Document set preferred. Works from a free-text source list. Output: evaluated-sources.md with relevance/reliability verdicts and reading notes."
discovery:
  goal: "Produce a ranked, evaluated source list with relevance verdicts, reliability assessments, and reading notes that capture arguments, data, and productive agreements/disagreements."
  tasks:
    - "Classify each source as primary, secondary, or tertiary relative to the research question"
    - "Apply the relevance skim protocol appropriate to each source type"
    - "Score each source on the reliability rubric (publisher, peer review, author, currency, apparatus)"
    - "Read high-priority sources using the two-pass active reading protocol"
    - "Identify creative agreements and disagreements to generate or refine the research problem"
    - "Record notes using the quote/paraphrase/summarize decision rule and context-preservation guidelines"
  audience: "researchers, students, academics, professionals conducting any evidence-based inquiry"
  when_to_use: "When you have candidate sources and need to decide which to read, how carefully to read them, and how to take notes that will hold up under scrutiny"
  environment: "Document set (source-list.md, research-question.md) or free-text description of the research question and sources"
  quality: placeholder
---

# Source Evaluator

## When to Use

You have a list of candidate sources — books, journal articles, websites, reports — and must decide which are worth reading, how deeply to read them, and how to capture what they say so you can use them reliably in your own argument. This skill applies when:

- You have search results or a bibliography and need to triage it to a manageable set
- You are unsure whether a source is credible enough to cite in a scholarly or professional context
- You are reading to find or sharpen a research problem, not just collect supporting evidence
- You need to understand where sources agree and disagree to position your own argument
- You have taken notes but worry they may misrepresent what a source actually claims

**The core pattern:** evaluate every source on two axes — relevance to your question, and reliability of its claims — before committing reading time. Then read high-priority sources in two passes: first generously (to understand), then critically (to respond). Record what the source says and what you think about it in clearly separated layers.

Before starting, confirm you have:
- A stated research question or working hypothesis (even a rough one)
- At least one candidate source to evaluate

---

## Context and Input Gathering

### Required Context

- **Research question:** What are you trying to find out? Even a topic ("desegregation in Midwest school boards, 1980s") is workable, but a specific question ("How did school boards in the Midwest respond to federal desegregation mandates?") makes relevance assessment more accurate.
- **Source list:** The sources you want to evaluate. URLs, titles, file paths, or a bibliography.

### Observable Context

If documents are provided, read them for:
- Keywords and named concepts central to the research question
- Source type signals (peer-reviewed journal, book from university press, Wikipedia, industry report, blog)
- Publication date and edition information
- Author credentials and institutional affiliation (if visible)
- Presence or absence of bibliography/references

### Default Assumptions

- If no research question is provided → ask for one before evaluating. Relevance cannot be assessed without it.
- If source type is ambiguous → default to the more skeptical classification (treat a non-peer-reviewed article as tertiary until proven otherwise)
- If currency norms for the field are unknown → use the conservative rule: flag anything older than 10 years for humanities fields, 5 years for social sciences, 2 years for sciences and technology

### Sufficiency Check

Before evaluating, confirm: "Do I know enough about the research question to judge whether a source addresses it?" If the question is still a vague topic, help the user sharpen it first.

---

## Process

### Step 1: Classify Each Source by Type

**ACTION:** For each source, assign it a type — primary, secondary, or tertiary — relative to the specific research question.

**WHY:** Source type determines how you use a source, not just how you find it. The same document can be primary for one project and secondary for another. A journal article analyzing Victorian swearing patterns is secondary if your question is about Victorian gender norms, but primary if your question is about how scholars have interpreted Victorian gender norms. Getting this wrong leads to misaligned reading strategies: you read a tertiary source (encyclopedia) as if it were evidence, or you miss that an article is primary data for your project.

**Classification rules:**
- **Primary source:** original materials that provide raw data or evidence for your question. In history: documents, artifacts, letters from the period. In literature/arts: the text or work you are interpreting. In social sciences: survey data, census records, interviews. In sciences: reports of original experiments.
- **Secondary source:** books, articles, or reports based on primary sources, written for scholarly or professional audiences. The best are peer-reviewed or published by university presses. These constitute the field's "literature" — the ongoing scholarly conversation you are entering.
- **Tertiary source:** syntheses of secondary sources for general audiences — textbooks, encyclopedias (including Wikipedia), mass-market publications. Useful for orientation and finding leads to secondary sources. Not appropriate as evidence in scholarly arguments.

**Key rule:** Classifications are relative to your project. If you change your focus, re-classify.

**Output format:**
```
Source: [title/URL]
Type: [primary / secondary / tertiary]
Reason: [one sentence]
```

---

### Step 2: Apply the Relevance Skim Protocol

**ACTION:** Skim each source using the protocol matched to its format. Do not read fully at this stage — the goal is a binary decision: pursue or set aside.

**WHY:** Full reading before relevance assessment wastes time on sources that turn out not to address your question. Skimming forces you to use the structural features that good scholarly sources provide precisely for this purpose — abstracts, indices, opening paragraphs, tables of contents. The skim takes 5-15 minutes per source. A full read takes hours. The investment ratio is decisive.

**See:** [Relevance Skim Protocols](references/relevance-skim-protocols.md) for the full per-format checklists.

**Summary by source format:**

**Book:**
1. Skim the index for your keywords; check the pages on which those words appear
2. Read first and last paragraphs of chapters where your keywords appear most densely
3. Skim prologues, introductions, and summary chapters
4. Read the last chapter's first and last 2-3 pages
5. For edited collections, skim the editor's introduction
6. Check the bibliography for titles relevant to your topic

**Journal article:**
1. Read the abstract if present
2. Skim introduction and conclusion; if headings are absent, read the first 6-7 paragraphs and last 4-5
3. Skim section headings; read first and last paragraphs of each section
4. Check the bibliography for related titles

**Online source:**
1. If it resembles a journal article, apply the article protocol plus keyword search
2. Look for sections labeled "introduction," "overview," "summary," or "about"
3. Check any site map or index for keywords
4. Use the site's search function if available

**Relevance verdict:** After skimming, assign one of three statuses:
- **Read:** clearly relevant; proceed to Step 3 reliability check and then full reading
- **Monitor:** potentially relevant; set aside but note for possible retrieval if other sources prove insufficient
- **Discard:** not relevant to this question

---

### Step 3: Apply the Reliability Rubric

**ACTION:** For each source you marked **Read**, assess reliability using the 5-criterion rubric below. Only proceed to full reading on sources that pass.

**WHY:** Relevance without reliability is dangerous — a source that addresses your question but makes unreliable claims can contaminate your argument. You cannot judge reliability fully until you read, but the rubric identifies red flags before you invest reading time. Importantly, reliability indicators are not guarantees: a peer-reviewed article in a reputable journal can still contain weak arguments or thin data. The rubric tells you where to be cautious, not where to trust uncritically.

**See:** [Reliability Rubric](references/reliability-rubric.md) for the expanded 10-criterion version.

**Core 5-criterion reliability check:**

| Criterion | What to look for | Red flags |
|---|---|---|
| **1. Publication venue** | University press, peer-reviewed journal, established professional publisher | Self-published, vanity press, no publisher listed, sensationalist claims on cover |
| **2. Peer review** | "Peer reviewed" label, journal review process, university press review | Commercial magazines, blogs without editorial process, preprints without review notice |
| **3. Author credentials** | Academic affiliation, track record in the field, cited in other reputable sources | No credentials listed, credentials irrelevant to topic, financial ties to interested parties on contested topics |
| **4. Currency** | Publication date relative to field norms (see below) | Textbooks (assume outdated); sources older than field norms allow; earlier editions when recent editions exist |
| **5. Scholarly apparatus** | Notes, bibliography, citations that can be checked | No bibliography in a book-length argument; website with no attribution, date, or sponsor |

**Field-specific currency norms:**
- Sciences and technology: 2 years is the outer limit for findings; foundational theory may be older
- Social sciences: 10 years is often cited as the outer limit; check recent literature for updates
- Humanities (literary criticism, history, philosophy): articles and books can remain relevant for decades; foundational works cited regardless of age
- **Rule of thumb:** Look at the citations in 2-3 recent articles in your target field. The oldest sources cited regularly establish the floor for acceptable age.

**Advanced criteria** (for experienced researchers or high-stakes projects):
- Has the book been reviewed favorably in field-specific review journals?
- Has the source been frequently cited by others? (High citation count = high impact factor; use forward citation search)

**Reliability verdict:**
- **Reliable:** passes all 5 criteria or has minor issues on 1 criterion with strengths on the others
- **Use with caution:** issues on 2 criteria; flag the specific concerns in your notes
- **Exclude:** fails 3 or more criteria, or fails on peer review and author credentials simultaneously

---

### Step 4: Two-Pass Active Reading

**ACTION:** Read each **Read + Reliable** source twice using the protocol below. Record bibliographic information completely before reading anything else.

**WHY:** Passive reading — absorbing what a source says without responding to it — produces notes that are dead data. You remember the source's argument but not your reactions to it, and your reactions are where your own research problem lives. Two-pass reading forces the separation: the first pass builds understanding on the source's own terms (so you do not misrepresent it later); the second pass builds your critical response (so you do not accept it uncritically). Experienced researchers do this naturally; the protocol makes it explicit for those building the habit.

**Step 4a: Record complete bibliographic information first**

Before reading a single page, record all bibliographic data for the source. (See [Bibliographic Recording Checklist](references/bibliographic-recording-checklist.md) for field-specific lists.) Record it in whatever format you choose for your own notes, but record it completely — incomplete bibliographic records have killed publications.

**Step 4b: First pass — read generously**

Read to understand. Pay attention to what sparks interest. Reread passages that puzzle or confuse you. Do not look for disagreements yet. Read in ways that help the source make sense. Resist the temptation to criticize while you are still learning what the source actually argues.

**During first pass, mark or note:**
- The source's main claim (thesis)
- The evidence it relies on most heavily
- Passages that are central versus qualifications or concessions the author makes in passing
- Passages you find compelling or surprising

**Understand before you respond:** If you cannot summarize the source's argument in 2-3 sentences, you do not yet understand it well enough to respond to it.

**Step 4c: Second pass — read critically**

Read slowly and with your own research question in mind. Think about how you would respond to each claim. Record your responses in your notes — clearly separated from what the source says (see Step 5).

During the second pass, look for productive responses in two directions:

**Creative agreements** (sources you believe — extend them):
1. **Offer additional support:** The source makes a claim; you can support it with better or newer evidence
2. **Confirm unsupported claims:** The source speculates or assumes something; you can prove it
3. **Apply more widely:** The source correctly applies a principle to one case; you can show it applies more broadly

**Creative disagreements** (sources you doubt — challenge them):
1. **Contradiction of kind:** The source categorizes something incorrectly (e.g., graffiti is vandalism; you argue it is public art)
2. **Part-whole contradiction:** The source mistakes how the parts of something relate to each other
3. **Developmental contradiction:** The source mistakes the origin or trajectory of something (e.g., a trend the source says is rising is actually falling)
4. **Causal contradiction:** The source mistakes a causal relationship (e.g., claims X causes Y; you show they are both caused by Z)
5. **Contradiction of perspective:** The source analyzes something from one framework; a different framework reveals a new truth

**See:** [Active Reading Response Taxonomy](references/active-reading-response-taxonomy.md) for templates and examples for each type.

---

### Step 5: Take Notes That Hold Up

**ACTION:** Record your notes using the four-layer structure below. Always separate what the source says from what you think about it.

**WHY:** The most common failure in research note-taking is blending the source's words with your paraphrase of them, or blending both with your own interpretation. This causes two problems: (1) you may later cite your paraphrase as a direct quotation, undermining your credibility; (2) you may attribute your own ideas to the source or the source's ideas to yourself, which is a form of inadvertent plagiarism. Keeping the layers separate is not just scholarly caution — it is how you maintain the distinction between what you know and what you think, which is the foundation of honest argument.

**Four-layer note structure:**

```
SOURCE: [Author, short title, page]
TOPIC TAGS: [keywords for sorting and retrieval]

SUMMARY: [2-4 sentences capturing the source's main argument in your own words]

QUOTATION(S): [exact words, in quotation marks, with page numbers]
"..." (p. XX)

PARAPHRASE(S): [source's meaning in your words; replace most words, not just a few]
[Paraphrase of p. XX]: ...

MY RESPONSE: [your reactions, agreements, disagreements, questions — clearly marked as yours]
[Response]: ...
```

**Quote, paraphrase, or summarize — when to use each:**
- **Summarize** when you need only the point of a passage, section, or whole source. Summary is useful for context or for sources only loosely relevant to your argument. A summary alone is never evidence.
- **Paraphrase** when the specific words matter less than the meaning. Replace most of the words and phrasing — not just a word or two. A paraphrase is weaker evidence than a direct quotation.
- **Quote exactly** when: (a) the words themselves are evidence (e.g., exact language in a historical document); (b) the source is an authority you will directly challenge or rely on; (c) the phrasing is so precise or striking that your paraphrase would lose it; (d) you are disagreeing with a specific claim and must represent it fairly.

**Never abbreviate a quotation expecting to reconstruct it later. You cannot.**

**Context-preservation rules (prevent misrepresentation):**
1. Record not just what a passage says but the line of reasoning it serves in the source. Note the pages before and after.
2. When you record a claim, note its role: main point, minor point, qualification, or concession. A concession reported as a main point is a misrepresentation — even if accidental.
3. Record the scope and confidence of claims precisely. "Chemicals in french fries cause cancer" and "some chemicals in french fries correlate with a higher incidence of some cancers" are not the same claim.
4. Note when a source is summarizing someone else's view rather than stating its own. Misattributing a summary as the source's own position is a common error.
5. Note why sources agree or disagree when you spot convergence or conflict — the basis of agreement matters as much as the fact of it.

---

## Output

Produce an `evaluated-sources.md` file with the following sections:

**1. Source Inventory**
For each candidate source: type classification, relevance verdict (Read / Monitor / Discard), reliability verdict (Reliable / Use with Caution / Exclude), and a one-line rationale.

**2. Reading Notes**
For each source you read fully: the four-layer note structure (summary, quotations, paraphrases, your response).

**3. Productive Tensions**
A brief section listing the most significant creative agreements and disagreements you identified across sources — these are the seeds of your own argument.

---

## Examples

**Example 1: Triaging a bibliography**

Research question: "How did social media use during the 2020 US election affect political polarization?"

Given 12 sources, the evaluator would:
- Classify Wikipedia and a Pew Research explainer as tertiary (useful for orientation, not citation)
- Mark a 2019 peer-reviewed article on social media and polarization as secondary + Read
- Flag a 2016 book on political communication as secondary + Reliable but flag for currency (pre-2020)
- Mark a personal blog post with no author or date as Exclude

**Example 2: Reading for a research problem**

A student reading a secondary source on urban desegregation finds the source argues that federal mandates uniformly increased white flight. The student applies creative disagreement type 3 (developmental contradiction): data from three Midwestern cities suggest the pattern peaked and then reversed after 5 years. That disagreement is the research problem.

**Example 3: Context-preserving note**

Poor note: "Jones (p. 123): The war was caused by Z."
Better note: "Jones argues the war was caused by X, Y, and Z (p. 123); he considers Z the most important cause (p. 123), for reasons on pp. 124-28. Note: X and Y are introduced as background context, not as Jones's main causal argument."

---

## References

- [Relevance Skim Protocols](references/relevance-skim-protocols.md) — full per-format checklists for books, articles, and online sources
- [Reliability Rubric](references/reliability-rubric.md) — expanded 10-criterion rubric with field-specific guidance
- [Active Reading Response Taxonomy](references/active-reading-response-taxonomy.md) — templates for 3 creative agreement types and 5 creative disagreement types
- [Bibliographic Recording Checklist](references/bibliographic-recording-checklist.md) — field-specific bibliographic data requirements

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/active-reading-response-taxonomy.md
# Active Reading Response Taxonomy

Templates for generating productive intellectual responses while reading secondary sources. Use these during the second pass (critical reading) to identify research problems and argument positions. The goal is not to force a response on every passage, but to recognize when a passage triggers a genuine reaction and give that reaction a precise form.

---

## Part 1: Creative Agreement Types

Use when you believe what the source argues. Agreeing is not passive — it is an opportunity to extend, strengthen, or broaden the source's claim into new territory.

### Agreement Type 1: Offer Additional Support

**Pattern:** The source makes a claim with existing evidence; you can support the same claim with better, newer, or different evidence.

**When to use:** The source's evidence is old, thin, or limited to a particular context and you know of stronger or more current evidence.

**Template:**
> [Source] supports [claim] using [type of evidence], but [stronger/newer evidence] provides better support for the same claim.

**Patterns:**
- Source supports claim with old evidence → you offer new evidence
- Source supports claim with weak evidence (anecdotes, single case) → you offer stronger evidence (meta-analysis, multiple cases, experimental data)
- Source supports claim in one domain → you offer evidence from a different domain confirming the same pattern

**Example:**
> Smith uses anecdotal accounts to argue that the Alamo story had mythic status beyond Texas, but editorial coverage in major Northern newspapers during the 1840s-1860s offers stronger systematic evidence for the same claim.

---

### Agreement Type 2: Confirm Unsupported Claims

**Pattern:** The source speculates about or assumes something; you can provide evidence to prove it.

**When to use:** The source flags a gap ("further research is needed," "we might expect that...") and you can fill it, or the source assumes a premise without demonstrating it.

**Template:**
> [Source] speculates/assumes that [X], but [evidence] shows that it is actually the case.

**Patterns:**
- Source speculates [X] might be true → you offer evidence that it is
- Source assumes [premise] is true without proving it → you can prove it
- Source identifies a pattern in one period → you can show it holds across time

**Example:**
> Smith recommends visualization techniques for improving athletic performance but provides no mechanism. Neuroimaging studies of athletes during mental rehearsal show activation of the same motor pathways used during physical practice — providing the mechanistic basis Smith assumes but does not demonstrate.

---

### Agreement Type 3: Apply a Claim More Widely

**Pattern:** The source correctly applies a principle or finding to one case; you can show it applies to additional cases.

**When to use:** The source's claim is restricted to a specific population, time period, domain, or type — and you can demonstrate it applies more generally (or in a new specific context).

**Template:**
> [Source] correctly applies [claim] to [context A]; the same principle applies to [context B].

**Patterns:**
- Source applies finding to one population → you show it applies to another
- Source claims [X] is true in a specific situation → you show it is true generally
- Source applies one explanatory framework → you apply the same framework to a new case

**Example:**
> Smith argues that learning through multiple metaphors improves conceptual retention in medical education; the same pattern appears in engineering and legal education, suggesting a general principle of analogical scaffolding rather than a domain-specific phenomenon.

---

## Part 2: Creative Disagreement Types

Use when you find something puzzling, inaccurate, or simplistic in a source. Do not dismiss disagreements — they often point directly to a research problem worth pursuing.

### Disagreement Type 1: Contradiction of Kind

**Pattern:** The source categorizes something as a type of X; you argue it is better understood as a type of Y (or not a type of X at all).

**When to use:** The source's classification misrepresents the nature of the phenomenon and the misclassification has consequences for the analysis.

**Templates:**
- Source claims [A] is a kind of [X], but it is better understood as a kind of [Y]
- Source claims [A] always has [feature] as one of its qualities, but it doesn't
- Source claims [A] is normal/significant/useful in the way [X] are, but it isn't
- Reverse: Source claims [A] is *not* a kind of [X], but it is

**Example:**
> Smith classifies graffiti as a form of vandalism and analyzes it accordingly; it is better understood as a form of public art, which changes what counts as quality, who counts as a legitimate practitioner, and what role the community plays as audience.

---

### Disagreement Type 2: Part-Whole Contradiction

**Pattern:** The source mistakes how the parts of something relate to the whole, or mistakes the relationship between sub-components.

**When to use:** The source treats a part as the whole, or claims one sub-component relates to another in a way that is incorrect or incomplete.

**Templates:**
- Source claims [A] is a part of [B], but it is actually [a whole / a distinct category]
- Source claims one part of [X] relates to another in [way], but the relationship is actually [different]
- Source claims every [X] has [feature] as one of its parts, but it doesn't (or only some do)

**Example:**
> Smith treats procedural memory as a component of declarative memory in his model of learning; current cognitive science treats them as distinct memory systems with different neural substrates — making them parts of a larger whole (long-term memory) rather than one being a part of the other.

---

### Disagreement Type 3: Developmental or Historical Contradiction

**Pattern:** The source mistakes the origin, history, or developmental trajectory of something.

**When to use:** The source claims something originated in a particular place or time, developed in a particular direction, or is currently changing in a particular way — and the evidence contradicts that account.

**Templates:**
- Source claims [X] is changing in [direction], but it isn't (or it is changing in the opposite direction)
- Source claims [X] originated in [place/time], but it actually originated in [different place/time]
- Source claims [X] develops in [way] over time, but evidence shows [different developmental pattern]

**Example:**
> Smith argues that public trust in newspapers has been declining continuously since the 1970s; survey data from 1985-2005 shows the decline plateaued and partially recovered before resuming — making the trajectory non-linear in ways that complicate Smith's causal account.

---

### Disagreement Type 4: External Cause-Effect Contradiction

**Pattern:** The source mistakes a causal relationship — claiming X causes Y when the relationship is absent, reversed, or more complex.

**When to use:** The source's causal claim is central to its argument and you have or can identify evidence that the causal mechanism is wrong, incomplete, or conflated with correlation.

**Templates:**
- Source claims [X] causes [Y], but evidence shows they are both caused by [Z]
- Source claims [X] is sufficient to cause [Y], but [Y] requires [X + other conditions]
- Source claims [X] causes only [Y], but [X] also causes [Z] (with different implications)
- Reverse: Source claims [X] does *not* cause [Y], but evidence shows it does

**Example:**
> Smith claims that legalizing marijuana increases use among teenagers; longitudinal data from states with early legalization shows no significant change in teen use rates compared to states without legalization — suggesting the causal relationship Smith posits is not supported by the available evidence.

---

### Disagreement Type 5: Contradiction of Perspective

**Pattern:** The source analyzes a topic from within one framework, but a different conceptual framework reveals aspects the source's framework cannot see.

**When to use:** The source's conclusions are limited not by the evidence but by the analytical lens it applies. A different framework — economic, historical, feminist, ecological, philosophical, etc. — would yield different and arguably more complete or accurate conclusions.

**Templates:**
- Source discusses [X] from the perspective of [framework A], but viewing [X] through [framework B] reveals [new truth]
- Source analyzes [X] using [theory/value system A], but [theory/value system B] yields a more accurate or complete account
- The standard view treats [X] as [type of problem]; reconceived as [different type of problem], [X] is better understood as [reframing]

**Example:**
> Smith treats advertising exclusively as an economic phenomenon, optimizing for product sales. Analyzed as a cultural form, advertising also functions as a medium for public art and social commentary — a function that the economic framework renders invisible but that matters for understanding advertising's social effects and regulation.

---

## Notes on Using This Taxonomy

**Not every passage calls for a response.** Use the taxonomy when you feel a genuine intellectual reaction — agreement that surprises you, or disagreement you want to pursue. Forced responses produce hollow arguments.

**Disagreements are more generative than agreements** for finding research problems. But do not discount creative agreements: extending a strong finding into new territory is a legitimate and often undervalued form of research contribution.

**Categories overlap.** A single response might involve both a causal contradiction and a perspective contradiction (e.g., the source's causal claim is wrong because it uses an incomplete framework). Use the taxonomy to clarify your thinking, not to force a single label.

**Keep your responses separate from the source's argument.** In your notes, mark your responses explicitly as [MY RESPONSE] so you do not later confuse your interpretation with what the source actually says.

FILE:references/bibliographic-recording-checklist.md
# Bibliographic Recording Checklist

Record complete bibliographic data before reading any source. This is the single most damage-preventing habit in research. Incomplete records have delayed and killed publications. Spend the 2 minutes now.

Record in whatever format works for your notes — you will adapt to your field's citation style when you write. But record everything listed here.

---

## Printed Books

- Author(s) — full names as they appear on the title page
- Title (including subtitle) — exactly as printed
- Editor(s) and translator(s) — full names, with role specified (ed., trans.)
- Edition — if not the first (2nd ed., revised ed., etc.)
- Volume — if part of a multi-volume work (vol. 3 of 5)
- Place of publication — first city listed if more than one
- Publisher — full name
- Date of publication — year on copyright page (not cover year if different)
- Page numbers of the specific chapters or articles you consulted
- Library call number — if you accessed the physical book in a library (for re-retrieval)
- ISBN — for future ordering or verification

---

## Electronic Books

Record everything listed for printed books, plus:

- URL or DOI — the persistent link if available; avoid recording the URL of a search results page
- Name of database or platform — if accessed through a licensed database (e.g., ProQuest Ebook Central)
- Date of access — month, day, year you retrieved it
- Electronic format of the book — PDF, EPUB, HTML, etc.

---

## Journal Articles (Print or Electronic)

- Author(s) — full names as they appear in the article
- Title of the article (including subtitle)
- Title of the journal — full title, not abbreviation
- Volume number
- Issue number (within the volume)
- Date of publication — year; add month or season if the journal uses them
- Page numbers of the article (first page — last page)
- Library call number — if accessed physically
- DOI — digital object identifier; the most stable identifier for electronic retrieval
- URL — if no DOI, record the URL; note that journal article URLs can change
- Name of database — if accessed through a licensed database
- Date of access — for online access

---

## Online Sources (Websites, Reports, Web Publications)

Record as much of the following as the source provides:

- Author(s) — full name(s) if identified; note if no author is listed
- Title of the specific page or document consulted
- Title of the website or organization hosting the content
- URL — copy the exact URL of the page you consulted, not just the homepage
- Name of database — if accessed through a licensed or institutional database
- Date of publication or last update — look for a "last updated" timestamp; note if absent
- Date of access — month, day, year you retrieved it
- Webmaster or site administrator — if identified (relevant for reliability assessment)
- Sponsor or organization — the body responsible for the site

**If you access a printed text through an online source:** Record bibliographic data from the original print publication AND your online access source.

---

## Photocopied or Scanned Materials

If you photocopy or scan a passage from a book or article:
- Also scan or photocopy the title page
- Also scan or photocopy the bibliographic information on the reverse of the title page (copyright page)
- Write the library call number on the copy if you know it
- Note the library where you found the physical item

The call number allows you to re-locate the source if you need to return to it.

---

## Why Every Field Item Matters

**Author names:** Required for citation; variant name spellings cause retrieval failures
**Subtitle:** Often contains the key terms that identify what the source is actually about
**Edition:** Earlier editions can have different page numbers, different content, and sometimes different conclusions — citing the wrong edition undermines verification
**Volume:** Without it, a multi-volume work citation is incomplete
**Publisher and press location:** Required for most citation formats; needed to assess credibility
**Date of publication:** Required for currency assessment and citation
**Page numbers:** Without specific pages, you cannot cite evidence; without the full article range, you cannot confirm you read the whole piece
**Call number:** Not required in citations, but enables re-retrieval during writing when you cannot remember where you found something
**DOI/URL + access date:** Online sources can move or disappear; the access date documents when you verified the content was there

FILE:references/relevance-skim-protocols.md
# Relevance Skim Protocols

Quick-reference checklists for assessing whether a source is worth reading fully. Complete each checklist in order — stop and mark the source **Read** as soon as you find clear relevance, or **Discard** as soon as you confirm it is not relevant. If evidence is mixed, mark **Monitor**.

---

## Protocol A: Books (Print or Digital)

1. **Index scan:** Find your primary keywords in the index. Check the pages listed under those keywords.
   - If your keywords do not appear → likely Discard (unless the topic uses different terminology; try 2-3 synonyms first)
   - If keywords appear on many pages → promising; continue

2. **Chapter-level keyword density:** Identify chapters with the most keyword hits. Read the first and last paragraph of each such chapter.
   - Do the opening and closing paragraphs address your question? → continue to Read verdict
   - Do they address related issues only tangentially? → Monitor

3. **Front matter skim:** Read the prologue, preface, or introduction. Read any "How to use this book" or overview section.

4. **Summary structures:** Read any dedicated summary chapters, epilogues, or conclusion chapters — specifically their first and last 2-3 pages.

5. **Edited collections:** Read the editor's introduction in full; it maps the book's intellectual territory and often explains which chapters are most central.

6. **Bibliography check:** Scan the bibliography or works cited for titles you already know are relevant to your topic. A source that cites your core sources is likely part of the same conversation.

**For e-books:** Follow steps 1-6, and also run a full-text search for your primary keywords. Review the surrounding context (1-2 paragraphs) around each hit to assess density.

**Verdict signals:**
- Read: keywords appear frequently; first/last paragraphs of relevant chapters directly address your question; bibliography overlaps significantly with your known sources
- Monitor: keywords appear but only in passing; source addresses a related but not directly relevant question
- Discard: keywords absent or appear only as one-word mentions; source's central question is clearly different from yours

---

## Protocol B: Journal Articles

1. **Abstract:** Read the full abstract if present. Does the article's question intersect with yours? Does it report findings or arguments you need to engage with?

2. **Introduction and conclusion:** If no abstract, or to confirm relevance after reading the abstract:
   - If headings mark introduction and conclusion clearly: read both in full
   - If headings are absent: read the first 6-7 paragraphs and the last 4-5 paragraphs

3. **Section headings:** Scan all headings. Do they suggest the article addresses your specific question or a related sub-question?

4. **Section skim:** For sections whose headings look relevant, read the first and last paragraph of each section.

5. **Bibliography check:** Scan the works cited for overlap with your source list.

**Verdict signals:**
- Read: abstract directly addresses your question; introduction sets up a problem you care about; conclusion reports findings you need to engage with
- Monitor: article addresses a related question; findings may be cited as context
- Discard: article is on the same general topic but a clearly different question with no intersection

---

## Protocol C: Online Sources

**First: identify what kind of source this is.** Many online sources resemble print articles or books — apply Protocol A or B after this initial classification step.

1. **Article-resembling source** (structured with introduction, sections, conclusion, citations):
   - Apply Protocol B (article protocol)
   - Also run a keyword search on the page or within the site

2. **Website or web resource** (no clear article structure):
   - Look for sections labeled: "introduction," "overview," "summary," "about this site," or similar
   - Check for a site map or index; scan it for your keywords
   - Use the site's search function if available; search for your primary keywords
   - Read the "About" page to understand the sponsor's purpose and credentials

3. **No structure visible and no search:** This is a red flag for both relevance assessment and reliability. Note it as a reliability concern (see Reliability Rubric).

**Special notes for online sources:**
- Treat online sources as tertiary or primary by default; prove secondary status before relying on them as scholarly sources
- For secondary-source online materials (preprints, institutional reports, published PDFs), apply the article protocol
- For primary sources online (historical documents, datasets, fan forums as primary data): apply Protocol A or B as appropriate to the format, not the medium

**Verdict signals:**
- Read: clearly scholarly or authoritative; directly addresses your question; has attribution, date, and sponsor
- Monitor: addresses your topic but unclear authority; may be useful for orientation
- Discard: no attribution; no date; sensationalist or advocacy tone without evidence; clearly not relevant

---

## When to Stop Skimming

Stop and commit to **Read** the moment you find a passage that directly addresses your research question. Do not skim further — you have confirmed relevance.

Stop and commit to **Discard** the moment you have completed all steps in the protocol and found no direct relevance. Do not second-guess yourself — you can always return to a Discarded source if later reading suggests you missed something.

FILE:references/reliability-rubric.md
# Reliability Rubric

A structured checklist for assessing the reliability of a source before committing to full reading. These criteria are warning signals, not guarantees. A source that passes all criteria can still have weak arguments or thin evidence. A source that fails one criterion may still be useful if its other qualities are strong. Use judgment.

---

## Core 5 Criteria (All Researchers)

### Criterion 1: Publication Venue

**Question:** Is the source published or posted by a reputable press or organization?

**What to look for:**
- University presses (Oxford, Cambridge, Chicago, MIT, etc.) are reliable for most purposes
- Established commercial publishers known in the field (Norton for literature, West for law, Ablex for sciences) are generally reliable
- Peer-reviewed journals (any field) — reliability depends partly on the journal's reputation
- Government agencies, major research institutions, established professional associations

**Red flags:**
- Self-published books or articles with no visible publisher
- Sensationalist claims on the cover or in the abstract that are out of proportion to the evidence
- Publisher or sponsor driven by ideology on contested social issues (gun control, climate, abortion) — check who funds the press or organization
- No publisher listed at all

**Note:** Libraries include materials from all kinds of publishers for the sake of coverage. Presence in a library catalog does not indicate reliability.

---

### Criterion 2: Peer Review

**Question:** Was the work peer-reviewed before publication?

**What to look for:**
- "Peer reviewed" designation in the database or journal description
- Journal articles in established academic journals (most are peer-reviewed)
- Books from university presses (typically reviewed by field experts before acceptance)

**What does not constitute peer review:**
- Editorial review alone (most commercial magazines, newspapers)
- Review only by the named editor of an essay collection (not the same as peer review)
- Preprints without clear peer-review notice
- Blogs and personal websites

**If you cannot determine peer-review status:** Treat the source as unverified and increase scrutiny on other criteria.

---

### Criterion 3: Author Credentials

**Question:** Is the author a reputable scholar or expert in this specific field?

**What to look for:**
- Academic affiliation listed (institution, department)
- Track record of publication in the field (check citations, CV if accessible)
- Cited by other reputable sources in the same field
- No undisclosed financial ties to parties with a stake in the outcome

**Complications:**
- Established scholars can have axes to grind, especially on politically or commercially contested topics
- Check acknowledgments and funding disclosures: if a pharmaceutical company funded the research, that is relevant context
- On highly contested social issues, even credentialed scholars may be advocates first; require stronger evidence from these sources

**If you are new to the field and cannot assess credentials:** Use citation count as a proxy — frequently cited authors are generally considered significant by the field. Use citation indexing (forward citation search) to check.

---

### Criterion 4: Currency

**Question:** Is the source current enough for this field and this question?

**Field norms:**

| Field | Outer limit (findings) | Notes |
|---|---|---|
| Sciences and technology | ~2 years | Foundational theory may be cited regardless of age |
| Social sciences | ~10 years | Check recent literature for updates or reversals |
| Humanities (literary criticism, history, philosophy) | Decades | Foundational works cited regardless of date; recent editions matter |

**Rules:**
- Assume textbooks are not current — they synthesize existing knowledge but typically lag the cutting edge by years
- Always use the most recent edition of a secondary or tertiary source; researchers revise (and sometimes reverse) views across editions
- For primary sources (novels, letters, historical documents), standard editions are preferred — usually not the most recent reprint
- **Rule of thumb for any field:** Look at the citations in 2-3 recent articles in your target area. The oldest sources cited regularly define the acceptable floor for age

**If you are unsure:** Flag the source as possibly outdated and continue; read for content and let the argument determine whether the age matters.

---

### Criterion 5: Scholarly Apparatus

**Question:** Does the source include notes, citations, and bibliography that allow you to verify its claims?

**What to look for:**
- Footnotes or endnotes with specific citations
- A bibliography or works cited list
- For websites: clear attribution (author name), date of publication or last update, named sponsor or organization

**Red flags:**
- A book-length argument with no bibliography or citations: you have no way to follow up on anything it claims
- A website with no stated author, no date, and no organizational sponsor: you cannot assess the source of its claims
- A website that has not been updated recently may be abandoned and no longer endorsed by its sponsor

---

## Advanced Criteria (Experienced Researchers or High-Stakes Projects)

### Criterion 6: Published Reviews

**Question:** Has the source been reviewed in field-specific review journals?

Indexes to published reviews exist for most fields. A source reviewed favorably by multiple field experts has an additional layer of vetting. A source that has attracted no reviews (for a book expected to be significant) or only negative reviews warrants extra scrutiny.

---

### Criterion 7: Citation Impact

**Question:** Has the source been frequently cited by other researchers?

High citation count = high impact factor. Use forward citation indexing (looking up who has cited a source) to assess influence. A source cited repeatedly by other scholars is considered significant by the field — even if it presents a contested view, you need to engage with it.

**Caution:** Citation count is a measure of influence, not necessarily accuracy. A highly cited source may be influential because it made a major claim that many subsequent papers argue against. Read beyond the count.

---

### Criterion 8: Website Topic Approach

**Question (for websites only):** Does the site approach its topic judiciously?

A site that engages in heated advocacy, attacks those who disagree, makes wild or unsubstantiated claims, uses abusive language, or contains frequent spelling and grammatical errors is unlikely to be accepted by careful readers as reliable evidence.

---

## Reliability Verdict

After assessing all applicable criteria, assign one of three verdicts:

| Verdict | Criteria profile |
|---|---|
| **Reliable** | Passes all 5 core criteria, or minor issue on 1 criterion offset by strengths on the others |
| **Use with Caution** | Issues on 2 criteria; use the source but flag the specific concerns in your notes and consider corroborating from a stronger source |
| **Exclude** | Fails 3 or more criteria; or fails simultaneously on peer review AND author credentials — the two criteria most central to scholarly vetting |

**Remember:** Even Reliable sources must be read critically. Reliability indicators lower the probability of errors; they do not eliminate it.

---
name: research-question-formulator
description: Transform a broad topic into a focused, answerable research question with built-in significance using the 3-step sentence-completion formula (topic → direct question → So What?). Use this skill when the user has a subject or area of interest but no specific question, has a vague topic like "climate change" or "leadership" and needs to narrow it to something researchable, says they don't know what their paper is really about, is collecting too many notes without a clear direction, wants to know if their research question is worth asking, needs to test whether their topic is too broad (4-5 words = too broad), has a yes/no question that won't drive analysis, is asking a settled-fact question that doesn't open inquiry, wants to move from aimless reading to targeted evidence gathering, has a draft thesis but lost track of what question it answers, or is stuck at the beginning of a research project and doesn't know where to start — even if they never use the phrase "research question formulation." This skill handles question generation and significance testing; it does NOT write the thesis statement (use a separate skill) or plan the full argument structure.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/research-question-formulator
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [3]
tags: [research-methodology, academic-writing, critical-thinking, research-questions]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "User's topic, subject area, draft notes, or early outline — even a single sentence describing what they are interested in"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user needs a text file or can supply topic verbally"
discovery:
  goal: "Turn a broad subject into a focused, significant research question that directs evidence gathering and gives the paper a reason to exist"
  tasks:
    - "Diagnose whether the current topic is too broad, too narrow, or not a question at all"
    - "Apply the 3-step sentence-completion formula to produce a testable research question"
    - "Run the So What? test to confirm the question has significance beyond personal curiosity"
    - "Identify and correct anti-patterns: settled-fact questions, yes/no questions, dead-end questions"
  audience: "Researchers, students, professionals, and analysts starting any evidence-based writing project"
  triggers:
    - "User has a topic but no question"
    - "User is collecting too much information without direction"
    - "User's draft question is too broad, too narrow, or answerable by a single lookup"
    - "User cannot explain why their question matters to anyone else"
    - "User asks 'what should I research about X?'"
    - "User says their paper is about a subject but cannot state what it argues"
---

# Research Question Formulator

## When to Use

You have a subject area, a general interest, or a draft topic, and need to sharpen it into a research question that:

1. Specifies what you do not yet know
2. Directs exactly which evidence you need
3. Answers the "So what?" test — others beyond you care about the answer

Typical triggers:

- "I want to write about [subject] but don't know what my question is"
- Draft notes are accumulating without a focal point
- A question exists but produces only yes/no or looks-it-up answers
- The topic can be stated in four or five words (a sure sign it is too broad)
- Advisor or peer feedback says "I don't know what you're arguing"

**Preconditions to verify:**

- Does the user have at least a rough subject area? (Even "something about climate change" is enough to start)
- Are there any existing notes, outlines, or a draft that might reveal unstated questions?

**This skill does NOT cover:**

- Writing the full thesis statement or claim (topic + question precede the claim)
- Structuring the full argument or paper outline
- Evaluating source quality or research methodology choices

## Context and Input Gathering

### Required (ask if missing)

- **The topic or subject area:** The raw material. A phrase, a sentence, or a paragraph describing what interests the user.
  -> Check prompt for: any subject noun or phrase the user mentions
  -> If missing, ask: "What subject or area are you exploring? Even a rough phrase like 'social media and teenagers' is enough to start."

- **Audience or field context:** Who will read and judge this research? Determines what counts as significant.
  -> Check prompt for: mentions of course, discipline, journal, professional context, or reader type
  -> If missing, ask: "Who is the intended audience — a course instructor, professional colleagues, a journal, or a general reader?"

### Useful (gather from environment if available)

- **Existing notes or draft:** May contain an implicit question the user has not yet named explicitly.
  -> Look for files like `notes.md`, `outline.md`, `draft.md` in the working directory
  -> If found, read them before proceeding — the actual question may already be buried in the material

- **Prior question attempts:** If the user has written a question already, evaluate it rather than starting from scratch.

## Process

### Step 1 — Diagnose the Current Topic

**WHY:** Most researchers begin with a subject, not a question. A subject is static (it can just *be*); only a question creates the obligation to gather evidence and answer something. Diagnosing the topic first prevents building a bad question on a bad foundation.

Check the topic against three diagnostic tests:

**Test A — The Four-Word Test**
If the topic can be stated in four or five words, it is almost certainly too broad to research manageably:

| Too broad (static) | Narrowed (action words added) |
|--------------------|-------------------------------|
| Free will in Tolstoy | The conflict between free will and inevitability in Tolstoy's depiction of three battles in *War and Peace* |
| History of commercial aviation | The military's contribution to developing the DC-3 in the early years of commercial aviation |
| Social media and teenagers | How Instagram's design features affect reported anxiety levels in teenage girls aged 13-17 |

The narrowing technique: add nouns derived from action verbs — *conflict*, *contribution*, *effect*, *development*, *influence*, *change*. These imply something is happening, not just existing.

**Caution:** Do not over-narrow. A topic so specific that almost no sources exist (e.g., "the decision to lengthen the wingtips on the DC-3 prototype for military use as a cargo carrier") is also unworkable.

**Test B — The Sentence Test**
Restate the topic as a full sentence. If the result is trivially true, the topic is still a subject, not a research direction:

- "Free will in Tolstoy" → "There is free will in Tolstoy's novels." (trivially true, no inquiry needed)
- "The conflict between free will and inevitability in Tolstoy's depiction of three battles" → "In *War and Peace*, Tolstoy describes three battles in which free will and inevitability conflict." (a claim worth investigating)

**Test C — Question Quality**
Identify which type of problem the current question (if any) represents:

| Problem type | Example | Why it fails |
|---|---|---|
| Settled fact | "Do the Inuit use masks in ceremonies?" | A single source answers it; no analysis needed |
| Dead end | "How many black cats slept in the Alamo the night before the battle?" | Even a correct answer leads nowhere significant |
| Pure speculation | "Would church attendance increase if congregants wore masks?" | No evidence could settle it |
| Yes/No | "Did Lincoln believe in predestination?" | Closes off inquiry the moment you find one source |

Good questions ask *how* and *why* — they invite deeper analysis than *who*, *what*, *when*, or *where*.

### Step 2 — Generate a Question Inventory

**WHY:** Before selecting one question, generating many opens research directions you would otherwise miss. The goal is not to answer yet — it is to map the possibility space so the best question can emerge. Researchers who jump to a single question too early often get stuck when that question dead-ends.

Using the focused topic from Step 1, generate questions across six heuristic categories. Apply each to the user's topic (the masks example from the book illustrates each category):

**History questions**
- How did [topic] come into being? Why? What came before it?
- How has [topic] changed over time? What caused those changes?

**Structure and composition questions**
- How does [topic] fit into a larger system or context?
- How do the parts of [topic] work together? Which parts are most significant?

**Category questions**
- What kinds of [topic] are there? How are they classified?
- How does [topic] compare to or contrast with related phenomena?

**Negative questions** (flip any positive question)
- Why has [topic] *not* changed in a particular way?
- Why does [topic] *not* appear in contexts where you would expect it?

**Speculative "what if" questions**
- What would be different if [topic] never existed, disappeared, or moved to a new context?
- What analogies connect [topic] to other domains?

**Source-suggested questions** (use after initial reading)
- What claim from a source could be extended or tested in a new context?
- Where do sources disagree, and which side has better evidence?
- What questions do sources explicitly leave open at the end?

Record all questions without filtering. The inventory is productive even if most questions are ultimately set aside.

### Step 3 — Select and Elevate the Best Question

**WHY:** Not all questions deserve research. Selecting based on significance — not just personal interest — is what separates research that contributes to a field from research that only satisfies personal curiosity.

**Evaluate each candidate question against these criteria:**

- Would the answer make a reader think about the topic in a new way? (If no: likely a settled fact or dead end)
- Can evidence plausibly exist that could settle it? (If no: purely speculative)
- Does the answer connect to something larger than the question itself? (If no: dead end)

Prefer questions that ask *how* or *why* over those that ask *who*, *what*, *when*, or *where* — the latter are more likely to produce catalog-style reports rather than analysis.

Once you have two or three promising candidate questions, consider combining them into a single broader question that encompasses them all. Multiple related questions often point to one larger organizing question.

Example (from the book): Three Alamo questions — *How have politicians used the story? How have storytellers' motives changed? Whose purposes does each story serve?* — all combine into: *How and why have users of the Alamo story given the event a mythic quality?*

### Step 4 — Apply the 3-Step Significance Formula

**WHY:** A question is not ready for research until you can articulate why someone beyond you cares about its answer. The 3-step formula forces this articulation before you invest significant time gathering evidence. Researchers who skip this step often finish a project and then cannot explain its contribution.

Complete all three parts of this sentence:

```
I am studying/working on [TOPIC]              <- Step 1: Name your topic
because I want to find out [QUESTION]         <- Step 2: State what you don't know
in order to help my reader understand [SO WHAT?]  <- Step 3: Articulate significance
```

**Step 1 — Name the topic** (use action-derived nouns):
```
I am studying the causes of the disappearance of large North American mammals...
I am working on Lincoln's beliefs about predestination and their influence on his reasoning...
```

**Step 2 — Add the direct question** (what you do not know):
```
...because I want to find out whether the earliest peoples hunted them to extinction...
...because I want to find out how his belief in destiny influenced his understanding of the causes of the Civil War...
```

**Step 3 — Answer So What?** (why readers beyond you should care):
```
...in order to help my reader understand whether native peoples lived in harmony with nature or helped destroy it.
...in order to help my reader understand how his religious beliefs may have influenced his military decisions.
```

The third step is the hardest and the most important. It is your ticket into the conversation of your research community. Advanced researchers are judged not by whether their question interests them, but by whether the answer to Step 3 matters to their field.

**If you cannot complete Step 3:** The question may be valid but you are not yet far enough into the topic to see its significance. Write "TBD — research in progress" and return to it after initial reading. Do not abandon the question; begin research, then revisit.

**Peer feedback prompt:** Ask a colleague, roommate, or friend to listen to your three-step sentence. If they respond "So what?" after Step 2, you have not yet answered Step 3 convincingly.

### Step 5 — Deliver the Formulated Question

**WHY:** The user needs three distinct outputs — not one blended answer — because topic, question, and significance serve different downstream functions. The topic goes into the introduction; the question guides source selection; the significance statement becomes the motivation section. Separating them explicitly prevents the common error of conflating topic with question or question with thesis.

Present the output in three parts:

1. **The focused topic** (narrowed from the original subject using action-derived nouns)
2. **The research question** (direct question form, asking how or why, not answerable by a single lookup)
3. **The significance statement** (Step 3 of the formula — why the answer matters to a reader)

Optionally include:
- The full 3-step sentence in one block for easy reference
- A brief note on what type of evidence would be needed to answer the question
- 1-2 anti-pattern warnings if the original topic exhibited them (e.g., "Your original topic 'Social media and teens' was too broad — the action-derived noun 'effect' is what focused it")

## Examples

### Example 1 — Undergraduate research paper

**Input topic:** "Climate change and agriculture"

**Step 1 diagnosis:** Four-word test fails (too broad). Sentence test: "Climate change affects agriculture" — trivially true.

**Narrowed topic:** The effect of shifting precipitation patterns on smallholder wheat yields in sub-Saharan Africa, 2000-2020.

**Research question:** How have changes in seasonal rainfall timing altered crop yields for smallholder wheat farmers in sub-Saharan Africa over the past two decades?

**3-step formula:**
```
I am studying the effect of shifting precipitation patterns on smallholder wheat yields in sub-Saharan Africa
because I want to find out whether yield losses are driven more by total rainfall reduction or by changes in seasonal timing
in order to help my reader understand whether climate adaptation programs should prioritize irrigation infrastructure or planting calendar adjustments.
```

---

### Example 2 — Graduate thesis (from the book)

**Input topic:** "Lincoln and religion"

**Step 1 diagnosis:** Can be stated in three words. Sentence: "Lincoln was religious" — settled fact, no inquiry.

**Narrowed topic:** Lincoln's beliefs about predestination and their influence on his Civil War reasoning.

**Research question:** How did Lincoln's belief in predestination and divine destiny influence his interpretation of the causes and meaning of the Civil War?

**3-step formula:**
```
I am working on Lincoln's beliefs about predestination and their influence on his reasoning
because I want to find out how his belief in destiny and God's will influenced his understanding of the causes of the Civil War
in order to help my reader understand how his religious beliefs may have influenced his military decisions.
```

---

### Example 3 — Professional report context

**Input topic:** "Remote work productivity"

**Step 1 diagnosis:** Yes/no question risk: "Is remote work more productive?" — answerable by a single meta-analysis citation.

**Narrowed topic:** The relationship between remote work policy design (flexibility vs. structure) and knowledge worker output quality in software teams.

**Research question:** How does the degree of schedule flexibility in remote work policies affect code quality and defect rates in software development teams?

**3-step formula:**
```
I am studying the relationship between remote work policy design and software team output quality
because I want to find out whether higher schedule flexibility is associated with better or worse code quality metrics
in order to help my reader understand which remote work policy features organizations should prioritize to maintain engineering quality.
```

---

## Anti-Pattern Quick Reference

| Anti-pattern | Diagnosis signal | Fix |
|---|---|---|
| Topic too broad | Stateable in 4-5 words; sentence version is trivially true | Add action-derived nouns: *effect*, *conflict*, *development*, *influence* |
| Settled-fact question | A single source can answer it | Reframe as *how* or *why* it happened, not *whether* it happened |
| Yes/No question | Question begins with "Does," "Is," "Did," "Are" | Replace with "How," "Why," "In what ways," "To what extent" |
| Dead-end question | Correct answer doesn't connect to anything larger | Add the Step 3 significance test — if you can't complete it, the question is a dead end |
| Pure speculation | No evidence could plausibly settle it | Ask: can I imagine finding data, documents, or observations that would answer this? If no, drop it |
| Over-narrowed | No sources exist; question is hyper-specific trivia | Broaden by one level; the topic should be researchable by a non-specialist |

## References

- See `references/question-inventory-heuristics.md` for the full six-category question generation framework with worked examples across multiple disciplines
- See `references/3-step-formula-templates.md` for fill-in-the-blank templates and discipline-specific variants of the significance formula

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/3-step-formula-templates.md
# 3-Step Significance Formula: Templates and Variants

Source: The Craft of Research, 4th ed., Chapter 3 (section 3.4)

The 3-step sentence-completion formula is the core mechanism for moving from a topic to a research question with built-in significance. This reference provides fill-in-the-blank templates, discipline-specific variants, and guidance on the most common failure mode (inability to complete Step 3).

---

## The Core Formula

```
Step 1: I am studying/working on/researching ___[TOPIC]___
Step 2: because I want to find out ___[who/what/when/where/whether/why/how + QUESTION]___
Step 3: in order to help my reader understand ___[SIGNIFICANCE — the larger issue the answer illuminates]___
```

The formula works because each step performs a distinct function:

| Step | Function | What it forces you to articulate |
|---|---|---|
| Step 1 | Names the topic with action-derived nouns | That something is happening or changing, not just existing |
| Step 2 | States what you do not know | The gap in knowledge your research fills |
| Step 3 | Connects your gap to something larger | Why someone other than you should care about the answer |

---

## Fill-in-the-Blank Templates

### Template A — Causal / Explanatory Research

```
I am studying the [effect / impact / role / influence] of [X] on [Y]
because I want to find out [how / why / whether] [specific causal mechanism or relationship]
in order to help my reader understand [larger principle, policy implication, or theoretical debate this illuminates].
```

**Example:**
```
I am studying the influence of social media algorithm design on political polarization
because I want to find out whether recommendation systems amplify extreme content more than moderate content
in order to help my reader understand whether platform design choices are a primary driver of political fragmentation.
```

---

### Template B — Historical / Developmental Research

```
I am studying the [development / transformation / origins / decline] of [X] during [time period / context]
because I want to find out [how / why] [specific change or event occurred]
in order to help my reader understand [what this reveals about larger historical forces or recurring patterns].
```

**Example (from the book):**
```
I am working on Lincoln's beliefs about predestination and their influence on his reasoning
because I want to find out how his belief in destiny influenced his understanding of the causes of the Civil War
in order to help my reader understand how his religious beliefs may have influenced his military decisions.
```

---

### Template C — Comparative Research

```
I am studying the [differences / similarities / contrast] between [X] and [Y] in [context]
because I want to find out [which factors explain the divergence / convergence]
in order to help my reader understand [what this comparison reveals about the underlying principles governing both cases].
```

**Example:**
```
I am studying the differences in antibiotic prescribing rates between primary care physicians and urgent care clinics
because I want to find out whether the institutional context predicts prescribing behavior independently of patient diagnosis
in order to help my reader understand whether reducing antibiotic resistance requires changing physician training or changing institutional incentive structures.
```

---

### Template D — Evaluative / Critical Research

```
I am studying [claims / arguments / practices] related to [X]
because I want to find out whether [specific claim or assumption] holds under [specific conditions or scrutiny]
in order to help my reader understand [what the evidence actually supports vs. what is assumed or oversimplified].
```

**Example:**
```
I am studying claims about the productivity benefits of open-plan offices
because I want to find out whether the research supporting those claims controlled for task type and collaboration frequency
in order to help my reader understand whether open-plan design is effective generally or only for specific kinds of work.
```

---

### Template E — Interpretive / Humanistic Research

```
I am studying [texts / artifacts / cultural practices] related to [X]
because I want to find out how [specific meaning, function, or significance] operates in [specific context]
in order to help my reader understand [what this reveals about the values, assumptions, or conflicts of the community that produced it].
```

**Example (adapted from the book):**
```
I am studying the political uses of the Alamo story in Texas public life
because I want to find out how and why storytellers have given the battle a mythic quality
in order to help my reader understand how communities construct and maintain founding myths that serve contemporary political purposes.
```

---

## Discipline-Specific Variants for Step 3

The significance statement in Step 3 should connect to the stakes that your specific research community cares about. Here are common Step 3 framings by field:

| Field | Common Step 3 openings |
|---|---|
| Natural sciences | "...in order to help my reader understand the mechanism by which..." / "...whether current models of X accurately predict..." |
| Social sciences | "...in order to help my reader understand which policy lever is most effective for..." / "...whether structural or individual factors better explain..." |
| Humanities | "...in order to help my reader understand what this reveals about the values of..." / "...how this challenges the received interpretation of..." |
| Engineering / Applied | "...in order to help my reader understand which design approach minimizes..." / "...whether current practice is based on evidence or assumption..." |
| Professional / Management | "...in order to help my reader understand which organizational practice produces better outcomes when..." / "...what the evidence suggests about the tradeoffs between..." |

---

## The Most Common Failure: Unable to Complete Step 3

If you cannot complete Step 3, it means one of three things:

**1. You are too early in the project.**
You have not read enough to see how your question connects to larger debates in the field. This is normal for beginning researchers. Write "TBD — return after initial reading" and proceed. Do not abandon the question; return to Step 3 after you have read three to five sources.

**2. The question is a dead end.**
Your question, even if answerable, leads nowhere. The answer would not change how anyone thinks about anything beyond itself. Symptoms: you answer it and feel done, with nothing further to say; the answer is a list or catalog with no analytical implications.

Fix: Ask "If I knew the answer, what would I do with it? What larger claim would it support or undercut?" If you cannot answer that, the question needs to be replaced, not developed.

**3. The question is about personal curiosity only.**
You are interested in the answer, but the interest is entirely private. This is fine for starting research but not for finishing it. The So What? question asks: "Beyond your own interest, why would others think it a question worth asking?"

Fix: Ask yourself what debate, policy question, recurring argument, or unresolved scholarly problem your answer would speak to. Then rewrite Step 3 to connect explicitly to that debate or problem.

---

## Quick Reference: Step 2 Question Word Guide

The question word in Step 2 determines the type of claim your research will ultimately make:

| Question word | Type of claim produced | Best for |
|---|---|---|
| **whether** | Existence or occurrence claim | Settling a factual dispute; establishing that something happened |
| **how** | Process or mechanism claim | Explaining the way something works or occurred |
| **why** | Causal claim | Explaining the reasons or causes behind something |
| **to what extent** | Degree claim | Measuring or calibrating the strength of a relationship or effect |
| **what [the best / most effective]** | Evaluative claim | Comparing options against explicit criteria |

Avoid question words that lead to catalog answers rather than analytical claims:

| Avoid | Why | Replace with |
|---|---|---|
| *what are all the examples of* | Produces a list, not an argument | *How do these examples collectively reveal...* |
| *does X exist* | Yes/No; settled by a single source | *How and why does X exist in context Y but not Z* |
| *what happened* | Narrative summary, not analysis | *How and why did this sequence of events produce outcome Z* |

---

## Peer Test Script

To test whether your 3-step sentence works, read it aloud to someone unfamiliar with your topic and ask them to respond honestly to each part:

After Step 1: "Do you know what I'm studying?"
After Step 2: "Do you know what question I'm trying to answer?"
After Step 3: "Do you care whether I find the answer? Why?"

If their answer to the third question is "not really," your Step 3 needs revision. Common fixes:
- Make the larger issue more explicit (connect to a debate they have heard of)
- Replace abstract stakes ("understand better") with concrete ones ("determine whether current policy is based on a false assumption")
- Ask them what *they* would need to hear for the answer to matter — then rewrite Step 3 toward that

FILE:references/question-inventory-heuristics.md
# Question Inventory Heuristics

Source: The Craft of Research, 4th ed., Chapter 3 (sections 3.3.1–3.3.6)

These six heuristic categories are designed to generate a broad inventory of possible research questions from any focused topic. Their purpose is not to guide analysis — it is to stimulate questions you would not otherwise think to ask. Use them early, before selecting a single research question.

---

## How to Use This Framework

1. Take your focused topic (already narrowed from a subject using action-derived nouns).
2. Apply each category below, generating 2-4 questions per category.
3. Write down every question without filtering — even silly-seeming questions occasionally yield unexpected insights.
4. After generating the inventory, evaluate questions using the criteria in the main skill (Step 3: Select and Elevate).

---

## Category 1: History Questions

Ask about the developmental context and internal history of your topic.

**Prompts:**
- How did [topic] come into being? What conditions produced it?
- What came before [topic]? What might come after it?
- How has [topic] changed over time? Why did it change?
- What internal history does [topic] have — how have its own components or practices shifted?

**Discipline examples:**

| Field | Topic | History question |
|---|---|---|
| Literature | Free will in Tolstoy | How did Tolstoy's conception of free will shift across his novels from early to late career? |
| Technology | Remote work software tools | How did video conferencing evolve from boardroom technology to consumer product? |
| Medicine | Antibiotic resistance | How did clinical prescribing norms contribute to the emergence of resistant strains? |
| Sociology | Halloween as a holiday | Why has Halloween grown into the second-largest commercial holiday in the United States? |

---

## Category 2: Structure and Composition Questions

Ask how the topic fits into a larger system and how its parts work together.

**Prompts:**
- How does [topic] fit into a larger structure or function within a larger system?
- What roles does [topic] play in different contexts?
- How do the parts of [topic] interact? Which parts are most significant?
- What would happen to the larger system if [topic] were removed?

**Discipline examples:**

| Field | Topic | Structure question |
|---|---|---|
| Organizational behavior | Remote work policies | How do flexibility and accountability components of remote work policies interact to produce different team outcomes? |
| History | The Battle of the Alamo | How did the Alamo story function within the larger mythology of Texas statehood? |
| Biology | Mammal extinction at end of Pleistocene | How did the removal of megafauna affect the ecological structure of North American grasslands? |
| Computer science | Code review practices | Which components of the code review process are most strongly associated with defect reduction? |

---

## Category 3: Category Questions

Ask how the topic is classified, grouped, or compared.

**Prompts:**
- What kinds of [topic] are there? How are they categorized?
- What criteria distinguish one type from another?
- How does [topic] compare to similar phenomena in other contexts?
- In what ways are apparently different instances of [topic] actually the same? Apparently similar instances actually different?

**Discipline examples:**

| Field | Topic | Category question |
|---|---|---|
| Anthropology | Masks | How do Native American ceremonial masks differ functionally from Japanese theatrical masks? |
| Economics | Remote work | How do knowledge worker remote work outcomes differ from manual-task remote work outcomes? |
| Literature | Alamo narratives | What categories of storyteller (politician, novelist, filmmaker, tourist guide) produce structurally different versions of the Alamo myth? |
| Medicine | Antibiotic prescribing | How do prescribing patterns differ between primary care physicians and specialists in treating common respiratory infections? |

---

## Category 4: Negative Questions

Flip any positive question to ask why something does *not* occur.

**Prompts:**
- Why has [topic] *not* changed in a particular way?
- Why does [topic] *not* appear in contexts where you would expect it?
- Why did [phenomenon] *not* spread to [context] even though conditions seemed similar?

**Why negatives matter:** Negative questions often reveal constraints, structural barriers, and competing forces that positive questions miss. The absence of an expected outcome is frequently more informative than its presence.

**Discipline examples:**

| Field | Positive version | Negative version |
|---|---|---|
| Sociology | Why are masks common in African religions? | Why are masks *not* common in Western Christian ceremonies? |
| History | Why did the Alamo story spread? | Why did other battles of similar scale in the Texas Revolution *not* become cultural myths? |
| Technology | Why did remote work adoption grow during COVID-19? | Why did some industries with high digital capability *not* sustain remote work post-pandemic? |

---

## Category 5: Speculative "What If" Questions

Ask what would be different under counterfactual or analogical conditions.

**Prompts:**
- What would be different if [topic] never existed, disappeared, or were placed in a completely new context?
- What if [topic] were applied to a domain where it is not normally found?
- What analogy connects [topic] to a different field — and what does the analogy reveal?

**Important caveat:** Speculative questions must be testable to be usable as research questions. Use them to generate hypotheses and analogies, but always ask: *Can I find evidence that would settle this?* If not, the question is a creative exercise, not a research question.

**Discipline examples:**

| Topic | Speculative question | Testable follow-up |
|---|---|---|
| Halloween masks | What if everyone wore masks in all public social contexts? | How do existing mask-wearing societies (medical, cultural) handle social recognition and identity? |
| Mammal extinction | What if large North American mammals had survived to the present? | What ecological changes followed their extinction that can be measured against analogous ecosystems where large mammals persist? |
| Remote work | What if office work were the exception rather than the norm? | How do societies where remote/distributed work is the historical default (e.g., agricultural or maritime economies) organize collaboration? |

---

## Category 6: Source-Suggested Questions

Use after initial reading on the topic. Ask questions that build on, extend, or challenge what your sources claim.

**Building on agreement:**
- A source makes a persuasive claim. What question extends its reach to a new context?
- A source supports a claim with one type of evidence. Would a different type of evidence support or undercut the same claim?
- A source analyzes [topic] from one disciplinary angle. What would a different disciplinary angle reveal?

**Building on disagreement:**
- Two sources contradict each other. Which has better evidence? What question would settle the dispute?
- A source makes a claim that seems too broad. What counterexample would test its limits?
- A source implies a claim without stating it. What question makes the implied claim explicit?

**Open questions in sources:**
- Many articles and books end with explicit suggestions for further research. These are ready-made research questions waiting for a researcher.
- Look for phrases like: "Further research is needed on...", "We did not examine...", "Future work might address...", "A limitation of this study is..."

---

## Combining Questions into One

After generating the full inventory, look for questions that share a common theme. Multiple related questions often point to a single larger organizing question that encompasses them all.

**Pattern to recognize:**
If three or four questions all ask about the interests, motives, or purposes of people involved with your topic, there is probably one large question about agency and motivation waiting to be articulated.

**Alamo example (from the book):**
- How have politicians used the Alamo story?
- How have storytellers' motives changed over time?
- Whose purposes does each version of the story serve?

Combined: *How and why have users of the Alamo story given the event a mythic quality?*

This combined question gives direction to the entire research project and begins to imagine readers who will judge whether the question is significant.

---
name: research-problem-framer
description: Transform a research question into a fully framed research problem that readers recognize as worth solving, using the condition+consequence structure and the So What? cascade test. Use this skill when the user has a research question but cannot explain why readers should care about the answer, has completed the 3-step significance formula but the Step 3 consequence still feels abstract or weak, is writing an introduction and the reader's "So what?" objection keeps coming back, cannot tell whether their project is pure or applied research and whether that matters for their introduction, wants to verify that solving their conceptual problem actually serves a practical one, has a research question that feels personally interesting but lacks community relevance, is being asked by an advisor or reviewer "why does this matter?", needs to state a research problem in formal proposal or introduction language, wants to understand the difference between a research question (condition) and a research problem (condition + consequence) and why the problem frame is what goes in the introduction, or is building a research argument but keeps getting feedback that readers don't feel the stakes. This skill handles problem framing and consequence articulation; it does NOT formulate the initial research question (use research-question-formulator for that) or write the claim/thesis statement.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/research-problem-framer
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [4]
tags: [research-methodology, academic-writing, research-problems, critical-thinking]
depends-on: [research-question-formulator]
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "User's research question, 3-step formula sentence, or partial introduction describing what they are studying and why"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user needs a text file or can supply context verbally"
discovery:
  goal: "Elevate a research question into a problem statement that readers recognize as worth solving — by naming the condition (what you do not know) and its consequence (what that gap prevents readers from understanding)"
  tasks:
    - "Classify the problem as practical or conceptual and identify the type of gap"
    - "Name the condition (the knowledge gap stated as the research question)"
    - "Identify the consequence (what more important question remains unanswerable until this one is solved)"
    - "Run the So What? cascade to test whether the consequence is specific enough and audience-grounded"
    - "Determine whether the project is pure or applied research and adjust the consequence framing"
    - "Compose a 2-3 sentence problem statement ready for a paper introduction"
  audience: "Researchers, graduate students, professionals, and analysts who have a focused question but need to articulate its community significance"
  triggers:
    - "User has a question but cannot explain why readers should care about the answer"
    - "User's Step 3 significance statement reads as 'interesting to me' rather than 'urgent to my readers'"
    - "Reviewer or advisor asks 'so what?' after hearing the research question"
    - "User cannot distinguish their question (condition) from their problem (condition + consequence)"
    - "Introduction does not create stakes that make readers want to keep reading"
    - "User asks 'is this pure or applied research?' or 'how do I justify this to a practical audience?'"
---

# Research Problem Framer

## When to Use

You have a focused research question (or can formulate one quickly) and need to complete the move from question to problem — the move that tells readers not just what you do not know, but why they should care whether you find out.

A research problem has two parts:

1. **Condition:** A gap in knowledge or understanding (stated as your research question)
2. **Consequence:** What more important question cannot be answered until this gap is closed

Without the consequence, your question is a private curiosity. With it, you enter a community conversation.

Typical triggers:

- "I have a good question but my advisor keeps asking 'so what?'"
- Your introduction describes what you will study but does not create a reason for readers to keep reading
- You finished the 3-step formula and Step 3 still feels thin or forced
- You are not sure whether your consequence is strong enough for your field
- You need to write an introduction that opens with stakes, not just a topic

**Preconditions to verify:**

- Does the user have a focused research question (a how/why question, not a topic)? If not: invoke `research-question-formulator` first, OR ask the user to state what they want to find out in one sentence.
- Does the user know their intended audience? (The consequence must land from the reader's point of view, not the researcher's.)

**This skill does NOT cover:**

- Formulating the research question from scratch (use `research-question-formulator`)
- Writing the full thesis, claim, or argument structure
- Evaluating source quality or selecting a methodology

## Context and Input Gathering

### Required (ask if missing)

- **The research question:** The condition of the problem — what the researcher does not yet know.
  -> Check prompt for: a direct how/why question, or the middle term of the 3-step formula
  -> If missing, ask: "What is your research question — the specific thing you want to find out? Even a rough sentence is enough to start."

- **The intended audience or field:** Who will judge whether the consequence is significant.
  -> Check prompt for: mentions of discipline, course, journal, professional setting, or reader type
  -> If missing, ask: "Who are the readers this research is for — a course, a scholarly field, a professional community, or a general audience?"

### Useful (gather from environment if available)

- **Existing 3-step formula sentence:** May already contain a draft consequence in Step 3 that just needs strengthening.
  -> Look for files like `notes.md`, `outline.md`, `draft-intro.md`, or `proposal.md` in the working directory
  -> If found, extract the Step 3 statement and evaluate it against the So What? cascade before rebuilding from scratch

- **Whether the project has a practical application:** Determines whether to frame the consequence as pure (understanding) or applied (doing).

## Process

### Step 1 — Confirm the Research Question Is Problem-Ready

**WHY:** A research problem cannot be framed until the question is genuinely focused. Too-broad questions produce consequences that are correspondingly vague ("helps readers understand a lot of things"), which fails the So What? cascade. A quick diagnostic here saves a wasted framing cycle.

Check the question against two fast tests:

**Test A — Is it a how/why question?**
Questions beginning with "Does," "Is," "Did," or "Are" invite yes/no answers that close inquiry. Problem framing requires an open question.

| Closes inquiry (not ready) | Opens inquiry (ready) |
|---|---|
| Did the Alamo story influence Texas voters? | How and why did nineteenth-century politicians use Alamo stories to shape public opinion? |
| Is remote work less productive? | How does schedule flexibility in remote work policies affect code quality in software teams? |

**Test B — Is it a single focused question (not a topic)?**
If the question can be answered by listing facts without analysis, it is still a topic masquerading as a question.

If either test fails: invoke `research-question-formulator` first, or work with the user to sharpen the question before continuing.

### Step 2 — Classify the Problem Type

**WHY:** The type of problem determines the type of consequence. Practical problems have tangible costs (money, health, safety, opportunity). Conceptual problems have epistemic consequences — what else we cannot understand because of this gap. Misclassifying the problem type leads to mismatched consequences: attaching a thin practical justification to a conceptual question reads as a stretch and weakens the problem statement.

**Practical problem:** A condition in the world that imposes tangible costs (time, money, safety, respect, opportunity). Solving it requires doing something.

Example:
- Condition: Costs are rising at the Omaha plant.
- Cost: The company is losing money.
- Research problem: What changed? (to find out so management can act)

**Conceptual (research) problem:** A condition of not knowing or not understanding something. Solving it requires answering a question.

Example:
- Condition: We do not know how politicians used Alamo stories to shape public opinion.
- Consequence: We cannot understand how regional self-images influence national politics.

Most academic, journalistic, and professional research projects are conceptual problems. Practical problems motivate the research but the conceptual problem is what the paper actually solves.

**Relationship:** Practical problems motivate conceptual ones. Research answers the conceptual question, which in turn helps solve the practical one. When they are linked, both should appear in the problem statement — the practical problem as motivation, the conceptual problem as what the research actually does.

### Step 3 — Name the Condition

**WHY:** The condition is simply the research question restated as a gap. Making it explicit in the problem statement — rather than leaving it implicit in the question alone — is what lets readers recognize that you are not just curious but investigating something that has not yet been understood. Naming the gap signals that the research is filling something, not just adding to an already-full pile.

Restate the research question as a gap in one of two forms:

**Direct form (question):**
> "How have romantic movies changed in the last fifty years?"

**Indirect form (noun clause, for use in the problem statement):**
> "We do not yet know how romantic movies have changed in the last fifty years."
> or
> "The ways romantic movies have changed in the last fifty years remain poorly understood."

The indirect form integrates more naturally into the introduction prose. Use it when composing the final problem statement.

### Step 4 — Identify the Consequence Using the So What? Cascade

**WHY:** The consequence is what makes the research matter to someone other than the researcher. Without a named consequence, readers have no reason to care whether you answer your question. The So What? cascade is a pressure test: you keep answering "So what if we can't answer this?" until you reach a question that your audience clearly thinks needs answering. If you can answer the cascade in one step, the consequence is probably strong. If you run out of answers before reaching something your audience cares about, the problem needs reconfiguration.

**Run the cascade:**

Start with the condition, then ask: "So what if we can't answer that question?"

> Condition: We do not know how romantic movies have changed in the last fifty years.
> So what? → We can't answer: How have our cultural depictions of romantic love changed?
> So what? → We can't answer: How does our culture shape young people's expectations about marriage and family?

Stop when you reach a question your audience clearly cares about. That is your consequence.

**Rules for a strong consequence:**

1. **The answer to Q1 (condition) must genuinely help answer Q2 (consequence).** The link must be logical, not merely thematic.
2. **Q2 must be more important than Q1 to your readers.** If Q2 is equally obscure or equally unimportant, keep cascading.
3. **The consequence must be stated from the reader's point of view** — it is a cost they pay if the condition is not resolved, not a reward you receive for answering your question.
4. **Stop the cascade before the question becomes so large it swamps the research.** "How does culture work?" is too broad to function as a consequence.

**If the cascade stalls:** You have not yet spent enough time in the field to see what depends on your answer. Provisional remedy: state "The significance of this question will emerge during research" and return to this step after initial reading.

### Step 5 — Determine Pure or Applied Framing

**WHY:** Pure and applied research have the same structure (condition + consequence) but different consequence types. Pure research consequences refer to understanding; applied research consequences refer to doing. Using the wrong consequence type misrepresents the project's goals and can create a mismatch between what the research delivers (knowledge) and what the introduction promises (action). Getting this right also determines whether to include a practical application as a fourth step.

**Test:** Look at your consequence (Step 3 of the 3-step formula). Does it refer to knowing/understanding, or to doing?

**Pure research:** Consequence refers to understanding a larger question.
> "...in order to help readers understand how regional self-images influence national politics."

**Applied research:** Consequence refers to doing something differently.
> "...in order to help astronomers use data from earthbound telescopes to measure electromagnetic radiation more accurately."

**Hybrid (add a fourth step):** If the project is fundamentally conceptual but has a plausible practical application, frame it as pure research and add a fourth step naming the potential practical consequence. The practical application goes in the conclusion, not the introduction.

> Step 3 (conceptual): "...in order to help readers understand how politicians use popular culture to advance political goals."
> Step 4 (practical, in conclusion only): "...so that readers might better recognize and resist manipulation by political narratives."

**Caution:** Do not attach a practical consequence to a conceptual question unless the link is genuine and specific. If readers must take a conceptual leap to see how the knowledge leads to action, the link is a stretch. A thin practical consequence weakens the problem, it does not strengthen it.

### Step 6 — Compose the Problem Statement

**WHY:** The problem statement is the payoff of this entire process — a 2-3 sentence formulation that goes into the introduction and creates the stakes that make readers want to keep reading. It must state both the condition (the gap) and the consequence (what the gap prevents) clearly enough that readers can recognize the problem as one that belongs to them, not just to the researcher.

**Template:**

```
[Condition sentence: name the gap]
[Consequence sentence: name what cannot be understood/done without closing that gap]
[Optional: name the practical application if the project is applied]
```

**Examples:**

Pure research (Alamo):
> We do not yet understand how nineteenth-century politicians used stories of the Alamo to shape public opinion. Without that understanding, we cannot fully explain how regional self-images become instruments of national political persuasion.

Applied research (Hubble telescope):
> We do not know how much the atmosphere distorts electromagnetic measurements taken by earthbound telescopes. Until we do, astronomers cannot use earthbound data to measure the density of electromagnetic radiation with confidence.

Hybrid (four-step):
> We do not understand how politicians used stories of great historical events to shape public opinion in the nineteenth century. Until we do, we cannot understand how popular culture functions as a tool of political goals — a gap that, once closed, may help readers recognize similar moves in contemporary politics.

**Composition rules:**

- Lead with the condition (the gap), not with background or context
- State the consequence before offering any solution
- Use "we do not know," "we do not yet understand," or "the X of Y remains unclear" for the condition
- Use "until we do, we cannot" or "without understanding X, we cannot address Y" for the consequence
- Keep the statement to two or three sentences; longer statements bury the stakes

## Examples

### Example 1 — Undergraduate humanities paper

**Input question:** How have cultural depictions of romantic love in American movies changed since 1970?

**Step 2 — Problem type:** Conceptual. The condition is not knowing how depictions have changed; there is no immediate practical cost, but there is an epistemic consequence.

**Step 3 — Condition:** We do not yet know how romantic movies have changed their depictions of love and relationships over the past fifty years.

**Step 4 — So What? cascade:**
- So what if we don't know this? → We can't understand how Hollywood storytelling shapes cultural expectations about romantic partnerships.
- So what? → We can't understand how mass media influences what young people expect from marriage and long-term relationships.
- *(Audience cares. Stop here.)*

**Step 5 — Pure or applied:** Pure. The consequence is about understanding cultural influence, not about making better movies or fixing relationship expectations. No practical fourth step needed.

**Problem statement:**
> We do not yet know how American romantic movies have changed their depictions of love and partnership over the past fifty years. Without that knowledge, we cannot fully understand how Hollywood storytelling shapes the expectations young people bring to romantic relationships — a question central to cultural psychology and media studies.

---

### Example 2 — Graduate policy research

**Input question (from 3-step formula):** How did the loss of NASA's Hubble instruments change the agency's approach to telescope design and redundancy planning?

**Step 2 — Problem type:** Hybrid. There is a conceptual problem (what we do not know about how institutional learning works after a major failure) linked to a practical one (how NASA and similar agencies should design future systems).

**Step 3 — Condition:** We do not know how the Hubble failure changed NASA's internal design review and redundancy protocols.

**Step 4 — So What? cascade:**
- So what? → We can't understand how large engineering organizations update institutional practices after high-profile failures.
- So what? → We can't explain what conditions cause those updates to stick versus fade within a decade.
- *(Applied consequence within reach: agencies cannot design reliable failure-response protocols.)*

**Step 5 — Framing:** Applied. The consequence refers to doing (designing better protocols), so the project is applied research.

**Problem statement:**
> We do not know how the Hubble Space Telescope's optical failure changed NASA's internal design review and redundancy planning processes. Without that understanding, large engineering agencies lack a documented model for how institutional learning after catastrophic failure translates — or fails to translate — into lasting procedural change.

---

### Example 3 — Professional context

**Input:** "I'm writing a report on why our engineering team's velocity dropped after switching to remote work. My manager wants to know if we should go back to in-office or just change our remote policies."

**Step 2 — Problem type:** Practical problem (velocity drop costs delivery time and morale) → motivates a conceptual problem (we don't know which aspect of remote work caused the drop).

**Step 3 — Condition:** We do not know whether the velocity drop is caused by collaboration friction, tooling gaps, scheduling misalignment, or some combination.

**Step 4 — So What? cascade:**
- So what? → Without knowing the cause, management cannot fix the right variable.
- So what? → The organization risks reverting to in-office work to solve a problem that actually requires policy redesign, not location change.
- *(Manager cares. Stop here.)*

**Step 5 — Framing:** Applied. Consequence refers to doing (making the right policy decision).

**Problem statement:**
> We do not yet know whether our team's velocity decline after the remote transition stems from collaboration friction, scheduling misalignment, tooling inadequacy, or a combination of factors. Without that diagnosis, any intervention — including a return to in-office work — risks fixing the wrong variable and leaving the actual cause unaddressed.

---

## Anti-Pattern Quick Reference

| Anti-pattern | Diagnosis signal | Fix |
|---|---|---|
| Researcher-centric consequence | Step 3 reads "to satisfy my curiosity about X" or "to complete my degree" | Reframe from reader's point of view: what does the reader gain or lose depending on your answer? |
| Artificial practical significance | Practical consequence requires a leap of logic from the conceptual question | Frame as pure research; add practical application as an optional fourth step in the conclusion only |
| Consequence equals condition | "We don't know X because we don't know X" — circular restatement | Run the cascade: ask "So what if we can't answer this?" to find a genuinely distinct, larger question |
| Consequence too large | "Until we answer this, we can't understand how culture works" | Narrow: stop the cascade one step earlier; link to a specific sub-question your audience already cares about |
| Mixed pure/applied framing | Introduction promises practical change but research only delivers understanding | Choose one: either frame as pure + optional practical fourth step, or redesign the project as genuinely applied |
| No audience grounding | Consequence is significant to the researcher but not named from reader's perspective | Imagine the reader asking "So what does that mean for me?" — the answer to that question is the consequence |

## References

- See `references/condition-consequence-structures.md` for worked condition+consequence pairs across disciplines (humanities, social sciences, natural sciences, professional research)
- See `references/pure-vs-applied-framing.md` for the four-step hybrid formula and guidance on when to add a practical application step

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-question-formulator`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/condition-consequence-structures.md
# Condition-Consequence Structures Across Disciplines

Each entry follows the structure:
- **Condition:** What we do not know (the research question restated as a gap)
- **Consequence:** What more important question we cannot answer until the condition is resolved
- **Problem statement:** Ready-to-use 2-sentence formulation

---

## Humanities

### Literary / Cultural Studies
**Condition:** We do not know how nineteenth-century American politicians used stories of the Alamo to shape public opinion.
**Consequence:** We cannot explain how regional self-images become instruments of national political persuasion.
**Problem statement:** We do not yet understand how nineteenth-century politicians used Alamo stories to mobilize public sentiment. Without that understanding, we cannot fully explain how regional cultural myths function as tools of national political influence.

---

### Media Studies
**Condition:** We do not know how American romantic movies have changed their depictions of love and partnership over the past fifty years.
**Consequence:** We cannot understand how Hollywood storytelling shapes the expectations young people bring to romantic relationships.
**Problem statement:** We do not yet know how American romantic movies have changed their depictions of love and partnership since 1970. Without that knowledge, we cannot assess how mass media storytelling shapes the relational expectations young adults bring into long-term partnerships.

---

### History
**Condition:** We do not know how the life experiences of nineteenth-century Russian peasants shaped their performance as soldiers.
**Consequence:** We cannot explain the patterns of battlefield effectiveness and collapse observed in Tsarist armies.
**Problem statement:** The relationship between Russian peasants' peacetime experience and their performance as soldiers remains poorly understood. Until we examine this connection, standard explanations of Tsarist military outcomes will remain structurally incomplete.

---

## Social Sciences

### Political Science
**Condition:** We do not know how a senator's public position on gun control affects vote loss when lobbied by firearms industry groups.
**Consequence:** We cannot model how industry lobbying constrains legislative behavior on high-salience issues.
**Problem statement:** We do not know how constituent opinion on gun control affects the electoral risk a senator faces when resisting industry lobbying. Without that understanding, models of legislative responsiveness to lobbying will underestimate the role of district-level opinion as a constraint.

---

### Psychology
**Condition:** We do not know whether Instagram's visual-comparison features cause or merely correlate with reported anxiety increases in teenage girls.
**Consequence:** We cannot determine whether platform redesign would reduce anxiety or whether the same users would seek equivalent comparison elsewhere.
**Problem statement:** We do not yet know whether Instagram's visual-comparison features cause anxiety increases in teenage girls or simply attract users who are already more anxious. Without a causal account, recommendations for platform regulation will remain unable to predict their actual psychological effects.

---

## Natural Sciences

### Astronomy (pure)
**Condition:** We do not know how many galaxies are in the observable sky.
**Consequence:** We cannot determine whether the universe will expand forever or eventually collapse.
**Problem statement:** We do not yet know the total number of galaxies in the observable universe. Until we have a reliable estimate, cosmological models of the universe's long-term fate — expansion, collapse, or equilibrium — lack a necessary empirical constraint.

### Astronomy (applied)
**Condition:** We do not know how much the atmosphere distorts electromagnetic measurements taken by earthbound telescopes.
**Consequence:** Astronomers cannot use earthbound telescope data to measure electromagnetic radiation density with confidence.
**Problem statement:** We do not know how much atmospheric interference distorts electromagnetic readings from earthbound telescopes. Until we do, astronomers cannot reliably use earthbound instruments to cross-check or substitute for space-based measurements.

---

## Professional / Applied Research

### Organizational Management
**Condition:** We do not know whether the velocity decline in a software team after a remote transition stems from collaboration friction, tooling gaps, or scheduling misalignment.
**Consequence:** Management cannot choose between location-change, policy-change, or tooling-change interventions — and risks fixing the wrong variable.
**Problem statement:** We do not yet know which operational factors are responsible for the team's velocity decline following the remote transition. Without that diagnosis, intervention decisions risk addressing the wrong cause and leaving the underlying problem unresolved.

### Public Health
**Condition:** We do not know whether ozone layer thinning in the high latitudes increases actual skin cancer incidence at ground level, controlling for behavioral adaptation.
**Consequence:** We cannot assess whether the current environmental policy response to ozone depletion is proportionate to the health risk it is meant to address.
**Problem statement:** We do not know whether observed ozone layer thinning has translated into measurable skin cancer incidence increases at affected latitudes. Without that evidence, policymakers cannot assess whether existing protective regulations are matched to actual health risk.

---

## Notes on Construction

**The condition** is always some version of: "We do not know / understand / have evidence for X."
- Use: "We do not yet know..."
- Use: "The X of Y remains poorly understood..."
- Use: "No systematic account exists of how..."
- Avoid: "Researchers have studied X" (that is background, not condition)

**The consequence** is always some version of: "Therefore, we cannot understand / explain / address Y."
- Use: "Without that understanding, we cannot..."
- Use: "Until we resolve this gap, we cannot..."
- Use: "This gap prevents us from..."
- Avoid: "This is interesting because..." (personal curiosity, not community consequence)

**The Q1 → Q2 test:** The answer to Q1 (condition) must genuinely help answer Q2 (consequence). Test it: "If I answered my research question, would that actually advance my readers' ability to answer the consequence question?" If no: the consequence is not the right one.

FILE:references/pure-vs-applied-framing.md
# Pure vs. Applied Research Framing

## The Distinction

**Pure research** addresses a conceptual problem that does not bear directly on any practical situation in the world. Its consequence is understanding: a community of researchers will understand something better.

**Applied research** addresses a conceptual problem whose answer has direct practical consequences. Its consequence is doing: someone will be able to act differently or more effectively because of the research.

The distinction is not about whether the research is valuable or rigorous — many researchers consider pure research more intellectually demanding precisely because it cannot lean on practical urgency to justify its significance.

---

## How to Tell Them Apart

Look at Step 3 of the 3-step formula. Does it end with *knowing* or *doing*?

| Step 3 ends with... | Research type |
|---|---|
| "...understand whether the universe will expand or collapse" | Pure |
| "...understand how regional self-images influence national politics" | Pure |
| "...measure electromagnetic radiation more accurately" | Applied |
| "...choose the right intervention for our team's velocity decline" | Applied |

---

## The Four-Step Hybrid Formula

When a conceptual project has a *plausible* practical application, add a fourth step. Keep Step 3 as the conceptual consequence. Step 4 names the practical consequence — hedged with "might" or "may" to signal that the application is not guaranteed.

```
1. Topic:     I am studying [topic]
2. Question:  because I want to find out [conceptual question]
3. Conceptual significance: in order to help readers understand [conceptual consequence]
4. Practical application: so that readers might [practical action or benefit]
```

### Example (Alamo story research)

```
1. Topic:     I am studying how nineteenth-century versions of the Alamo story differ
2. Question:  because I want to find out how politicians used stories of great events to shape public opinion
3. Conceptual: in order to help readers understand how politicians use popular culture to advance political goals
4. Practical:  so that readers might better protect themselves from unscrupulous political narratives
```

**In the introduction:** Present the project using Steps 1-3 only (as a pure conceptual problem). The practical application is explicitly hedged — it goes in the conclusion, not the introduction.

**Why:** Readers of a conceptual paper will feel misled if the introduction promises political protection advice and the paper delivers historical analysis. Save the practical application for the conclusion, where it functions as one possible implication rather than a core promise.

---

## The Plausibility Test for Applied Research

Before claiming your project is applied, test whether the link from your conceptual question to the practical consequence is genuine.

Ask:
> **(a)** If my readers want to achieve [Step 3 goal], would they think they could do it if they found out [Step 2 answer]?

**Astronomy (applied) — passes the test:**
> (a) If astronomers want to measure electromagnetic radiation density accurately, would they think they could do it better if they knew how much the atmosphere distorts earthbound measurements?
> Answer: Yes. Knowing the distortion allows them to correct for it.

**Alamo history (fails as applied):**
> (a) If readers want to protect themselves from unscrupulous politicians, would they think they could do it by knowing how nineteenth-century politicians used Alamo stories?
> Answer: Probably not. The connection is too distant and too historically specific.

If the answer is "probably not" or "maybe, with a large leap," frame the project as pure and add the practical consequence as a hedged Step 4 in the conclusion only.

---

## Common Framing Errors

### Error 1: Thin practical consequence bolted onto a conceptual question
**Symptom:** Step 3 reads "...in order to help people make better decisions about X" but the research is a historical or theoretical analysis.
**Fix:** Move the practical consequence to Step 4. Frame Step 3 as the genuine conceptual contribution.

### Error 2: Purely conceptual consequence for an applied project
**Symptom:** The research explicitly aims to help a specific community act differently (e.g., improve a software team's process), but the problem statement only mentions understanding.
**Fix:** Rewrite Step 3 with a doing-oriented consequence that reflects the actual purpose of the research.

### Error 3: Overclaiming application scope
**Symptom:** A study of one team in one company claims to help "all software organizations" or "the tech industry" make policy decisions.
**Fix:** Scope the practical consequence to what the evidence can actually support. "This study may help similar teams in comparable contexts" is more honest and more credible than an overreaching universal claim.

### Error 4: Forgetting to hedge the practical application
**Symptom:** Step 4 reads "so that readers can protect themselves" rather than "so that readers might better recognize and resist..."
**Fix:** Add "might," "may," or "could" to all practical applications that are potential downstream uses of the research, not guaranteed outcomes of it.

---

## Quick Decision Tree

```
Does your research answer a conceptual question (what we don't know)?
└── Yes → Does answering it directly enable a specific group to DO something different?
    ├── Yes, and the link is direct and logical → Applied research (3 steps, Step 3 = doing)
    ├── Yes, but the link requires a substantial leap → Hybrid (3 steps conceptual + Step 4 hedged practical)
    └── No, it only improves understanding → Pure research (3 steps, Step 3 = understanding)
```

---
name: research-paper-planner
description: "Build a storyboard-based plan for a research paper and turn it into a first draft. Use this skill when the user has assembled a research argument — a main claim with supporting reasons, evidence, and acknowledgments — and now needs to organize it into a coherent, reader-ready structure before writing. Triggers include: user has a thesis and evidence but does not know how to order the sections; user suspects their draft is organized as a research narrative (what they found first) rather than as an argument (what readers need); user's draft summary-hops across sources without asserting their own claim (patchwork writing); user wants a working introduction sketch to start drafting; user is staring at a blank page and cannot begin; user needs to decide where to state their main point — end of introduction or end of paper; user wants to know whether to order reasons by importance, complexity, familiarity, or contestability. This skill does NOT build the underlying argument from scratch — use research-argument-builder for claim, reason, evidence, acknowledgment, and warrant assembly before using this skill."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/research-paper-planner
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [12]
tags: [research-methodology, academic-writing, paper-planning, drafting, organization]
depends-on: [research-argument-builder]
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "User's assembled argument: main claim (point), supporting reasons, evidence per reason, and any acknowledgments of objections"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user supplies their argument in text form or answers guided questions"
discovery:
  goal: "Translate a research argument into a sequenced storyboard — one page per section — that the user can draft from; detect and correct the three structural anti-patterns before drafting begins"
  tasks:
    - "Verify the argument is assembled (claim + reasons + evidence); redirect to research-argument-builder if not"
    - "Sketch a working introduction: context summary, gap statement, So what? consequence, and tentative main point"
    - "Identify 4-6 key concepts that must run through every section"
    - "Plan the body: background page, one storyboard page per major reason, suitable ordering heuristic"
    - "Plan each section: section point at top, evidence placement, acknowledgments, warrants, and section summaries if needed"
    - "Sketch the conclusion: restate point, articulate significance"
    - "Diagnose the three flawed-plan anti-patterns: Research Journey Narrative, Source Patchwork, Assignment-Prompt Order"
    - "Produce a draft protocol: quick vs. careful drafting guidance, keyword tracking, heading strategy"
  audience: "Researchers, students, analysts, and professionals who have a completed argument and need to build a reader-oriented structure before writing"
  triggers:
    - "User has a claim and reasons but does not know how to sequence sections"
    - "Draft organizes events in the order they occurred to the researcher, not by reader need"
    - "Draft reads as a tour of sources rather than an assertion of the researcher's own view"
    - "User copied assignment-prompt language into the opening paragraph"
    - "User cannot decide whether to state the main point early or late"
    - "User is stuck at the blank page and cannot begin drafting"
---

# Research Paper Planner

## When to Use

You have a research argument — a main claim supported by reasons, with evidence for each reason and responses to likely objections. The argument exists; what you need now is a reader-oriented plan for presenting it.

A plan is not the same as an argument. Your argument captures what you know. Your plan arranges that knowledge into the order that helps *readers* follow, evaluate, and accept it.

This skill guides you through five planning moves, then helps you turn the plan into a draft.

**Preconditions to verify:**

- Does the user have a main claim (point/thesis) and at least two supporting reasons? If not, invoke `research-argument-builder` first, or ask the user to state their claim and reasons before continuing.
- Does the user know their intended audience? Ordering decisions depend on what readers already know and believe.

**This skill does NOT cover:**

- Building the claim, reasons, evidence, or acknowledgments from scratch (use `research-argument-builder`)
- Writing the final polished introduction or conclusion (see chapter 16 of the source)
- Incorporating and citing sources (separate skill)

---

## Context and Input Gathering

### Required (ask if missing)

- **Main claim (point):** The thesis sentence — what the whole paper asserts.
  -> Ask: "In one or two sentences, what is the main claim your paper will defend?"

- **Supporting reasons:** The major reasons why the claim is true — these become sections.
  -> Ask: "What are the two to four main reasons a skeptical reader should accept your claim?"

- **Evidence per reason:** Data, examples, cases, or quotations that support each reason.
  -> Ask: "For each reason, what is your strongest evidence?"

### Useful (gather if present)

- Acknowledgments of objections and responses to them
- The field or discipline (affects ordering conventions and heading norms)
- Deadline and draft length constraints (affects depth of storyboard)

---

## Execution

### Step 1 — Sketch a Working Introduction

**Why:** You need a starting place that orients both you and your readers. A working introduction is not polished writing — it is a map of your opening move. Most writers revise it substantially; that is expected. Writing it now prevents you from drifting during drafting.

The working introduction has four parts, sketched in this order on the first storyboard page:

**1a. Context: summarize only the most relevant prior work**

Write a brief summary (even just bullet notes) of the sources you intend to challenge, modify, or build on. Order them in a way that is useful to readers — chronologically, by significance, by point of view — not in the order you read them and not exhaustively.

> Example: For a paper on concussion risk in youth sports, note the three major epidemiological studies, ordered from earliest population-level finding to most recent biomechanical evidence.

Omit marginal sources. An account of sources that are only loosely related wastes reader attention and dilutes your framing.

**1b. Gap statement: rephrase your question as a flaw or gap in that prior work**

Transform your research question into a statement of what the existing literature fails to explain.

Pattern: "Few of these researchers, however, have explained [your question restated as an absence]."

> Example question: "Why do female athletes under eighteen sustain concussions at nearly twice the rate of male athletes in the same sports?"
> Gap statement: "That research reveals little about the *causes* of this discrepancy."

**1c. Consequence (So what?): sketch why readers cannot afford to ignore this gap**

Even a rough answer is enough to start. Ask: "Until we understand [your gap], we cannot know [more important thing]."

> Example: "Until we understand why female athletes suffer proportionally more concussions, we cannot know the most effective ways to protect them."

If you cannot articulate a consequence at all, pause here. A gap with no consequence is a private curiosity, not a research problem. (If this is the sticking point, consult `research-argument-builder` for the So What? cascade.)

**1d. Main point placement — choose one:**

- **State your claim at the end of your introduction.** This puts readers in charge: they know the problem and the answer, so they can read selectively, challenge you earlier, and remember your argument longer. This is the default preference for most readers and most fields.
- **Reserve your claim for the conclusion.** Use this if building suspense is a disciplinary convention or if your argument only makes sense after the reader has moved through all the evidence. You still need a "launching sentence" at the end of your introduction that previews the key concepts.

> Note: Stating the claim early does not bore readers. If the question is interesting, readers want to see how well you can answer it.

Place your tentative main point at the bottom of the introduction page. You can move it later.

---

### Step 2 — Identify Key Concepts That Will Run Through Your Paper

**Why:** Readers experience a paper as coherent when a small set of concepts recurs across all sections. Without shared key terms, sections feel disconnected even if each section is internally sound.

From your main claim and gap statement, circle four to six words or phrases that name the central concepts your paper depends on.

> Example (immigration and employment paper): employment, job satisfaction, recent Southeast Asian immigrants, cross-cultural, length of residence, prior economic level.

These are not your topic's general vocabulary. They are the specific terms tied to your particular question.

- Write these key concepts at the top of every storyboard page. They are your coherence anchor.
- If you cannot find four to six distinct terms, your claim may be too general. Review your question.
- Check after drafting: does each section use most of these terms? Sections that do not may belong elsewhere or not at all.

---

### Step 3 — Plan the Body

**Why:** The order you assembled your argument follows your research logic. The order your draft follows must serve your readers' cognitive needs. These are rarely the same.

**3a. Sketch a background page (if needed)**

After the introduction page, add a storyboard page for necessary background: definitions of key terms, delimitation of scope, review of a major prior position, historical or social context. Keep it short. Background that exceeds what readers need to follow your argument is delay, not context.

**3b. Create one storyboard page per major reason**

At the top of each page, write the point that the section supports, develops, or explains. This point is almost always one of your supporting reasons. Label it explicitly — "This section argues that..." — even if you delete that label from the final draft.

**3c. Choose a body-ordering heuristic**

When you are unsure how to sequence your reasons, choose from the following eight options. The first two are based on your topic; the remaining six are based on your readers' knowledge and understanding:

| Heuristic | Use when |
|---|---|
| **Part-by-part** | Your topic divides naturally into components; deal with each component in an order that reflects their functional relationships or hierarchy. |
| **Chronological** | Earlier to later, or cause to effect — the simplest option when sequence is intrinsic to the subject. |
| **Short to long, simple to complex** | Readers prefer to settle simple issues before working through complex ones. |
| **More familiar to less familiar** | Readers prefer to read about known territory before venturing into new ideas. |
| **Less contestable to more contestable** | Start with what readers already accept; move toward claims they are more likely to resist. |
| **More important to less important** | Readers prefer important reasons first, but those reasons may have more impact if they come last — weigh both. |
| **Earlier understanding to prepare for later understanding** | Some concepts, definitions, or events must be understood before other sections make sense. |
| **General analysis to specific applications** | Readers need the overall position before they can evaluate how it applies to particular cases, texts, or situations. |

These principles often cooperate. A familiar reason is usually also short and simple. But they can conflict: a crucial reason may be the least familiar. When they conflict, prioritize the principle that most directly serves your readers' ability to follow your core argument.

Mark your chosen principle at the top of each section page with a word that signals the ordering: *First ... Second ... Later ... More important ... A more complex issue ... As a result ...* These markers are for your benefit during drafting; revise or remove them in the final draft.

---

### Step 4 — Plan Each Section and Subsection

**Why:** A section point at the top of a storyboard page is a claim. Like all claims, it needs support. Planning section internals before you draft prevents you from filling a section with whatever material you happened to find, rather than with evidence organized around the section's specific assertion.

For each section page, add:

**4a. Key terms for this section**

From your global key concepts, circle the two or three that uniquely distinguish *this* section from all others. They should appear in the sentence stating the section point. If you cannot find terms that distinguish this section, the section may not contribute independently to your argument.

**4b. Evidence placement**

List the evidence that supports this section's point. If you have multiple evidence items:
- Group items of the same kind together
- Order groups in a way that is logical for readers (not the order you found them)
- Note where you must explain the evidence — its source, reliability, or exact connection to the section's point

**4c. Acknowledgments and responses**

Anticipate the objection most likely to arise in this section. Sketch a response. Responses can be sub-arguments with their own claims, reasons, and evidence.

**4d. Warrants (when needed)**

If readers might not see why a reason supports your claim, add a warrant before the reason. A warrant makes explicit the principle that connects evidence to claim.

> Without warrant: "Most Oxford University students in 1580 signed documents with only their first and last names, so most were commoners."
> With warrant: "In late sixteenth-century England, when someone was not a gentleman but a commoner, he did not add 'Mr.' or 'Esq.' to his signature. Most Oxford University students in 1580 signed documents with only their first and last names, so most must have been commoners."

State the warrant *before* the reason and claim — not after. If readers might question the warrant itself, add a brief argument supporting it.

**4e. Section summaries (long or fact-heavy papers)**

If your paper is long and dense with dates, names, or numbers, end each major section with a one- or two-sentence summary: "This section has established [X]. The argument so far supports [Y]." Cut these from the final draft if they feel clunky; they serve drafting more than reading.

---

### Step 5 — Sketch a Working Conclusion

**Why:** Knowing where you are heading before you start keeps you from discovering your conclusion only after you have written 5,000 words pointing somewhere else.

On the final storyboard page:

1. **Restate your main point.** Not word for word — restate it in light of everything the body has established.
2. **Articulate significance.** Return to the "So what?" question from your introduction. Now that you have argued the point, what does the reader understand about a larger issue that they could not before? What should they do or think differently?

If you still cannot articulate significance, leave a placeholder and return to it after drafting. The act of drafting often surfaces the answer.

---

## Diagnosing the Three Flawed-Plan Anti-Patterns

Before you begin drafting, run these three diagnostic checks. First drafts organized around any of these three patterns require significant reorganization — catching them now costs far less than catching them after drafting.

### Anti-Pattern 1: Research Journey Narrative

**What it is:** The paper follows the order in which you conducted the research — what you found first, the dead ends you hit, the problems you overcame — rather than the order that serves readers.

**Diagnostic signals:** Scan your storyboard for language that refers to *how you did the research* rather than *what the research shows*:
- "The first issue I investigated was..."
- "Then I compared..."
- "Finally, I concluded..."
- Sections labeled by research activity ("Literature Review," "Data Collection") rather than by the argument being made

**Correction:** Reorganize around the elements of your argument: your claim and the reasons that support it. Readers do not need to know how you found the answer; they need to be persuaded by the answer itself.

---

### Anti-Pattern 2: Source Patchwork

**What it is:** The paper strings together quotations, summaries, and paraphrases of sources without asserting your own analysis. Structure follows the sources; your voice is absent.

**Diagnostic signals:**
- Section points at the top of storyboard pages name what a source says rather than what *you* claim
- Key terms running through the paper are the same as those in one or more major sources — you may be mimicking their organization rather than building your own
- Paragraphs begin with a source name rather than your claim
- You could swap two sections without changing the argument because no section has a distinct, necessary contribution

**Correction:** Rewrite each section point in your own analytical language. Your sources are evidence *for* your reasons; they do not *constitute* your reasons. If the key terms in your plan match a source's key terms, ask: am I making my own argument or summarizing theirs?

Note: Patchwork writing is especially common when most research is done online and cutting-and-pasting is easy. Readers recognize it, and it invites charges of plagiarism.

---

### Anti-Pattern 3: Assignment-Prompt Order

**What it is:** The paper's structure echoes the language or sequence of the assignment prompt, substituting the instructor's framing for your own argument.

**Diagnostic signals:**
- Opening paragraph contains phrases lifted from the assignment
- Sections follow the order of items listed in the prompt
- The paper compares Subject A and Subject B in two sequential halves (one half on A, one half on B) — a structure imposed by the assignment's compare-and-contrast framing rather than by conceptual logic

> Example: An assignment asks you to "compare Freud and Jung on the imagination and the unconscious." The flawed response organizes into Part 1 (Freud) and Part 2 (Jung), producing two disconnected summaries. The corrected response breaks the topic into conceptual parts — definitions of the unconscious, mechanisms of imagination, therapeutic implications — and orders those parts by reader need.

**Correction:** Break the prompt's topics into their conceptual elements. Order those elements in the way that best serves your readers' understanding of your argument.

---

## Turning the Storyboard into a Draft

### Draft in the mode that works for you

- **Quick drafting:** Let words flow. Skip quotations and data you will plug in later. When you cannot remember a detail, insert `[?]` and keep writing. Do not stop to perfect a sentence. Start early enough to leave time for revision.
- **Careful drafting:** Get each sentence right before writing the next. Requires a detailed storyboard or outline. If this is your mode, invest more time in Steps 3 and 4 above before you begin.

Most writers draft quickly and revise carefully. But either mode works — the important thing is that you draft *now*, not after conditions are perfect. The perfect paper has never been written and never will be.

### Use keywords to stay on track

Keep your four to six key concepts visible as you draft. Check periodically: does each section use most of them? If a section wanders, return to the key terms and the section point at the top of its storyboard page.

If you find yourself following a trail that diverges from your plan, follow it for a while — you may be discovering something. But return to the storyboard afterward and decide whether the discovery belongs in this paper.

### Use headings during drafting

Even if your field does not use headings in the final paper, use them while drafting. Headings built from section-specific key terms (e.g., "Sam Houston as a Hero in Newspapers Outside of Texas") show the structure at a glance and keep each section anchored to its point. Delete them before final submission if convention requires.

### On procrastination and writer's block

If you cannot start at all, the cause is usually one of three things:

- **Intimidation by the whole:** The task feels too large to begin. Solution: divide the storyboard into the smallest possible steps — one page, one section point. Work only on that step.
- **Goals set too high:** You are trying to draft a polished sentence when a rough one will do. Solution: lower the bar explicitly. Tell yourself you are writing notes to think with, not sentences for readers. Every researcher compromises perfection to get the job done.
- **Perfectionism sentence by sentence:** You cannot move to the next sentence until the current one is perfect. Solution: write informally, remind yourself the draft is for your benefit, and accept that revision is where polishing happens.

If you are stuck but have time, set the draft aside for a day or two. Some writer's block is the subconscious integrating material. Return to find the path clearer.

---

## Worked Examples

### Example A: Sports Science Paper

**Claim:** The higher concussion rate among female youth athletes is due primarily to differences in protective equipment standards and monitoring protocols, not to physiological factors alone.

**Working introduction sketch:**
- Context: Three epidemiological studies (ordered chronologically) establishing the rate disparity
- Gap: "That research reveals little about the causes of this disparity"
- Consequence: "Until we understand the causes, we cannot design effective protective interventions"
- Main point: stated at end of introduction

**Key concepts:** concussion rate, female youth athletes, protective equipment, monitoring protocols, physiological factors

**Body ordering chosen:** Less contestable to more contestable — the equipment-standards argument is more data-supported; the monitoring-protocols argument is more contested among practitioners.

**Conclusion sketch:** Restate claim. Significance: policy recommendations for equipment standards and mandatory monitoring protocols in youth sports governing bodies.

---

### Example B: History Seminar Paper

**Claim:** The Alamo story became a national legend not because of its military significance but because of how Texas boosters and Eastern journalists co-constructed it in print media between 1840 and 1880.

**Working introduction sketch:**
- Context: Historians who have documented the Alamo narrative (ordered by significance to the argument)
- Gap: "Few of these historians, however, have explained *why* the story became so important in national mythology"
- Consequence: "Until we understand how such stories become national legends, we cannot understand what values a culture selects to represent itself"
- Main point: reserved for conclusion (disciplinary convention in narrative history)

**Body ordering chosen:** Chronological (1840s newspaper coverage → 1860s dime novels → 1880s monumental histories) because sequence is intrinsic to how the narrative evolved.

---

## Quick Reference: Storyboard Page Structure

```
PAGE: [Section name]
Key concepts active here: [2-3 from your global list]
Section point: [One sentence asserting what this section claims]
Ordering signal: [First / More important / Earlier understanding needed / etc.]

Evidence:
  - [Item 1 — note source and what it proves]
  - [Item 2 — note source and what it proves]
  (group by kind; order by reader logic)

Warrant needed? [Yes/No — if yes, state it before reason]
Acknowledgment: [Likely objection]
  Response: [Sub-claim + reason + evidence]
Section summary (if long paper): [One sentence — what has this section established?]
```

For deeper detail on the final polished introduction and conclusion structure, see references/working-introduction-conclusion.md.

---

## References

- `references/working-introduction-conclusion.md` — Full structure for final (not working) introductions and conclusions
- `references/body-ordering-heuristics.md` — Extended guidance on each of the eight ordering principles with field-specific examples
- `references/storyboard-anti-patterns.md` — Diagnostic checklist and correction exercises for all three flawed-plan patterns

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-argument-builder`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/body-ordering-heuristics.md
# Body Ordering Heuristics

Source: *The Craft of Research, 4th Edition*, Chapter 12, Section 12.1.3

When you are not sure how to sequence the sections of your paper, choose from the eight heuristics below. The first two are based on the nature of your topic; the remaining six are based on your readers' cognitive state.

---

## Topic-Based Heuristics

### 1. Part-by-Part

Break your topic into its component parts and deal with each in turn. The order of parts should reflect their functional relationships or hierarchical structure — not the order that seemed obvious to you.

Use when: Your subject is a system, an artifact, a theory with distinct components, or a phenomenon that divides naturally into elements (biological systems, software architecture, governmental structures).

Caution: Do not default to this heuristic for compare-and-contrast assignments. A "Part 1 = Subject A, Part 2 = Subject B" structure usually produces two disconnected summaries instead of an integrated argument. Break subjects into their *conceptual* components and organize those.

### 2. Chronological

Arrange sections from earlier to later, or from cause to effect.

Use when: Sequence is intrinsic to your subject — historical development, narrative events, causal chains, experimental procedures.

This is the simplest heuristic and the right choice when the subject itself has an inherent temporal or causal logic. Do not use it merely because you researched the material in a particular order (that is the Research Journey Narrative anti-pattern).

---

## Reader-Based Heuristics

### 3. Short to Long, Simple to Complex

Present the shorter or simpler issues before working toward longer or more complex ones.

Use when: Your reasons vary significantly in depth of treatment required. Settling simpler issues first gives readers the conceptual foundation they need to handle more demanding material.

### 4. More Familiar to Less Familiar

Place sections dealing with concepts readers already know before sections introducing new or unfamiliar ideas.

Use when: Your argument requires readers to extend their existing understanding. Starting with the familiar earns readers' trust and builds the context in which the unfamiliar makes sense.

### 5. Less Contestable to More Contestable

Start with claims readers are most likely to accept and move toward claims they are most likely to resist.

Use when: Your argument contains a spectrum of claims ranging from well-established to genuinely controversial. Establishing agreement early creates goodwill and a shared basis for the contested claims that follow.

### 6. More Important to Less Important (or the reverse)

Place your most important reason first so readers encounter it at maximum attention — or place it last for maximum impact at the moment of closure.

Use when: Your reasons differ significantly in weight. This heuristic has a genuine tension: important first is more efficient; important last is more memorable. If in doubt, place the most important reason last when building toward a conclusion that depends on cumulative force.

### 7. Earlier Understanding to Prepare for Later Understanding

Sequence sections so that each one provides what readers need to follow the next.

Use when: Your argument depends on a chain of concepts, definitions, or events — each section builds on what was established previously, and the sections cannot be reordered without losing comprehension.

This is distinct from chronological ordering. The criterion here is conceptual dependency, not time.

### 8. General Analysis Followed by Specific Applications

Present the overall framework, position, or theory before applying it to specific cases, texts, or situations.

Use when: Your argument involves a theoretical claim applied to multiple examples. Readers need to understand the general model before they can evaluate whether specific instances confirm or complicate it.

---

## When Principles Conflict

These heuristics often cooperate: a familiar reason is usually also simple, and an early-understanding dependency often maps onto a simple-to-complex progression. But they can conflict:

- A crucial reason may be the *least* familiar (important first vs. familiar first)
- A highly contested claim may also be the most important (contestable last vs. important last)

When principles conflict, ask: **which principle most directly serves my readers' ability to follow my core argument?** The answer depends on your field, your audience's prior knowledge, and the specific nature of your evidence.

Whatever order you choose, signal it explicitly at the top of each storyboard page with a transition word: *First... Second... Later... More important... A more complex issue is... As a result...*

These signals are for drafting purposes. Revise or remove them in the final draft as needed by your field's conventions.

FILE:references/storyboard-anti-patterns.md
# Storyboard Anti-Patterns: Diagnostic Checklist and Corrections

Source: *The Craft of Research, 4th Edition*, Chapter 12, Section 12.2

Three structural errors appear with high frequency in first drafts. All three share a common cause: the plan tracks the researcher's activities or the sources' organization rather than the readers' experience. Use the checklists below to diagnose your storyboard before drafting.

---

## Anti-Pattern 1: Research Journey Narrative

### What It Is

The paper tells the story of how you conducted your research — what you found first, the dead ends you hit, the problems you solved — rather than presenting the results organized as a reader-oriented argument.

### Why It Fails

Readers want to know what you found, not how you found it. A research narrative forces readers to slog through the history of your project before reaching your main point, which you have "saved for the end." Experienced readers find this structure condescending and inefficient.

### Diagnostic Checklist

Scan your storyboard page points and any draft text for these signals:

- [ ] Section titles name research activities rather than argument claims ("Literature Review," "Data Collection," "Analysis")
- [ ] Section points contain phrases like: "The first issue I investigated was...", "Then I compared...", "Finally, I concluded..."
- [ ] The main claim appears only on the final page of the storyboard
- [ ] Sections could be reordered without affecting the argument's logical progression
- [ ] The introduction describes what you *did* rather than the *gap* you are addressing

### Correction

Rebuild the storyboard around the elements of your argument:
1. Identify your main claim (point)
2. Identify two to four reasons why the claim is true
3. Make each reason a section point
4. Order sections by reader need, not research chronology

Delete or radically compress any section whose only function is to narrate your research process. Readers need the destination, not the journey.

---

## Anti-Pattern 2: Source Patchwork

### What It Is

The paper strings together quotations, summaries, and paraphrases of sources in a way that substitutes the sources' content for the researcher's own analysis. The researcher's voice is absent or minimal.

### Why It Fails

Readers want *your* analysis, not a tour of other people's ideas. Patchwork writing invites the charge "This is all summary, no analysis." It is also a plagiarism risk, particularly when research is done online and cutting-and-pasting is easy — experienced readers recognize it immediately.

A subtler version: the paper is not literally a patch of quotations, but its organization maps onto the structure of a major source rather than onto the researcher's own argument. The key terms that run through the paper are the same as those in one major source.

### Diagnostic Checklist

- [ ] Section points name what a source says rather than what *you* claim ("Smith argues that...", "According to the report...")
- [ ] Removing any source from a section would eliminate the section's content, not just its evidence
- [ ] The key concepts running through your paper match the key concepts of one or more major sources
- [ ] Paragraphs in the draft begin with source names rather than your own claims
- [ ] Sections could be swapped without changing your argument because each is independently about a different source

### Correction

1. Rewrite every section point in your own analytical language, as a claim about your subject — not about what a source says about it
2. Compare your key concept list against your major sources' key concepts. If they match, ask: "Am I making my argument or imitating their structure?"
3. Demote sources from section organizers to evidence: sources appear inside sections to support your reasons, not to constitute them
4. Apply the test: "If I removed this source, would my argument still exist?" If yes, the source is evidence. If no, you are summarizing, not arguing.

---

## Anti-Pattern 3: Assignment-Prompt Order

### What It Is

The paper's structure mirrors the language or sequence of the assignment prompt, using the instructor's framing as a substitute for the researcher's own argument structure.

### Why It Fails

Echoing the prompt language signals to the reader that you have not developed your own analytical framework — you are executing instructions rather than making an argument. Assignments that list topics in sequence (compare X and Y; discuss A, B, and C) invite structures that produce disconnected summaries rather than integrated arguments.

### Diagnostic Checklist

- [ ] The opening paragraph contains phrases or vocabulary lifted from the assignment wording
- [ ] Sections follow the order in which topics were listed in the prompt
- [ ] A compare-and-contrast assignment produced a two-part structure: Part 1 = Subject A, Part 2 = Subject B
- [ ] Section headings or points use the same terms as the assignment without extending or reframing them
- [ ] The paper reads as if you are checking off assignment requirements rather than defending a position

### Correction

1. Do not open your paper with the prompt's language. Start with context that frames your specific question.
2. Break the prompt's topics into their *conceptual* components. A "compare Freud and Jung" assignment is not asking for a summary of Freud followed by a summary of Jung — it is asking you to identify the analytical dimensions on which comparison is interesting (e.g., definitions of the unconscious, mechanisms of the imagination, therapeutic implications).
3. Order those conceptual components by reader need, not by the sequence the assignment listed them.
4. Develop your own section points, in your own language, organized around your own claim.

---

## Combined Quick Diagnostic

Run this scan on your storyboard before beginning any draft:

| Question | If YES, check for: |
|---|---|
| Do any section points begin with "I found...", "I compared...", "I then..."? | Research Journey Narrative |
| Do any section points begin with a source's name? | Source Patchwork |
| Are my key concept terms the same as a major source's key terms? | Source Patchwork (subtle form) |
| Does my paper open with the assignment's wording? | Assignment-Prompt Order |
| Do my sections follow the order the assignment listed topics? | Assignment-Prompt Order |
| Is my main claim absent from all pages except the last? | Research Journey Narrative |

Two or more YES answers in the same storyboard usually indicate that the plan needs to be rebuilt from the argument up — starting with claim, reasons, and evidence — before drafting begins.

FILE:references/working-introduction-conclusion.md
# Working Introduction and Conclusion Structure

Source: *The Craft of Research, 4th Edition*, Chapter 12, Sections 12.1.1 and 12.1.5

This reference covers the four-part structure of a working introduction and the two-part working conclusion used during storyboard planning. These are planning sketches — not polished drafts. Expect to revise them substantially before final submission.

---

## Working Introduction: Four Parts

A working introduction has four distinct moves. Each serves a different function for your readers.

### Part 1: Context (Summary of Relevant Prior Work)

**Function:** Orient readers within the existing conversation your paper enters.

**What to include:** Brief summaries of only the sources you intend to challenge, modify, or build on. Not a comprehensive literature review — only the most directly relevant prior work.

**Ordering options** (choose based on what serves your readers):
- Chronologically: trace development of the conversation over time
- By significance: most important contributions first
- By position: group sources that share a point of view
- By relevance to your specific question: most directly related first

**What to exclude:** Sources that are only marginally related to your specific question. An account of loosely connected references wastes reader attention and dilutes the stakes of your gap.

**Do not follow** the order in which you read the sources, or the order in which you recorded them in your notes.

### Part 2: Gap Statement (Question Rephrased as a Flaw or Absence)

**Function:** Signal to readers where the existing conversation has left something unresolved — and why your paper addresses that gap.

**Pattern:**
> "Few of these researchers, however, have explained [your question restated as an absence in the literature]."

**Transform your research question into a gap claim:**

| Research question | Gap statement |
|---|---|
| Why is the Alamo story so important in national mythology? | Few of these historians have explained *why* the Alamo story became so important in our national mythology. |
| How does length of residence affect job satisfaction among recent immigrants? | That research reveals little about how length of residence shapes job satisfaction in this population. |
| What caused the shift in Roman monetary policy after 200 CE? | These accounts, however, do not explain what drove the specific shift in monetary policy after 200 CE. |

The gap statement is the pivot between the context (what others have done) and your paper (what you will do).

### Part 3: Consequence Statement (So What?)

**Function:** Tell readers why the gap matters — what they cannot understand or decide until your question is answered.

**Pattern:**
> "Until we understand [your gap], we cannot know [more important question or practical implication]."

**Examples:**

> "Until we understand why female athletes suffer proportionally more concussions, we cannot know the most effective ways to protect them."

> "If we understood how such stories become national legends, we would better understand our national values, perhaps even what makes us distinct."

**If you cannot articulate a consequence at all:** Your gap may not yet be a problem — it is still a question. Use the So What? cascade from `research-argument-builder` to locate the consequence before planning the introduction.

**Rough answers are acceptable:** At the planning stage, an imprecise consequence is better than no consequence. You are writing this for yourself, not yet for readers.

### Part 4: Main Point (Claim / Thesis / Point)

**Function:** State what your paper argues, or launch readers into the body of the argument.

**Two placement options:**

**Option A — State at end of introduction:**
Place your claim as the final sentence of the introduction section. Readers then know the problem and the solution from the start. They can read selectively, challenge you on specific points, and follow the argument with purpose. Most readers prefer this. Stating your claim early does not bore readers who find the question interesting.

**Option B — Reserve for conclusion:**
If your disciplinary convention or argumentative strategy calls for suspense, withhold the explicit claim until the conclusion. You still need a launching sentence at the end of your introduction. This launching sentence names the key concepts that will run through the paper, signaling the territory without revealing the destination.

> Example launching sentence: "The difference lies not in biology but in the institutional structures that govern how female athletes are equipped and monitored."

For planning purposes, write your claim on the introduction page of your storyboard even if you plan Option B — you can move it to the conclusion page later.

**Road maps (optional):**

Some fields (especially social sciences) expect a road map at the end of the introduction:
> "In Part 1, I discuss... Part 2 addresses... Part 3 examines..."

If this is a convention in your field, add it to your storyboard. If it is not expected, consider adding it for drafting purposes only (it helps you stay on track) and cut it before final submission.

---

## Working Conclusion: Two Parts

The conclusion is not a summary of what you have already said. It is a forward-looking move that answers the reader's final question: "So what? Given everything you have established, what does it mean?"

### Part 1: Restate the Main Point

State your main claim again — but in light of everything the body has established. This is not a verbatim repetition. The point sounds different after the body has filled it in with evidence, acknowledged objections, and refined its implications.

Place this restatement at the top of your conclusion storyboard page.

### Part 2: Articulate Significance

Return to the "So what?" question you sketched in the introduction and answer it more fully now that you have established your point.

Questions to answer in this section:
- What can readers now understand about a *larger* issue that they could not before reading this paper?
- What should practitioners, policymakers, scholars, or citizens do or think differently?
- What question does your answer open up that others should now investigate?
- What is the broader implication of this particular finding for the field?

**If you cannot articulate significance:** Leave a placeholder in your storyboard. The act of drafting the body often surfaces the answer. Return to the conclusion after completing your first draft.

---

## Storyboard Structure: Complete Overview

```
PAGE 1: Introduction
  Context: [Brief notes on 2-4 directly relevant sources, ordered for readers]
  Gap: "Few researchers have explained..."
  Consequence: "Until we understand X, we cannot know Y"
  Main point: [Your claim — move to conclusion page if using Option B]
  Road map (optional): "Part 1... Part 2... Part 3..."

PAGE 2: Background (if needed)
  Background needed: [Definitions, scope limits, context, prior position review]
  Section point: [What this background page establishes]
  Keep it short.

PAGES 3-N: Body (one page per major reason)
  Key concepts active here: [2-3 from global list]
  Section point: [The claim this section supports or establishes]
  Ordering signal: [First / More important / A more complex issue / etc.]
  Evidence: [Items grouped by kind, ordered by reader logic]
  Warrant: [If needed — state before reason]
  Acknowledgment + response: [Likely objection and sub-argument response]
  Section summary: [Optional — one sentence for long/fact-heavy papers]

PAGE N+1: Conclusion
  Main point restated: [Claim in light of body's argument]
  Significance: [Larger implication — what readers now understand]
```

---
name: research-introduction-architect
description: Draft a complete research introduction and matching conclusion using the Context→Problem→Response architecture. Use this skill when the user has a framed research problem (condition + consequence) and needs to write or revise the opening and closing sections of a research paper; when an introduction exists but reads as a flat topic announcement instead of a problem-driven argument; when the user cannot decide whether to state the main point in the introduction (point-first) or withhold it for the conclusion (point-last) and needs to understand the trade-offs; when the first sentence of the introduction is a dictionary definition, a grand universal claim ("Throughout history…"), or a repetition of the assignment prompt; when the conclusion merely restates the introduction without adding new significance or calling for further research; when the user needs guidance on how much context to provide — neither too sketchy nor encyclopedic — based on the audience's prior knowledge; when the pacing of an introduction (fast vs. slow context setup) needs to match audience expertise level; or when the user wants a checked draft that correctly omits the context element (problem well-known) or consequence element (widely understood) rather than including them by default. This skill outputs a draft introduction and conclusion. It does NOT frame the research problem from scratch — use research-problem-framer for that.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/research-introduction-architect
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [16]
tags: [research-methodology, academic-writing, introductions, conclusions, argumentation]
depends-on: [research-problem-framer]
execution:
  tier: 1
  mode: full
  inputs:
    - type: document
      description: "User's research problem statement (condition + consequence), main point or thesis, and a description of the intended audience and field"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user provides problem statement and main point verbally or in a file"
discovery:
  goal: "Draft a complete introduction (Context + Problem + Response) and a matching conclusion (main point + new significance + call for research) that compels the audience to read and leaves them with renewed appreciation of the work's significance"
  tasks:
    - "Gather the research problem statement, main point, audience description, and field"
    - "Decide whether to include or omit each element (context, consequence, main point in intro)"
    - "Determine pace: fast (expert audience) or slow (general/introductory audience)"
    - "Select the opening sentence strategy: striking fact, quotation, or anecdote"
    - "Draft the introduction with correct Context→Problem→Response sequencing"
    - "Draft the conclusion as a mirror reversal: main point → new significance → call for research"
    - "Run structural self-check against anti-patterns and omission rules"
  audience: "Researchers, graduate students, and professionals drafting or revising introductions and conclusions for academic papers, reports, or proposals"
  triggers:
    - "User has a research problem but needs to write the introduction"
    - "Introduction exists but reads as topic announcement, not problem argument"
    - "First sentence is a dictionary definition, grand universal claim, or assignment repetition"
    - "User cannot decide whether to place the main point in the intro or the conclusion"
    - "Conclusion merely repeats the introduction without adding significance or new questions"
    - "Introduction pace (context length) does not match audience expertise"
---

# Research Introduction Architect

## When to Use

You have a framed research problem — a condition (what is not known) and a consequence (what that gap prevents) — and a main point or thesis. Now you need to write the introduction and conclusion that will move readers from curiosity to engagement and leave them with a clear sense of why the work matters.

This skill covers the full drafting cycle:

1. **Introduction:** Context → Problem → Response
2. **Conclusion:** Main point → New significance → Call for further research

**Preconditions to verify:**

- Does the user have a problem statement with both a condition and a consequence? If not: invoke `research-problem-framer` first.
- Does the user have at least a draft main point or thesis? If not: ask them to state in one sentence what their research concludes.
- Do you know the intended audience (field, expertise level, general vs. specialist)?

**This skill does NOT cover:**

- Framing the research problem from scratch (use `research-problem-framer`)
- Building the argument structure of the body (use `research-argument-builder`)
- Editing prose clarity (use `prose-clarity-reviser`)

## Context and Input Gathering

### Required (ask if missing)

- **The problem statement:** A condition + consequence in 2-3 sentences. This is the core of the Problem element.
  -> Check for files like `notes.md`, `outline.md`, `draft-intro.md`, or `proposal.md`
  -> If missing, ask: "What is your research problem — the gap in knowledge and why it matters to readers?"

- **The main point or thesis:** The conclusion the research reaches.
  -> If missing, ask: "What does your research conclude? Even a rough one-sentence answer is enough to start."

- **The intended audience:** Determines pace (how much context to build before disrupting it) and whether consequences need to be spelled out.
  -> If missing, ask: "Who are your readers — specialists in a field, a general educated audience, or a specific professional community?"

### Useful (gather from environment if available)

- **Existing draft introduction or outline:** Diagnose which elements are present and which are missing before rewriting.
- **Field or discipline conventions:** Sciences and social sciences typically allow faster context setup and explicit road-maps; humanities introductions often open more slowly and withhold the main point until the conclusion.

## Architecture: The Three-Element Introduction

Every introduction, regardless of field, follows the same underlying grammar:

**Context → Problem → Response**

| Element | Function | When to omit |
|---|---|---|
| Context | Establish common ground — the stable, unproblematic state of knowledge — so that the Problem can disrupt it | Omit when the problem is so well-known that readers already share the disruption |
| Problem | State the condition (gap in knowledge) and consequence (what the gap prevents) | Never omit — this is the engine of the introduction |
| Response | State the main point directly (point-first) or promise a main point to come (point-last) | Never omit entirely — at minimum, provide a launching point |

The conclusion reverses this order: **Response (main point) → new significance → call for further research**.

## Process

### Step 1 — Gather and Confirm the Problem Statement

**WHY:** Everything in the introduction grows from the research problem. A problem statement that lacks a consequence, or that is still framed as a topic rather than a gap, will produce a flat introduction no matter how well the surrounding elements are written. Confirming the problem statement before drafting prevents rewriting the introduction after fixing the problem.

Confirm the problem statement has two parts:

- **Condition sentence:** Names the gap. Uses formulations like "We do not yet know…," "X remains poorly understood…," or "The ways Y has changed remain unclear…"
- **Consequence sentence:** Names what that gap prevents. Uses formulations like "Without that understanding, we cannot…" or "Until we do, readers cannot address…"

If the consequence is missing, run the So What? cascade (see `research-problem-framer`) before continuing.

### Step 2 — Decide the Structural Configuration

**WHY:** Choosing the wrong structural configuration — omitting context the audience needs, or laboriously establishing context the audience already knows — damages credibility. Experts who see two paragraphs of background they already know may conclude the writer underestimates them. Novices who see no context may not understand why the problem is a problem.

Work through three decisions:

**Decision A — Include or omit the Context element?**

Include context when:
- The audience may not immediately recognize why the existing state of knowledge is insufficient
- The problem requires a stable baseline to make the disruption legible

Omit context (open directly with the Problem) when:
- The problem is already widely familiar in the field (e.g., a well-known open question in mathematics or physics)
- Opening with context would be condescending to specialist readers

**Decision B — Spell out consequences or omit them?**

Spell out consequences when:
- The audience works in a field where the problem is new or unfamiliar
- The connection between the condition and the larger gap is not self-evident

Omit consequences (or state them briefly) when:
- The audience knows the field well and would find explicit consequence-statements redundant
- The consequence is genuinely obvious once the condition is named

**Decision C — Point-first or point-last?**

| Choice | What it does | When to use |
|---|---|---|
| **Point-first** (state main point in intro) | Reader knows the destination from the start; controls their own reading | Sciences, social sciences, professional reports; readers who skim for answers |
| **Point-last** (launch point in intro, main point in conclusion) | Writer controls revelation; suspense builds through the argument | Humanities, complex arguments where the conclusion requires the full argument to be persuasive |

The minimum for point-last is a **launching point**: a concrete plan or outline of the argument, not just "This paper will discuss X." Vague topic announcements ("This study investigates processes leading to ozone depletion") give readers no reason to trust the journey.

### Step 3 — Determine the Pace

**WHY:** Pace signals how much the writer assumes the reader already knows. A fast opening (one or two sentences of context before the disruption) implies a peer audience. A slow opening (extended context built over several sentences or paragraphs) implies a less expert audience. Mismatched pace is a credibility signal: too slow for experts reads as patronizing; too fast for novices reads as inconsiderate.

**Fast pace (expert audience):**
- One sentence naming the established consensus or current practice
- Immediately disrupt with the condition of the problem
- State the consequence in one sentence if needed

**Slow pace (general or introductory audience):**
- Multiple sentences building the background, defining terms, and establishing the stakes
- Disrupt only after the reader has enough grounding to feel the disruption

**Diagnostic question:** If you read the first sentence to someone in your target audience, would they immediately recognize it as describing their world? If yes: fast pace is safe. If no: build more context first.

### Step 4 — Select the Opening Sentence Strategy

**WHY:** The first sentence establishes tone, signals audience, and creates the first impression of the writer's authority. Three patterns work; three common substitutes do not.

**Patterns that work:**

1. **Striking fact relevant to the problem**
   > "Those who think that tax cuts for the rich stimulate the economy should contemplate the fact that the top 1 percent of Americans control one-third of America's total wealth."
   — Works because the fact is specific, surprising, and directly anticipates the problem

2. **Striking quotation** (only if key terms recur through the introduction)
   > "From the sheer sensuous beauty of a genuine Jan van Eyck there emanates a strange fascination not unlike that which we experience when permitting ourselves to be hypnotized by precious stones." — Edwin Panofsky
   — Works because specific language from the quotation ("fascination," "strange") drives the rest of the introduction

3. **Relevant anecdote** (only if its language vividly illustrates the problem)
   > "On a park bench in July 1996, Cynthia, Laurie, and other senior officers of the Black Sisters United — Chicago's largest federation of 'girl gangs' — reflected on their efforts…"
   — Works because the scene concretizes the problem before any abstract statement

**Anti-patterns to avoid:**

| Anti-pattern | Why it fails |
|---|---|
| Dictionary definition ("Webster's defines ethics as…") | Signals the writer does not yet know what the word means in their field; too generic for any audience |
| Grand universal claim ("Throughout history, philosophers have…") | Wrong community: tries to speak to all of humanity rather than the specific audience who cares about this problem |
| Assignment repetition ("This paper will analyze…") | Signals the audience is a single teacher, not a research community; treats the paper as a task, not an argument |

### Step 5 — Draft the Introduction

**WHY:** Writing the full draft — rather than an outline — forces every element to be resolved. Gaps that seem small in outline form (a missing consequence, a vague launching point) become visible as structural problems in a full draft.

Assemble the elements in order:

```
[Opening sentence: striking fact, quotation, or anecdote]
[Context: 1-4 sentences establishing the stable, unproblematic state of knowledge]
    — OR omit if problem is well-known —
[Condition: the gap in knowledge, stated as "We do not yet know…" or equivalent]
[Consequence: what the gap prevents, stated as "Without that…, we cannot…"]
    — OR omit if consequence is self-evident to the audience —
[Response: main point stated directly (point-first) OR launching point with plan (point-last)]
```

Target length: 150–350 words for most academic and professional introductions. Longer for book chapters or proposals; shorter for short reports.

### Step 6 — Draft the Conclusion

**WHY:** A conclusion that merely restates the introduction tells readers nothing they do not already know. The conclusion earns its place by doing two things the introduction cannot: it states the main point in full (rather than promising it), and it extends the conversation — naming new significance, implications, or open questions that the research has made visible.

Structure the conclusion as a mirror reversal of the introduction:

```
[Main point, stated more fully than in the introduction — not word-for-word repeated]
[New significance or application: a new answer to "So what?" that was not in the introduction]
    — e.g., an additional implication, a practical consequence, or a broader theoretical connection —
[Call for further research: what questions remain open; what studies would follow naturally]
    — frame as keeping the conversation alive, not admitting defeat —
```

**On the new significance:** State it in the conclusion, not the introduction, because it suggests further questions the paper does not take up. Introducing it in the introduction would create an obligation to answer it in the body.

**On the call for research:** Imagine a reader fascinated by your work who wants to follow up. What would you tell them to investigate? This frames the call as generative, not as an admission of incompleteness.

### Step 7 — Structural Self-Check

**WHY:** Small structural errors in introductions and conclusions — missing consequences, vague launching points, word-for-word repeated conclusions — are easy to miss in the flow of drafting. A brief checklist catches them before the draft is shared.

Run through each item:

- [ ] Does the introduction open with a specific first sentence (fact, quotation, or anecdote) — not a dictionary definition, universal claim, or assignment repetition?
- [ ] Is the Context calibrated to audience expertise (not too thin, not encyclopedic)?
- [ ] Is the Problem stated with both a condition AND a consequence — or is the omission of the consequence justified by audience familiarity?
- [ ] Is the Response either a direct main point (point-first) or a concrete launching point with a plan (point-last) — not merely a vague topic announcement?
- [ ] Does the conclusion state the main point more fully than the introduction — not word-for-word repeated?
- [ ] Does the conclusion add new significance (a new answer to "So what?") not present in the introduction?
- [ ] Does the conclusion call for further research that extends rather than merely closes the argument?

If any item fails, revise before delivering the draft.

## Examples

### Example 1 — Humanities paper (point-last, full context)

**Input:**
- Condition: We do not know why the quasi-persons in American science fiction (androids, half-aliens, cyborgs) are almost invariably depicted as white and male.
- Consequence: Without understanding this pattern, we cannot fully analyze how popular culture defines personhood against non-Western, non-male norms.
- Main point: These figures implicitly define the "model person" by Western, male criteria — a definition that excludes most of humanity.
- Audience: Cultural studies readers, familiar with Star Trek and Frankenstein but not necessarily with critical race theory.
- Pace: Moderate (lay out examples before disrupting).

**Introduction draft:**

> In almost every episode of *Star Trek: The Next Generation*, the android Data wonders what makes a person a person. In the original *Star Trek*, similar questions were raised by the half-Vulcan Mr. Spock, whose status as a person was undermined by his machine-like logic and lack of emotion. Data and Spock are only the most recent "quasi-persons" in American popular fiction who have explored the nature of humanity — a lineage that runs from Frankenstein's monster to the Terminator. But the real question is not whether these characters achieve personhood. It is why these quasi-persons who struggle to be persons are almost always white and male. As cultural interpreters, do they tacitly reinforce destructive stereotypes about what it means to be "normal"? This article argues that the model person in American science fiction is consistently defined by Western, male criteria — a definition that, taken seriously, excludes most of the people in the world.

**Why this works:**
- Opening anecdote (Data wondering) anchors the problem concretely before naming it abstractly
- Context builds two paragraphs of examples, then the "But" disrupts them
- Consequence is embedded in the rhetorical question ("do they reinforce destructive stereotypes?")
- Response states the main point directly (for a short article); an essay version might end with a launching point instead

---

### Example 2 — Professional report (point-first, fast pace, no explicit consequence)

**Input:**
- Condition: Motodyne has no data showing which of its online help icons are self-explanatory to users.
- Consequence: Without such data, the design team cannot determine which icons to redesign.
- Main point: Five of eleven icons tested are not self-explanatory; those five should be priority redesigns.
- Audience: Engineering and UX team at Motodyne — knows the product, unfamiliar with the specific test data.
- Pace: Fast (one sentence of context, then directly to the problem).

**Introduction draft:**

> As part of its Continuous Quality Improvement program, Motodyne plans to redesign the user interface for its Unidyne online help system — specifically, the icons that should be self-explanatory without verbal labels. Motodyne has three years of experience with its current icon set, but it has no data showing which icons users actually understand without labels. Lacking such data, we cannot determine which icons to redesign. This report provides data for eleven icons and shows that five of them are not self-explanatory.

**Why this works:**
- Fast pace: one sentence of organizational context, then immediately to the gap
- Consequence is explicit ("cannot determine which icons to redesign") because the team is not a specialist research audience
- Response states the main point directly — appropriate for a technical report where readers skim for findings

---

### Example 3 — Academic paper with conclusion (legal history)

**Input:**
- Condition: It is unclear whether Supreme Court decisions rejecting mandatory death sentences in civilian cases apply to the military, creating ambiguity in the Uniform Code of Military Justice.
- Consequence: Until resolved, Congress cannot know whether it must revise military law.
- Main point: The Supreme Court's decisions do apply, and Congress must revise the Uniform Code.
- New significance (for conclusion): Revising the Code will challenge a foundational value of military culture — that ultimate betrayal requires ultimate penalty — which will force Congress to adjudicate military values, not just legal rules.
- Audience: Legal scholars and policy readers, familiar with the Uniform Code but not necessarily with the specific Supreme Court cases.

**Introduction draft:**

> In today's society, would Major John André — a British spy captured behind American lines in 1780 — be hanged? Though considered a noble patriot by the British, he suffered the punishment mandated by military law. Over time our traditions of punishment have changed, but the punishment for spying has not: it is the only offense that still mandates death. Recently, however, the Supreme Court has rejected mandatory death sentences in civilian cases, creating an ambiguity in their application to military law. If those decisions apply to the military, Congress may have to revise the Uniform Code of Military Justice. This article concludes that it will.

**Conclusion draft:**

> In light of recent Supreme Court decisions rejecting mandatory capital punishment, the mandatory death penalty for espionage is apparently unconstitutional and must be revised by Congress. More significantly, if the Uniform Code of Military Justice is changed, it will challenge the fundamental value of military culture that ultimate betrayal requires the ultimate penalty. Congress will then have to deal not just with a legal inconsistency, but with the military's deeply held sense of what justice demands — a question that extends well beyond the narrow issue of one penalty's constitutionality. Future research should examine how military culture has historically adapted its core values when external legal constraints required change, and whether those adaptations were perceived as legitimate by the communities they governed.

**Why this works:**
- Conclusion restates the main point ("must be revised") more fully than the introduction's one-sentence prediction
- New significance (the cultural challenge to military values) was not in the introduction — introducing it there would have created an unanswered obligation
- Call for research ("how military culture has historically adapted…") keeps the conversation alive without implying the article failed to answer its own question

## Anti-Pattern Quick Reference

| Anti-pattern | Diagnosis signal | Fix |
|---|---|---|
| Dictionary opening | "Webster's defines X as…" in sentence 1 | Replace with a striking fact, quotation, or anecdote specific to the problem |
| Grand universal opening | "Throughout history…" or "Since the dawn of civilization…" | Narrow to the specific community of readers; open with a fact or example from their field |
| Assignment repetition | "In this paper, I will discuss…" mirrors the prompt language | Rewrite for a reader who has not seen the assignment; treat the introduction as an argument, not a task report |
| Context-only introduction | Background runs for several paragraphs with no disruption | Add the condition (gap) and consequence before the response; the introduction must contain a problem |
| Missing consequence | Condition stated, then immediately jumps to response | Ask So what? after the condition; state what remains unresolvable if the gap is not closed |
| Vague launching point | "This paper will investigate X" with no plan | Replace with a concrete outline of the argument: what the paper evaluates, compares, or demonstrates |
| Word-for-word repeated conclusion | Conclusion uses same sentences as introduction | Restate the main point more fully; add new significance not previewed in the introduction |
| Conclusion as summary only | Conclusion lists findings without adding significance | Add a new answer to "So what?" — an implication, application, or further question the research has opened |

## References

- See `references/introduction-configurations.md` for full worked examples of all structural configurations (omit context, omit consequence, point-first vs. point-last) across three disciplines
- See `references/conclusion-patterns.md` for worked conclusion drafts with and without calls for further research

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-problem-framer`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/conclusion-patterns.md
# Conclusion Patterns: Worked Examples

A conclusion is the mirror reversal of the introduction. Where the introduction moves from stability to disruption to promise, the conclusion moves from resolution to new significance to invitation for further inquiry.

---

## The three-part conclusion structure

```
[Main point — stated more fully than in the introduction, not word-for-word repeated]
[New significance — a new answer to "So what?" not previewed in the introduction]
[Call for further research — what questions the work has opened, not closed]
```

---

## Pattern 1: Main point + new implication + research call

**Use when:** The research answers a conceptual question and opens further theoretical territory.

**Example — Legal history:**
> In light of recent Supreme Court decisions rejecting mandatory capital punishment, the mandatory death penalty for espionage is apparently unconstitutional and must be revised by Congress. More significantly, if the Uniform Code of Military Justice is changed, it will challenge the foundational value of military culture that ultimate betrayal requires the ultimate penalty — a value that extends well beyond the question of one penalty's constitutionality. Congress will then have to adjudicate not just a legal inconsistency, but the military's deeply held sense of what justice demands. Future research should examine how military culture has historically adapted its core values when external legal constraints required change, and whether those adaptations were perceived as legitimate by the communities they governed.

**Why it works:**
- Main point ("must be revised") is stated more fully than the introduction's single prediction
- New significance (the cultural challenge to military values) was withheld from the introduction intentionally — raising it there would have created an obligation to answer it in the body
- Research call names a specific question (how military culture adapts values under legal constraint) that follows naturally from the article's findings

---

## Pattern 2: Main point + practical application + research call

**Use when:** The research is fundamentally conceptual but has a concrete practical implication the body did not fully develop.

**Example — Medical diagnostics:**
> These differences between novice and expert diagnosticians define the trajectory of their maturation and development. Novices rely on presenting symptoms and surface-level pattern matching; experts rely on differential weighting of symptom clusters and active hypothesis elimination. While we now understand how novices and experts think differently, we do not yet understand which elements in the social experience of novices — mentoring relationships, feedback quality, case load — contribute most to the development of expert reasoning. Longitudinal studies on how coaching and active critique affect diagnostic outcomes would let medical educators design residency programs that accelerate, rather than merely wait for, the emergence of expert judgment.

**Why it works:**
- Main point (the novice-expert distinction) is stated more fully, not just repeated
- New significance is practical (implications for residency program design) — a step the body did not take
- Research call is specific (longitudinal studies, mentoring variables) rather than generic ("more research is needed")

---

## Pattern 3: Main point + extended significance (no separate research call)

**Use when:** The main point itself opens a question significant enough to function as both new significance and implicit research call. Used in shorter papers where a three-part conclusion would feel padded.

**Example — Cultural studies:**
> The quasi-persons of American science fiction — androids, half-aliens, engineered beings — consistently struggle toward a model of personhood defined by Western, male, rational criteria. That the struggle is presented as universally human, rather than as the projection of one culture's criteria onto all humanity, is itself the cultural work these figures perform. Recognizing this projection does not resolve it; but it makes visible the assumptions about personhood that American popular culture has exported globally for half a century — assumptions that audiences in the rest of the world have been invited to share, celebrate, or resist.

**Why it works:**
- Main point is restated and deepened ("the struggle is presented as universally human… is itself the cultural work")
- New significance is embedded in the extension: the global export of these assumptions opens questions about audience reception that the article did not address
- No explicit research call needed — the final sentence implies it without stating it mechanically

---

## Common conclusion failures

| Failure | What it looks like | Fix |
|---|---|---|
| Summary-only conclusion | "In this paper, I have argued X, shown Y, and demonstrated Z." | Restate the main point more fully; add one new implication or question |
| Word-for-word repetition | Conclusion uses the same sentences as the introduction | Expand the main point; the conclusion should feel like an arrival, not a rerun |
| Premature new argument | A significant new claim appears that the body did not build toward | Move the new claim to a future research call, or expand the body to support it |
| Deflation ("more research needed") | Conclusion closes with a generic call that signals the paper is incomplete | Make the research call specific: name the question, the method, or the type of evidence needed |
| Missing new significance | Conclusion ends at the main point without asking "So what?" one more time | Add one new implication — an application, a theoretical extension, or a further consequence |

FILE:references/introduction-configurations.md
# Introduction Configurations: Worked Examples

This reference covers every structural configuration of the Context→Problem→Response architecture, with worked examples across three disciplines.

---

## Configuration 1: Full structure (Context + Problem with consequence + Point-first Response)

**When to use:** General or introductory audience; problem is new to readers; writer wants to state the main point clearly at the end of the introduction.

**Template:**
```
[Opening: striking fact, quotation, or anecdote]
[Context: stable state of knowledge, 2-5 sentences]
[Condition: the gap — "We do not know…" / "X remains unclear…"]
[Consequence: what the gap prevents — "Without that…, we cannot…"]
[Response: main point stated directly]
```

**Example — Environmental science, general audience:**
> As our understanding of acid rain and carbon dioxide has improved, scientists have grown more confident in describing their effects on the biosphere. But recently, the chemical processes that thin the ozone layer have been found to be less well understood than once thought. We may have labeled hydrofluorocarbons as the chief cause incorrectly — meaning current regulatory strategies may be targeting the wrong compounds. This study reports a hitherto unexpected chemical bonding pattern between chlorine radicals and stratospheric water vapor that accounts for a larger share of ozone depletion than previously modeled.

---

## Configuration 2: Omit context (Problem + Response only)

**When to use:** Problem is already well known in the field; audience does not need the stable baseline established before seeing the disruption.

**Template:**
```
[Condition: the gap — stated directly without preamble]
[Consequence: what the gap prevents]
[Response: main point or launching point]
```

**Example — Mathematics, specialist audience:**
> The distribution of prime gaps in intervals beyond 10^18 remains poorly characterized by existing sieve methods. Without a tighter bound, probabilistic models of prime twin density cannot be validated. This paper derives an improved upper bound using a modified Cramér model that reduces the error margin by a factor of three.

**Why context is omitted:** Number theorists already know that prime gap distribution is an open problem; restating the history of the problem would be condescending.

---

## Configuration 3: Omit consequence (Context + Condition + Response, consequence implicit)

**When to use:** Specialist audience that would find the consequence self-evident; the connection between the condition and its larger implications is common knowledge in the field.

**Template:**
```
[Context: stable state of knowledge]
[Condition: the gap — stated without spelling out the consequence]
[Response: main point or launching point]
```

**Example — Molecular biology, specialist audience (after Crick and Watson):**
> A structure for nucleic acid has been proposed by Pauling and Corey, consisting of three intertwined chains with the phosphates near the fiber axis. In our opinion, this structure is unsatisfactory. We wish to suggest a structure for DNA that has novel features of considerable biological interest.

**Why consequence is omitted:** Every molecular biologist in 1953 understood that not knowing DNA's structure meant not understanding genetics. Spelling out the consequence would have seemed redundant and condescending.

---

## Configuration 4: Point-last (Context + Problem + Launching Point)

**When to use:** The conclusion requires the full argument to be persuasive; humanities and complex social science arguments; writer wants to control revelation.

**Template:**
```
[Opening: striking fact, quotation, or anecdote]
[Context: stable state of knowledge]
[Condition + Consequence]
[Launching point: concrete plan of the argument, NOT a vague topic announcement]
```

**Example — Legal history, academic audience:**
> In today's society, would Major John André — captured as a spy in civilian clothes behind American lines in 1780 — be hanged? Spying remains the only military offense mandating death. Recently, however, the Supreme Court has rejected mandatory death sentences in civilian cases, creating an ambiguity in their application to military law. If those decisions apply to the military, Congress may have to revise the Uniform Code of Military Justice. This article examines three Supreme Court cases, the legislative history of the Uniform Code, and two military tribunal rulings to determine whether that revision is constitutionally required.

**Why the launching point works:** It does not state the conclusion ("Congress must revise the Code") but it gives readers a concrete map of what the article will evaluate — three cases, legislative history, two rulings. Readers know the terrain before entering it.

**What NOT to do (vague topic announcement):**
> This article examines the relationship between Supreme Court decisions and the Uniform Code of Military Justice.
(No plan, no stakes, no reason to keep reading.)

---

## Discipline-specific conventions

| Discipline | Typical pace | Consequence spelling | Main point placement | Plan/outline |
|---|---|---|---|---|
| Natural sciences | Fast | Often omitted (specialists) | Introduction (point-first) | Often explicit |
| Social sciences | Moderate | Usually stated | Introduction (point-first) | Common |
| Humanities | Slow | Usually stated | Conclusion (point-last) | Less common; seen as heavy-handed |
| Professional reports | Fast | Stated explicitly | Introduction (point-first) | Standard |
| Legal writing | Moderate | Stated | Introduction (point-first) | Standard |

---
name: research-argument-builder
description: Build a complete, structured research argument from a framed problem — assembling all five elements (claim, reasons, evidence, acknowledgment/response, warrant) using the Claim→Reason→Evidence chain. Use this skill when the user has a research problem or framed question and needs to construct the supporting argument that justifies their answer, has a working thesis or claim but does not know how to assemble the reasons and evidence that make it hold, needs to identify which of the five claim types (fact, definition, cause, evaluation, policy) their main claim is and what kind of evidence each type demands, wants to evaluate whether their claim is specific and significant enough to anchor an argument, cannot tell whether a statement is a reason or evidence and keeps treating soft generalizations as hard data, has evidence but cannot determine whether it meets the quality standards (accurate, precise, sufficient, representative, authoritative) their readers will apply, needs to plan their argument visually using a storyboard (claim + reasons + evidence cards) before drafting, or wants to thicken a thin argument by identifying where acknowledgments and warrants are needed. This is the hub skill for research argumentation — use it before counterargument-handler (which handles detailed acknowledgment/response), warrant-tester (which tests whether reasons are genuinely relevant to claims), and research-paper-planner (which turns the completed argument structure into a paper outline).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/research-argument-builder
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [7, 8, 9]
tags: [research-methodology, academic-writing, argumentation, critical-thinking, evidence-evaluation]
depends-on: [research-problem-framer]
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "User's research problem statement, research question, working thesis/claim, or notes describing what they want to argue"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user should have a framed research problem (from research-problem-framer) or be able to state what they want to argue in one or two sentences"
discovery:
  goal: "Construct a complete, reader-ready research argument by assembling the five structural elements — claim, reasons, evidence, acknowledgment/response, warrant — in a logical order that answers every question a skeptical reader can predictably ask"
  tasks:
    - "Identify or sharpen the main claim and classify it by type (fact, definition, cause, evaluation, or policy)"
    - "Evaluate the claim for specificity and significance using the opposite-claim heuristic"
    - "Generate a set of reasons that support the claim, ordered logically"
    - "Distinguish reasons from evidence and identify the evidence needed to ground each reason"
    - "Evaluate available evidence against five quality criteria: accurate, precise, sufficient, representative, authoritative"
    - "Identify where acknowledgments and responses are needed and flag potential warrants"
    - "Produce a storyboard (claim + reason/evidence pairs + acknowledgment slots) ready for drafting"
  audience: "Researchers, graduate students, analysts, and professionals who have a research problem or framed question and need to assemble the argument that justifies their answer"
  triggers:
    - "User has a thesis but does not know how to build the argument around it"
    - "User's draft feels like a data dump with no argumentative structure"
    - "User conflates reasons and evidence, treating generalizations as hard data"
    - "User cannot tell whether their claim is specific or significant enough"
    - "User asks 'what kind of evidence do I need?' without knowing what type of claim they are making"
    - "User needs a logical plan for their argument before they start drafting"
    - "User asks 'is my evidence good enough?' without a quality framework to apply"
---

# Research Argument Builder

## When to Use

You have a research problem — a question plus the stakes of answering it — and you have done enough research to have a working answer. Now you need to assemble the argument that justifies that answer to skeptical readers.

A research argument is not a heated dispute. It is a cooperative inquiry: you and your readers working together to find the best answer to a question you both think matters. Your job is to anticipate every question a careful reader can ask and answer it before they can object.

Those questions reduce to five:

1. **Claim** — What do you want me to believe? What is your point?
2. **Reasons** — Why do you say that? Why should I agree?
3. **Evidence** — How do you know? Can you back it up?
4. **Acknowledgment and Response** — But what about...?
5. **Warrant** — How does that follow? What is your logic?

This skill walks you through all five elements in order, then produces a storyboard you can use as your drafting plan.

**Preconditions to verify:**

- Does the user have a framed research problem (condition + consequence)? If not, invoke `research-problem-framer` first.
- Does the user have a working answer — even a rough one — to their research question? If not, ask them to state what they think the answer is before continuing. The argument assembles around a claim; without one, there is nothing to build.

**This skill does NOT cover in full:**

- Developing detailed strategies for each objection or alternative view (use `counterargument-handler`)
- Testing whether a reason is genuinely relevant to a claim via warrant analysis (use `warrant-tester`)
- Turning the completed argument structure into a full paper outline with sections and order (use `research-paper-planner`)

## Context and Input Gathering

### Required (ask if missing)

- **The research question or problem:** What the researcher is trying to answer.
  - Check prompt for: a how/why question, a problem statement, or a framing from `research-problem-framer`
  - If missing, ask: "What is your research question — the specific thing you want to find out?"

- **The working claim:** The researcher's current best answer to the question.
  - Check prompt for: thesis statement, main argument, or any sentence that asserts something disputable
  - If missing, ask: "What do you currently think the answer is? Even a rough sentence is enough."

### Useful (gather from environment if available)

- **Notes, draft sections, or annotated sources:** May contain candidate reasons and evidence
  - Look for files like `notes.md`, `draft.md`, `outline.md`, `sources.md`, or any `.txt`/`.md` files in the working directory
  - If found, scan for claims, reasons, and data points before asking the user to supply them from scratch

- **Intended audience or field:** Determines what counts as authoritative evidence and which claim type fits
  - If not specified, ask: "Who are the readers this argument is for — a course, a scholarly field, a professional community?"

## Process

### Step 1 — Identify and Classify the Main Claim

**WHY:** The type of claim determines the type of argument you need to build and the type of evidence readers will demand. Trying to build an argument without knowing your claim type is like trying to construct a building without knowing whether it is a bridge, a house, or a dam. Each type requires different structural support.

State the main claim clearly as a single sentence (or two at most). Then classify it:

| Claim type | Core question answered | Evidence implication |
|---|---|---|
| Fact / existence | Does X exist or occur? | Observations, measurements, documented occurrences |
| Definition / classification | What kind of thing is X? | Criteria, comparison to established category, similarity/difference analysis |
| Cause / consequence | What caused X? What does X cause? | Causal mechanism, correlation + ruling out alternatives, time sequence |
| Evaluation / appraisal | Is X good or bad? Better or worse? | Explicit criteria of judgment, evidence that X meets or fails those criteria |
| Policy / action | What should be done? | Chain of conceptual sub-claims (problem exists → cause identified → solution addresses cause → solution is feasible + cost-effective) |

Most academic research produces **conceptual claims** (fact, definition, cause, or evaluation). Policy claims are **practical claims** — they require a chain of conceptual sub-arguments, not just one.

**Practical claim caution:** If the claim calls for action, it needs four sub-arguments: that the problem exists, what causes it, that the proposed solution addresses the cause, and that it is feasible and cheaper than alternatives. Do not build a policy argument as if it were a single conceptual claim.

See `references/claim-types-and-evidence.md` for worked examples of each claim type with corresponding evidence requirements.

### Step 2 — Evaluate Claim Strength: Specificity and Significance

**WHY:** A vague claim produces a vague argument. A trivial claim produces an argument readers do not think needs making. Both failures waste research. Checking specificity and significance before assembling evidence saves a complete rebuild later.

**Test 1 — Specificity**

Compare:
- Vague: "TV affects people's views of crime."
- Specific: "Graphic violence on local TV news leads regular viewers to overestimate neighborhood crime rates by up to 150 percent."

The specific version gives readers a richer set of concepts to interrogate. It also tells you exactly what evidence you need.

**How to sharpen specificity:** Write a working version of the claim with these two additions:
- An *although* clause acknowledging the main qualification readers might raise: `Although [widely held view you are challenging or limiting condition], [your claim]...`
- A *because* clause forecasting your main reason: `...[your claim] because [key reason].`

This is not your final draft claim — it is a working claim that makes your argument's logic explicit.

**Test 2 — Significance (opposite-claim heuristic)**

Flip your claim to its opposite. If the opposite seems obviously false or trivially unimportant, readers will not think your original claim needs arguing.

- Original: "Hamlet is not a superficial character." Opposite: "Hamlet is a superficial character." — Obviously false to anyone who has read the play. This claim does not need an argument.
- Original: "Graphic TV violence distorts viewers' risk perception." Opposite: "Graphic TV violence does not distort viewers' risk perception." — Contestable. Worth arguing.

**Significance proxy:** Ask — if readers accept this claim, how many other beliefs must they change? More belief-revision required = more significant claim.

### Step 3 — Generate Reasons and Order Them

**WHY:** Reasons are the logical spine of the argument. They are the assertions that, when placed between the claim and the evidence, explain *why* the evidence supports the claim. Without well-ordered reasons, an argument is a pile of data — readers cannot follow the logic even if they accept each individual fact.

**Key distinction — reasons vs. evidence:**

- **Reason:** A statement you think up. You use your mind to generate it. It explains *why* the claim follows. "We should leave — *it looks like rain*." (reason)
- **Evidence:** Data that exists in the world. You have to go find it. It anchors a reason in fact. "Barometric pressure has dropped 15 millibars in the past two hours." (evidence)

The direction of dependency is: reasons are *based on* evidence. Evidence does not *follow from* reasons.

**How to generate reasons:**

1. Ask: "Why should readers believe my claim?" Write every answer you can think of.
2. Ask for each answer: "Is this a reason I am asserting, or a piece of data I found?" Sort accordingly.
3. Check whether soft generalizations are reasons masquerading as evidence. "A majority of students leave college with a crushing debt burden" is a reason — it is a general assertion that still needs hard data to support it.

**Ordering reasons:**

Arrange reasons in a logical sequence — not a random list, not chronological unless the argument is about a sequence of events. Read just the reasons, without the evidence, to see if their order makes sense. If not, try other orders until the logical flow is clear.

### Step 4 — Distinguish Evidence from Reports of Evidence

**WHY:** Researchers almost never present the evidence itself — they present reports of it. A data table, a quoted passage, a case study description, a cited statistic: these are representations of evidence, shaped by those who collected and compiled them. Acknowledging this gap is not academic pedantry; it is what allows you to assess how far you are from the original data and how much trust you can ask readers to place in your report. The further you are from the original data, the more you must justify the chain of custody.

**The test:** Can readers plausibly ask "How do you know that?" more than once before reaching unquestionable bedrock? If yes, you have not yet reached evidence — you are still in the chain of reasons.

- Claim: "Higher tuition is harming educational access."
- Reason 1: "College has become unaffordable for low-income families."
- Soft reason 2 (treated as evidence): "Most students graduate with crushing debt." — *readers can still ask "How do you know?"*
- Harder evidence: "In 2013, 70 percent of students borrowed for college, averaging $30,000 in loans." — *harder to question, but still a report of data collected by someone else*

**Practical rule:** Cite as close to the original source as possible. When you use secondary sources, acknowledge it and explain why you could not get closer to the primary data.

### Step 5 — Evaluate Evidence Quality

**WHY:** Evidence that fails any of the five quality criteria will be challenged by careful readers, and even one failed criterion can discredit an otherwise strong argument. Running this checklist before drafting surfaces gaps early, when fixing them is still feasible.

The five criteria:

| Criterion | Question to ask | Common failure |
|---|---|---|
| **Accurate** | Is this data reported correctly and completely? | Misquotation, selective omission, rounding errors |
| **Precise** | Is it specific enough to mean something? | Vague quantities: "a great deal," "high probability," "many" |
| **Sufficient** | Is there enough evidence to support the reason? | One quotation or one number for a broad claim |
| **Representative** | Does it reflect the full range of the relevant data? | Cherry-picking, small unrepresentative samples |
| **Authoritative** | Is the source one readers will accept without question? | Wikipedia, uncited generalizations, non-expert opinion |

Apply this screen to every piece of evidence before adding it to your storyboard. Evidence that fails a criterion is not automatically disqualifying — but you must acknowledge the weakness and either supplement it or qualify the reason it supports.

See `references/evidence-quality-rubric.md` for detailed guidance on each criterion with cross-field examples.

### Step 6 — Identify Where Acknowledgments and Warrants Are Needed

**WHY:** An argument that only presents its own case looks one-sided. Careful readers will question every element, and if you have not anticipated their objections, they conclude you have not thought hard enough. Acknowledging objections before readers raise them, and responding to them, is what distinguishes scholarly argument from advocacy. Warrants — the general principles connecting a reason to a claim — are needed specifically when readers might accept a reason as true but deny it is *relevant* to the claim.

**Acknowledgment slots:** For each reason, ask: "What will readers say against this reason?" Identify the two or three most important objections or alternatives and note where you will acknowledge and respond to them.

**Warrant check:** For each reason→claim connection, ask: "Could a reader accept this reason as true but still deny it supports my claim?" If yes, you need a warrant — a general principle that makes the logical connection explicit.

Example: "We face higher health costs *because* the hard freeze line is moving north." A reader might accept both as true but not see the connection. The warrant: "When an area has fewer hard freezes, diseases carried by subtropical insects become more prevalent, raising medical costs." The warrant states the general principle that makes the specific reason relevant to the specific claim.

**When to state warrants explicitly:** Only when readers in your field might ask how a reason is relevant, or when you are explaining your field's reasoning to a general audience. Do not state warrants that your readers already take for granted — this wastes their time and can seem condescending.

For detailed warrant development and testing, use `warrant-tester`.
For detailed acknowledgment and response strategies, use `counterargument-handler`.

### Step 7 — Build the Storyboard

**WHY:** The storyboard externalizes the logical structure of the argument before you write prose. When the structure is on paper (or screen), you can see whether the reasons are in the right order, whether each reason has enough evidence, and where the acknowledgment gaps are — all without being tangled in the sentence-level decisions of drafting. Fixing the structure at this stage costs minutes; fixing it in a full draft costs hours.

Produce a storyboard with this format:

```
MAIN CLAIM: [one to two sentences]
Claim type: [fact / definition / cause / evaluation / policy]

REASON 1: [one sentence]
  Evidence: [source or data type needed]
  Evidence quality check: [any concerns with accuracy, precision, sufficiency, representativeness, authority]
  Acknowledgment needed: [yes/no — brief description of the objection]

REASON 2: [one sentence]
  Evidence: [source or data type needed]
  Evidence quality check: [any concerns]
  Acknowledgment needed: [yes/no]

[continue for each reason]

WARRANTS NEEDED: [list reason→claim connections that require a stated warrant]
```

Read through the reasons alone, without the evidence, to verify logical order. If the sequence does not make sense, reorder before finalizing.

## Examples

### Example 1 — Undergraduate humanities paper

**Input:** "My paper argues that Shakespeare's *Hamlet* develops the theme that indecision is more destructive than action, even wrong action."

**Step 1 — Claim classification:** Evaluation claim — judging *Hamlet* against a criterion (indecision as a form of destruction).

**Step 2 — Specificity/significance check:**
- *Although* clause: "Although Hamlet is often read as a play about moral paralysis caused by excessive thought..."
- *Because* clause: "...indecision is more destructive than action because every delay Hamlet makes produces a concrete death he could have prevented."
- Opposite: "Indecision in *Hamlet* is not more destructive than action." — Contestable. Worth arguing.

**Step 3 — Reasons:**
1. Each time Hamlet delays, the direct result is a preventable death (Polonius, Ophelia, Laertes, Gertrude, himself).
2. The characters who act decisively — Fortinbras, Laertes, even Claudius — achieve their immediate objectives.
3. Hamlet explicitly diagnoses his own problem as over-thinking, not lack of moral clarity.

**Step 4/5 — Evidence check:**
- Reason 1 needs: textual evidence from specific scenes (Acts 3–5), showing causal chain from delay to death.
- Reason 2 needs: textual examples of decisive action and its outcomes; must be *representative* (not cherry-picked scenes).
- Reason 3 needs: direct quotations from Hamlet's soliloquies; accuracy check — quote completely, not out of context.

**Storyboard fragment:**

```
MAIN CLAIM: In Hamlet, indecision is more destructive than action because
every delay produces a preventable death while decisive action — however
morally compromised — consistently achieves its objective.
Claim type: evaluation

REASON 1: Each of Hamlet's delays directly precedes a death he could have prevented.
  Evidence: Scene-by-scene textual analysis (Acts 3–5)
  Quality: Sufficient only if all major deaths are covered; representative
  Acknowledgment needed: Yes — "Hamlet could not have acted without more information earlier"
```

---

### Example 2 — Policy research

**Input:** "I want to argue that universities should require a one-semester research methods course for all undergraduates."

**Step 1 — Claim classification:** Policy claim — requires a chain of sub-arguments.

**Step 2 — Sub-claims needed:**
1. Most undergraduates currently lack basic research skills (fact)
2. The lack is caused by no structured instruction in research methodology (cause)
3. A required methods course would close that gap (cause/consequence)
4. The course is feasible and its benefits outweigh its costs (evaluation)

**Step 3 — Reasons for sub-claim 1:**
- Surveys show graduates cannot evaluate source credibility (fact sub-claim)
- Employers report new hires struggle to conduct independent research (fact sub-claim)

**Step 4/5 — Evidence:**
- Employer survey data: check *authoritativeness* (peer-reviewed or institutional?) and *representativeness* (industry range, not one sector)
- Graduate skills assessments: check *precision* (what exactly was measured, not just "critical thinking")

---

## Output

Produce a written storyboard following the format in Step 7, then provide:

1. **Claim diagnosis:** type, specificity assessment, significance check result
2. **Reasons list:** ordered sequence, with reason/evidence distinction confirmed for each
3. **Evidence inventory:** what you have, what you still need, quality concerns flagged
4. **Acknowledgment map:** where objections are anticipated and what the responses will cover
5. **Warrant flags:** any reason→claim connections that need a stated principle

If critical evidence is missing: name exactly what data type would fill the gap and where the researcher might find it. Do not recommend building the argument on soft reasons in place of unlocated evidence.

## Anti-Pattern Quick Reference

| Anti-pattern | Signal | Fix |
|---|---|---|
| Soft reason treated as evidence | "Most students..." "Many researchers..." — no source, no data | Ask "How do you know?" until you reach unquestionable data; cite the source |
| Policy claim without sub-argument chain | Single claim: "We should do X" | Break into 4 sub-claims: problem exists, cause identified, solution addresses cause, feasible + cheaper |
| Vague claim | "Technology affects education" | Apply *although/because* template to force specificity |
| Evidence fails opposite-claim test | Opposite claim is obviously false | The original claim is not worth arguing; sharpen or change the claim |
| Representative failure | One or two examples for a broad claim | Either narrow the claim to match the evidence or gather a representative sample |
| Warrant gap | Reason is true but readers say "so what?" | State the general principle that makes the specific reason relevant to the specific claim |

## References

- `references/claim-types-and-evidence.md` — Five claim types with worked examples and corresponding evidence requirements per type
- `references/evidence-quality-rubric.md` — Detailed guidance on all five evidence quality criteria with cross-field examples and failure modes

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-problem-framer`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/claim-types-and-evidence.md
# Claim Types and Evidence Requirements

Reference for `research-argument-builder`, Step 1.

This document provides worked examples of each claim type and specifies what kind of evidence readers expect for each.

---

## The Five Claim Types

### 1. Claims of Fact or Existence

**Core question answered:** Does X exist? Does X occur? Is X the case?

**Structure:** Asserts that something is (or is not) true, present, or happening.

**Examples:**
- Average global temperatures have risen to unprecedented levels within the past decade.
- Graphic violence on local TV news leads regular viewers to overestimate neighborhood crime rates by up to 150 percent.
- The rate of second-language acquisition declines significantly after puberty.

**Evidence required:**
- Direct observations, measurements, counts, or documented occurrences
- Experimental data, survey data, or field data that establishes the fact
- Primary sources (records, original studies, official statistics)
- Evidence must be **accurate** (correctly recorded), **precise** (specific enough to be falsifiable), and **sufficient** (enough data points to rule out coincidence)

**Common failure:** Treating a widely held belief as a fact without supplying data. "Most people believe X" is not evidence that X is true.

---

### 2. Claims of Definition or Classification

**Core question answered:** What category does X belong to? What kind of thing is X? How is X best characterized?

**Structure:** Assigns an entity to a broader class or distinguishes it from other entities based on shared or differing features.

**Examples:**
- Birds, not reptiles, are the direct descendants of dinosaurs.
- The 2008 financial crisis was a systemic failure of regulatory capture, not a random market event.
- The practice of using historical analogies in political speeches is a form of deliberate myth-making.

**Evidence required:**
- Established criteria that define the category (what makes something a "myth," a "systemic failure," etc.)
- Evidence that the subject meets (or fails to meet) those criteria
- Comparison to similar and contrasting cases to show why your classification is more accurate than alternatives
- Definitional sources: how the field defines the category (dictionaries are rarely sufficient; field-specific definitions required)

**Common failure:** Assuming the category is self-evident and skipping the criteria step. The argument must show why the subject qualifies, not just assert that it does.

---

### 3. Claims of Cause or Consequence

**Core question answered:** What caused X? What does X cause? What will X lead to?

**Structure:** Connects two situations or events, arguing that one produces or produces (or produced or will produce) the other.

**Examples:**
- Exposure to asbestos is a leading contributor to lung cancer.
- The Great Recession of 2008–2011 was caused largely by negligent financial regulation that let banks take on too much risk.
- Climate change is increasing the range of diseases carried by subtropical insects, raising healthcare costs in northern regions.

**Evidence required:**
- Evidence of a plausible causal mechanism (not just correlation)
- Evidence that X preceded Y in time (for causal claims about the past)
- Evidence ruling out alternative causes (especially for contested causal claims)
- Statistical correlation data plus mechanistic explanation
- Longitudinal data or controlled experimental data where available

**Common failure:** Confusing correlation with causation, or presenting a single mechanism without ruling out competing explanations. Readers of causal claims are especially alert to this failure.

**Note on consequence claims:** "X will cause Y" (a forward-looking claim) requires a warrant connecting the known mechanism to the predicted outcome, plus evidence that the conditions for Y are present.

---

### 4. Claims of Evaluation or Appraisal

**Core question answered:** Is X good or bad? Is X better or worse than Y? Does X succeed or fail by some standard?

**Structure:** Judges something against explicit criteria.

**Examples:**
- Shakespeare's greatest comedy is *As You Like It*.
- The CDC's pandemic response was inadequate because it prioritized data collection over early public communication.
- *Hamlet* develops the theme that indecision is more destructive than action.

**Evidence required:**
- Explicit statement of the evaluative criteria (why does this standard apply?)
- Evidence that the subject meets or fails those criteria (textual examples, performance data, outcome measurements)
- Evidence that the criteria are appropriate to the subject and accepted by the intended audience
- Comparison to other entities evaluated by the same criteria (optional but strengthens the judgment)

**Common failure:** Stating an evaluation without making the criteria explicit. "This was a bad policy" is not an argument until the reader knows what standard of "bad" is being applied.

---

### 5. Claims of Policy or Action

**Core question answered:** What should be done? What action should be taken?

**Structure:** Argues for (or against) a course of action or policy. Requires a chain of four conceptual sub-claims.

**Example:**
- Pennsylvania should increase extraction fees on natural gas drilling to fund education.

**Required sub-argument chain:**

| Sub-claim | Type | Evidence required |
|---|---|---|
| A problem exists | Fact | Data showing the problem is real and significant |
| This cause is responsible | Cause | Causal mechanism evidence |
| The proposed solution addresses the cause | Cause/consequence | Evidence that the intervention targets the right variable |
| The solution is feasible and cost-effective | Evaluation | Cost analysis, implementation precedents, comparison to alternatives |

**Common failure:** Arguing the policy claim as if it were a single assertion rather than four linked sub-claims. If any link in the chain is missing, readers can reject the whole argument.

**Relationship to practical vs. conceptual claims:** Policy claims are **practical** claims. Most academic research produces **conceptual** claims (types 1–4). Do not attach a policy claim to a conceptual argument unless you have built all four sub-claims. Suggesting a policy application is appropriate in a conclusion as an optional extension, not as the main claim.

---

## Cross-Cutting Note: Mixing Claim Types

Most research arguments contain multiple claim types. A common structure:

1. Establish that a phenomenon exists (fact claim)
2. Classify or define it precisely (definition claim)
3. Explain its cause or consequence (cause claim)
4. Evaluate it against relevant criteria (evaluation claim)
5. Optionally propose a response (policy claim)

When you identify your **main claim**, you are identifying the claim that anchors the argument — the claim your paper's conclusion restates. The other claim types may appear as sub-claims or supporting arguments.

FILE:references/evidence-quality-rubric.md
# Evidence Quality Rubric

Reference for `research-argument-builder`, Step 5.

Five criteria for evaluating the quality of evidence before adding it to your argument storyboard. Apply each criterion to every piece of evidence. Evidence that fails a criterion is not automatically discarded — but the failure must be acknowledged and either supplemented or used to qualify the reason it supports.

---

## The Five Criteria

### 1. Accurate

**Question:** Is this data reported correctly and completely?

**What it means:** Evidence is accurate when it faithfully represents what was observed, measured, or recorded — no errors in transcription, no selective omission of contradictory data, no rounding that changes meaning.

**Common failures:**
- Misquotation (even a single changed word can alter meaning)
- Selective quotation (a passage that, in context, says the opposite)
- Numerical errors from copying data across sources
- Reporting a statistic from an earlier version of a dataset when a corrected version exists

**How to check:**
- Quote directly and completely; do not paraphrase where exact wording matters
- Cross-check numbers against the original source, not a secondary citation
- If using data from a secondary source, note the original collection date and whether it has been superseded

**Field note:** In quantitative fields, accuracy includes documenting units, measurement instruments, and margin of error. In humanities, accuracy means the textual source is identified precisely enough for readers to verify it.

---

### 2. Precise

**Question:** Is the evidence specific enough to be meaningful?

**What it means:** Precise evidence uses specific quantities, timeframes, populations, or terms that let readers assess the claim's substance. Vague language functions as a hedge that cannot be evaluated.

**Warning language to watch for:**
- Quantities: "a great deal," "many," "some," "most," "often," "frequently," "several"
- Probability: "a high probability," "likely," "tends to"
- Magnitude: "large," "significant," "considerable," "substantial"

These words can appropriately limit a claim's scope (see hedges, Ch. 8.3), but when used as evidence — rather than as qualifiers — they fudge a claim the researcher did not work hard enough to quantify.

**Examples:**

| Imprecise | Precise |
|---|---|
| "The Forest Service spent a great deal of money on fire prevention." | "The Forest Service spent $2.7 billion on fire suppression in 2022, up from $1.0 billion in 2000." |
| "Many students struggle with debt." | "In 2013, 70 percent of students who borrowed for college averaged $30,000 in loans at graduation." |
| "Hamlet delays frequently." | "Hamlet has five direct opportunities to act against Claudius in Acts 3–4 and delays in all five." |

**Field note:** Precision standards vary by discipline. A physicist measures in nanoseconds; a historian might place an event in a decade. Both can be appropriately precise *by their field's standards*. Precision means specific enough that no reader can reasonably say "what does that actually mean?"

---

### 3. Sufficient

**Question:** Is there enough evidence to support the reason?

**What it means:** A single example, quotation, or data point is rarely enough to support a significant reason — although sometimes a single counter-example is sufficient to disprove a universal claim. Sufficient evidence means the body of data is large enough (in scope or quantity) that readers cannot dismiss it as coincidence or cherry-picking.

**Common failures:**
- Citing one quotation to support a claim about all of an author's works
- Using one year's data to support a trend claim
- Drawing a conclusion about a population from a sample that covers one subgroup

**Diagnosis:** Ask "Could a skeptical reader dismiss this as coincidence or a single outlier?" If yes, more evidence is needed or the claim must be narrowed to match the available evidence.

**Narrowing the claim vs. finding more evidence:** When you cannot gather more evidence, narrow the claim to fit what you have. "Shakespeare's *Hamlet* and *Macbeth* portray women as dangerous" is supported by two plays; "Shakespeare systematically portrays women as dangerous" requires evidence from the full corpus.

---

### 4. Representative

**Question:** Does this evidence reflect the full range of relevant variation?

**What it means:** Evidence is representative when it is not systematically skewed toward cases that support your claim and away from cases that complicate it. Representativeness is the criterion most easily violated by selective reading.

**Common failures:**
- Using only examples that confirm your hypothesis while ignoring counterexamples
- Sampling from a population in a way that over-represents one subgroup (e.g., surveying only urban respondents for a national claim)
- Using anecdotal evidence (one case, one exception) as if it stood for a pattern

**The anecdote risk:** A single well-chosen case study or exception can be persuasive in ways statistical data are not — the vivid example sticks in readers' minds. But if your argument depends on one or two examples, however compelling, it is vulnerable to the charge of cherry-picking. Use anecdotal evidence to *illustrate* a pattern, not to *establish* one.

**How to demonstrate representativeness:**
- For survey data: describe your sampling method and the population it represents
- For textual/case evidence: explain why the cases you selected represent the full range (or acknowledge that they are illustrative examples of a broader pattern established by other evidence)
- Acknowledge the cases that do not fit your pattern; explain why they are exceptions rather than refutations

---

### 5. Authoritative

**Question:** Is this source one my readers will accept without question?

**What it means:** Readers assign degrees of credibility to sources based on the source's reputation for rigor, expertise, and objectivity. Evidence drawn from a non-authoritative source will be dismissed regardless of its content.

**Authority is audience-dependent:** What counts as authoritative varies by field, community, and reader. The Centers for Disease Control is authoritative on disease transmission for most audiences. Wikipedia is not authoritative in academic contexts (because it aggregates rather than produces original verified data). A peer-reviewed journal article is more authoritative than a blog post by the same author.

**Common failures:**
- Citing Wikipedia for factual claims
- Citing a popular press summary when the original study is available
- Citing an expert's opinion in a field outside their expertise
- Citing data from an organization with a financial or political interest in the outcome without acknowledging that interest

**Levels of authority (general to specific):**
1. Your own primary data (most direct, requires methodology disclosure)
2. Peer-reviewed studies in the relevant field (high authority, standard in academic work)
3. Government and institutional reports (authority depends on the institution and its methods)
4. Expert opinion from credentialed specialists (appropriate for evaluation and policy claims; not sufficient for factual claims alone)
5. Journalism from reputable outlets (secondary authority; use when primary data is unavailable)
6. Aggregators, wikis, anonymous sources (not authoritative in most serious arguments)

**Chain of custody:** If you are reporting data from a secondary source that itself cites a tertiary source, you are at three or four removes from the original evidence. Acknowledge the chain and explain why you could not get closer to the primary source.

---

## Quick-Screen Checklist

Before adding a piece of evidence to your storyboard, run through:

- [ ] **Accurate** — Have I verified this against the original source? Am I quoting or citing completely and in context?
- [ ] **Precise** — Can readers tell exactly what this data means? No vague quantifiers hiding imprecision?
- [ ] **Sufficient** — Do I have enough data points to rule out coincidence? Or do I need to narrow my claim?
- [ ] **Representative** — Am I selecting evidence fairly? Have I accounted for counterexamples?
- [ ] **Authoritative** — Will my intended readers accept this source without question? If not, can I supplement or upgrade it?

A "no" on any criterion is a flag, not an automatic rejection. Acknowledge the weakness explicitly and adjust: either find better evidence, qualify the reason the evidence supports, or acknowledge the limitation directly in your paper.

---
name: prose-clarity-reviser
description: "Revise dense, unclear prose into clear, readable sentences by applying four diagnostic principles — characters-as-subjects, actions-as-verbs, old-before-new information flow, and complexity-last sentence endings. Use this skill whenever the user shares a draft passage, paragraph, or document and asks you to make it clearer, more readable, easier to follow, less dense, less academic-sounding, or better written — even if they don't use those words. Also triggers on: \"can you revise this,\" \"this feels clunky,\" \"my advisor said this is unclear,\" \"make this flow better,\" \"my writing sounds stilted,\" or any request to improve prose style in research papers, reports, essays, grant proposals, or professional documents. Apply all four principles end-to-end; return a revised version with brief annotations showing what changed and why."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/prose-clarity-reviser
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [17]
tags: [research-methodology, prose-revision, writing-style, clarity, academic-writing]
depends-on: []
execution:
  tier: 1
  mode: full
  inputs:
    - type: document
      description: "One or more paragraphs of draft prose to revise — a sentence, paragraph, section, or full document excerpt"
  tools-required: []
  tools-optional: [Read, Write]
  mcps-required: []
  environment: "Any agent environment. Works on pasted text or referenced files."
---

# Prose Clarity Reviser

## When to Use

The user has written prose — in any research, academic, or professional context — and wants it to be clearer or more readable. Typical triggers:

- User pastes a paragraph and says "this is confusing" or "make this clearer"
- User shares a draft section and asks for a style pass or readability revision
- A supervisor, editor, or reviewer has flagged the writing as dense, abstract, or hard to follow
- The user says their writing sounds too "academic," too stiff, or too indirect
- The user asks why their sentences don't "flow" or why readers keep re-reading them

**Do not apply these principles as the user writes new text.** These are revision tools — apply them only after a draft exists. Applying them during drafting causes paralysis.

**Scope:** This skill handles sentence-level and paragraph-level style revision only. It does not restructure arguments, reorganize sections, or evaluate evidence. For those concerns, use a different skill.

## Context

### Required (must be present before proceeding)

- **The text to revise:** One or more sentences or paragraphs. If the user says "revise my paper" without sharing text, ask them to paste the passage.

### Useful but not required

- **Audience:** Who are the intended readers? Specialists, generalists, students, peers? Knowing the audience determines whether field-specific nominalizations can stay as-is (see Principle 2 nuance below).
- **Purpose of the passage:** Is this a methods section, an argument, an introduction? This affects how active vs. passive voice should be used (see Principle 4).

## The Four Principles

Every unclear sentence violates one or more of these four structural rules. Diagnose before revising.

---

### Principle 1: Characters as Subjects

**The rule:** The grammatical subject of every clause should name the main character in that clause's story — the entity doing, experiencing, or driving the action.

**Why it matters:** Readers orient themselves by asking "who or what is this sentence about?" When the subject is an abstraction, a nominalization, or an empty phrase, readers lose the thread. They have to work backward to reconstruct the actual agent.

**Diagnostic — underline the first 6–7 words of every clause.** Then ask:
1. Does the subject name a concrete character (person, organization, organism, system, concept your field treats as an active agent)?
2. If the subject is abstract, is it an abstraction your readers already treat as a familiar actor (e.g., "inflation," "the immune system" in a medical paper)?

**If no to both — revise:**
1. Identify who or what the sentence is really about. If you cannot name a character, invent one (use "we," "researchers," "the study").
2. Make that character the grammatical subject.
3. Keep subjects short, specific, and concrete. Long whole-subjects (e.g., "The standardization of indices for the measurement of mood disorders") signal that the real character has been buried.

**Field-specific note:** In your discipline, the main characters may be abstractions — "monetary policy," "institutionalization," "gene expression." This is acceptable when those abstractions are already familiar to your audience and can be treated as acting entities. The problem occurs when you layer additional abstractions around them as additional nominalizations.

---

### Principle 2: Actions as Verbs (Nominalization Detection)

**The rule:** The key action in every clause should appear as a verb, not as a noun derived from a verb or adjective (a nominalization).

**Why it matters:** When actions become nouns, three things go wrong simultaneously: (1) you lose the specific verb and replace it with a vague one like *make*, *have*, *do*, or *be*; (2) you bury the characters as modifiers or objects of prepositions; (3) you clutter the sentence with articles and prepositions that would not be needed if the action were a verb.

**Nominalization markers to scan for:**
- Suffixes: *-tion, -ness, -ment, -ence, -ity, -ance, -al* (when used as noun)
- Common nominalizations: *analysis, assumption, consideration, development, establishment, examination, implementation, improvement, indication, investigation, provision, requirement, utilization*
- Some nominalizations look identical to their verb: *change, delay, report, plan, review, shift, study* — context determines which they are

**Diagnostic — for each nominalization you find, ask:**
1. Is this the subject of a verb? If so, and if it derives from a verb, that verb has been buried.
2. Is this from my field's standard vocabulary — a term readers recognize as a concept, not just a buried verb? (e.g., "hospitalization," "measurement" in a clinical paper) If yes, it may stay.

**To revise:**
1. Find the buried verb inside the nominalization (e.g., *consideration* → *consider*, *establishment* → *establish*, *utilization* → *use*).
2. Make that verb the main verb of the clause.
3. Move the character (now freed from modifier position) back to subject position.

**The exception:** Do not convert every nominalization into a verb. When a nominalization refers to a concept that your readers treat as familiar and established — especially if you introduced it earlier in the passage — leave it as a noun. Abstract nouns used to echo earlier content serve the old-before-new principle (Principle 3). Revise unnecessary nominalizations; keep necessary ones.

---

### Principle 3: Old Before New (Information Flow)

**The rule:** Begin each sentence with information readers already know (old) — either from the previous sentence or from shared context — and place new information at the end.

**Why it matters:** Readers build meaning incrementally. When each sentence opens with something familiar, readers have an anchor. They can attach the new information at the end to something they already hold in mind. When sentences open unpredictably, readers cannot connect them; the passage feels disjointed even if every individual sentence is well-formed.

**This principle overrides Principle 1 when they conflict.** If you must choose between starting with a character or starting with familiar information, always choose the familiar information.

**Diagnostic — underline the first 6–7 words of every sentence (skip short introductory phrases like "At first," "For the most part").** Then run your eye down the page and ask:
1. Do the sentence-opening words form a consistent, related set? They do not need to be identical, but they should name people or concepts that your readers will see as clearly related.
2. Does each sentence begin with a word or phrase that echoes something from the previous sentence or from shared context?

**If no — revise:**
1. Reorder: move the familiar information to the front of the sentence.
2. Use pronouns, synonyms, or summary nouns to refer back to what was just mentioned.
3. Move new, complex, or technical information to the end of the sentence. This often requires using the passive voice (see Principle 4).

**Active vs. passive and information flow:** The passive voice is not a flaw. Its primary function is to flip old and new information — to move older, familiar information from the end of one sentence to the beginning of the next. When familiar information would naturally be the object of an active verb, use the passive to promote it to subject position. Do not use passive or active based on a blanket rule; use whichever puts the right information at the beginning.

---

### Principle 4: Complexity Last (Sentence Endings)

**The rule:** Place the newest, most complex, most technically demanding information at the end of a sentence, never at the beginning.

**Why it matters:** Readers can handle complexity better when they arrive at it from stable ground. If a sentence opens with a long, complex subject or an unfamiliar technical phrase, readers must decode it before they even know what the sentence is doing. Placing complexity last lets readers approach it with context already built.

**Three contexts where this principle is especially important:**
1. **Introducing a new technical term** — the term should appear in the last few words of the sentence, not the first.
2. **Presenting a long, complex bundle of information** — if a clause requires a long phrase or subordinate clause, push it to the sentence's end.
3. **Opening a paragraph** — the last few words of the paragraph's opening sentence should name the key terms that the rest of the paragraph will develop.

**Diagnostic — underline the last 5–6 words of every sentence.** Ask:
1. Are those words the most important, complex, or emphatic in the sentence?
2. Are they technical terms being introduced for the first time?
3. Are they concepts that the next several sentences will develop?

**If no — revise:** Move the most important, newest content to the sentence's end. This may require splitting the sentence, using a passive verb, or reordering clauses.

---

## Active vs. Passive: The Decision Criterion

Use active voice when:
- The writer/researcher is performing an action that only they can perform: *argue, conclude, suggest, claim, design, prove*
- The action you want to emphasize is something only the authors did (rhetorical or intellectual ownership)

Use passive voice when:
- Describing a process that any qualified person could repeat (standard in methods sections of scientific writing)
- You need to move familiar information to the front of a sentence (information flow rule)

Do not use active or passive based on a blanket stylistic rule. The passive is not a sign of weak writing — it is a structural tool for managing information flow.

---

## Revision Protocol

Apply in this order. Do not try to apply all four principles simultaneously.

**Pass 1 — Diagnosis (mark up the text)**
1. Highlight the first 6–7 words of every clause. Skip short introductory phrases.
2. Scan for nominalizations: look for words ending in *-tion, -ness, -ment, -ence, -ity*. Mark each one.
3. Ask of each highlighted opening: Is this a character? Is this familiar information?
4. Highlight the last 5–6 words of every sentence. Ask: Is this the most important content?

**Pass 2 — Character and verb revision (Principles 1 and 2)**
5. Find sentences where the subject is not a character. Identify the real character.
6. Find nominalizations. Determine which are field-standard (keep) and which bury a verb (convert).
7. Rewrite: make real characters the subjects of specific, concrete verbs.

**Pass 3 — Information flow revision (Principle 3)**
8. Reorder sentence openings to begin with familiar information.
9. Shift new, complex, or technical content to sentence ends.
10. Adjust active/passive as needed to serve information flow — not as a separate stylistic choice.

**Pass 4 — Sentence endings (Principle 4)**
11. Check sentence endings. If the most important content is buried mid-sentence, reorder.
12. Check paragraph-opening sentences. Confirm that the last few words introduce the paragraph's key term.

**Pass 5 — Output**
13. Produce the revised text.
14. For each paragraph, add a brief annotation listing: what changed and which principle was applied. Format:
    - `[P1]` — character moved to subject
    - `[P2]` — nominalization converted to verb
    - `[P3]` — old-before-new reordering
    - `[P4]` — complexity moved to sentence end
    - `[P3+voice]` — passive/active adjusted for information flow

---

## Examples

### Example 1: Heavy Nominalization + Abstract Subject

**Before:**
> The standardization of indices for the measurement of mood disorders has now made possible the quantification of patient response as a function of treatment differences.

**After:**
> Having standardized indices for measuring mood disorders, we now can quantify patients' responses to different treatments.

**Annotations:**
- `[P2]` *standardization* → *standardized*, *measurement* → *measuring*, *quantification* → *quantify*, *response* → *responses*
- `[P1]` Subject changed from the abstract nominalized phrase to *we*
- `[P2]` Vague verb *made possible* replaced by specific verb *can quantify*

---

### Example 2: Information Flow Broken Across Sentences

**Before:**
> Locke frequently repeated himself because he did not trust the power of words to name things accurately. Seventeenth-century theories of language, especially Wilkins's scheme for a universal language involving the creation of countless symbols for countless meanings, had centered on this naming power. A new era in the study of language that focused on the ambiguous relationship between sense and reference begins with Locke's distrust.

**After:**
> Locke often repeated himself because he distrusted the naming power of words. This naming power had been central to seventeenth-century theories of language, especially Wilkins's scheme for a universal language involving the creation of countless symbols for countless meanings. Locke's distrust begins a new era in the study of language, one that focused on the ambiguous relationship between sense and reference.

**Annotations:**
- `[P3]` Sentence 2 rewritten to open with *This naming power* — picks up the phrase from sentence 1's end
- `[P3]` Sentence 3 rewritten to open with *Locke's distrust* — picks up the concept from sentence 2's end
- `[P2]` *distrust of* → *distrusted* (minor nominalization)
- `[P4]` New technical concept (*ambiguous relationship between sense and reference*) moved to sentence 3's end

---

### Example 3: Complexity at the Front, Technical Term Placement

**Before:**
> The monoamine hypothesis has been the leading biological account of depression for over three decades.

**After:**
> For over three decades, the leading biological account of depression has been the monoamine hypothesis.

**Annotation:**
- `[P4]` Technical term *monoamine hypothesis* moved to sentence end, where new technical terms belong when first introduced. The familiar framing (*for over three decades*) now anchors the sentence's opening.

---

## Field-Specific Notes

**Sciences (natural, social, clinical):** Use passive in methods sections to describe repeatable processes. Use active and first person (*we conclude*, *we argue*, *we designed*) when describing intellectual contributions unique to the authors. Do not use passive to appear "objective" — it only changes whose story is being told.

**Humanities:** Main characters are often real people or abstract concepts (*democracy*, *modernity*, *the text*). These abstractions can stay as subjects if they are familiar to your audience and you are not layering additional nominalizations around them.

**Technical and policy writing:** Characters are often institutions, systems, or policies. Keep them as subjects. Avoid nominalizing verbs that describe what those institutions actually do (*implement* vs. *implementation of*, *regulate* vs. *regulation of*).

---

See `references/before-after-sentence-pairs.md` for 15+ additional before/after sentence pairs illustrating all four principles across different academic disciplines.

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/before-after-sentence-pairs.md
# Before/After Sentence Pairs: Prose Clarity Revision

Source: Booth, Colomb, Williams, Bizup, FitzGerald. *The Craft of Research*, 4th ed., Chapter 17 (pp. 288–309).

Each pair is labeled with the primary principle(s) applied.

---

## Principle 1: Characters as Subjects

**Pair 1** — Abstract subject buried in long whole-subject phrase

Before:
> The reason for Locke's frequent repetition lies in his distrust of the accuracy of the naming power of words.

After:
> Locke frequently repeated himself because he did not trust the power of words to name things accurately.

Primary principle: `[P1]` Subject changed from *the reason for Locke's frequent repetition* to *Locke* (concrete, active character).

---

**Pair 2** — Stripping action vs. deforestation as subject

Before:
> The stripping of rain forests in the service of short-term economic interests could result in damage to the earth's biosphere.

After:
> If rain forests are stripped to serve short-term economic interests, the earth's biosphere may be damaged.

Primary principle: `[P1]` Two concrete characters (*rain forests*, *the earth's biosphere*) replace the abstract nominalized subject. Also `[P2]` *stripping* → *stripped*, *service* → *serve*.

---

**Pair 3** — Choosing the character intentionally (developer, logger, or nation)

Neutral (no clear actor named):
> If rain forests are stripped to serve short-term economic interests, the earth's biosphere may be damaged.

With named actors:
> If developers strip rain forests to serve short-term economic interests, they may damage the earth's biosphere.
> If loggers strip rain forests to serve short-term economic interests, they may damage the earth's biosphere.
> If Brazil strips its rain forests to serve short-term economic interests, it may damage the earth's biosphere.

Principle: `[P1]` The choice of character (developer, logger, Brazil) changes the story the sentence tells. Choose the character your argument is actually about.

---

## Principle 2: Actions as Verbs (Nominalization Conversion)

**Pair 4** — Four nominalizations in one sentence

Before:
> The standardization of indices for the measurement of mood disorders has now made possible the quantification of patient response as a function of treatment differences.

After:
> Having standardized indices for measuring mood disorders, we now can quantify patients' responses to different treatments.

Conversions: *standardization* → *standardized*, *measurement* → *measuring*, *quantification* → *quantify*, *response* → *responses*.
Also: vague verb *made possible* replaced by specific verb *can quantify*. `[P2]` + `[P1]`.

---

**Pair 5** — Action-as-noun creates pronoun ambiguity and article clutter

Before:
> The reason for Locke's frequent repetition lies in his distrust of the accuracy of the naming power of words.

After:
> Locke frequently repeated himself because he did not trust the power of words to name things accurately.

Conversions: *repetition* → *repeated*, *distrust* → *did not trust*. Note how eliminating the nominalizations also removes the articles and prepositions that cluttered the original. `[P2]` + `[P1]`.

---

**Pair 6** — Keeping necessary nominalizations (field-standard terms)

Medical passage (the nominalizations in sentence 2 are field-standard and should stay):
> Standardized indices to measure mood disorders help us quantify how patients respond to different treatments. These measurements suggest that treatments requiring long-term hospitalization are no more effective than outpatient care for most patients.

Note: *measurements*, *treatments*, *hospitalization*, *care* are all familiar to medical readers as conceptual terms, not buried verbs. Revising them into verbs would not improve clarity for the intended audience.

However, this version would benefit from revision:
Before:
> The hospitalization of patients without appropriate treatment results in the unreliable measurement of outcomes.

After:
> We cannot measure outcomes reliably when patients are hospitalized but not treated appropriately.

`[P2]` *hospitalization* → *are hospitalized*, *measurement* → *measure*. `[P1]` Subject changed to *we*. `[P3]` Old information (*patients*) moved to front of dependent clause.

---

## Principle 3: Old Before New (Information Flow)

**Pair 7** — Disjointed paragraph vs. chained paragraph

Before (each sentence opens with unpredictable new information):
> Because the naming power of words was distrusted by Locke, he repeated himself often. Seventeenth-century theories of language, especially Wilkins's scheme for a universal language involving the creation of countless symbols for countless meanings, had centered on this naming power. A new era in the study of language that focused on the ambiguous relationship between sense and reference begins with Locke's distrust.

Opening words (unpredictable): "Because the naming power" / "Seventeenth-century theories" / "A new era"

After (each sentence opens with information from the previous sentence):
> Locke often repeated himself because he distrusted the naming power of words. This naming power had been central to seventeenth-century theories of language, especially Wilkins's scheme for a universal language involving the creation of countless symbols for countless meanings. Locke's distrust begins a new era in the study of language, one that focused on the ambiguous relationship between sense and reference.

Opening words (predictable): "Locke" / "This naming power [from sentence 1's end]" / "Locke's distrust [from sentence 2's end]"

`[P3]` full paragraph chain reconstruction.

---

## Principle 4: Complexity Last — Active/Passive Decision for Information Flow

**Pair 8** — Passive used correctly to serve old-before-new

Before (active — breaks information flow):
> The quality of our air and even the climate of the world depend on healthy rain forests in Asia, Africa, and South America. But the increasing demand for more land for agricultural use and for wood products for construction worldwide now threatens these forests with destruction.

After (passive — restores flow):
> The quality of our air and even the climate of the world depend on healthy rain forests in Asia, Africa, and South America. But these rain forests are now threatened with destruction by the increasing demand for more land for agricultural use and for wood products used in construction worldwide.

In the "after" version, sentence 2 opens with *these rain forests* — picking up the character from the end of sentence 1 (*rain forests in Asia, Africa, and South America*). The passive verb *are threatened* is what makes this possible. `[P3+voice]`.

---

**Pair 9** — Active voice appropriate for unique authorial actions

Before (passive — hides authorial ownership):
> It can be concluded that the fluctuations result from the Burnes effect.

After (active — claims the conclusion):
> We conclude that the fluctuations result from the Burnes effect.

The active first-person form is appropriate here because *concluding* is something only the authors can do — it is an intellectual act that belongs to them. Scientists use first person and active voice when describing what they argued, designed, proved, or concluded. `[P3+voice]` — but driven by authorship logic, not just information flow.

---

**Pair 10** — Passive appropriate for repeatable scientific processes

Before (active — implies only the authors could do this):
> We measured eye movements at tenth-of-second intervals.

After (passive — signals any qualified person could replicate this):
> Eye movements were measured at tenth-of-second intervals.

Both sentences convey equal information. The difference is the story: the active version is about the authors; the passive version is about the process. When describing a repeatable experimental procedure, the passive is not just acceptable — it is the right choice. `[P3+voice]`.

---

## Principle 4: Complexity Last — Technical Term Placement

**Pair 11** — Technical term buried at sentence start

Before:
> The monoamine hypothesis has been the leading biological account of depression for over three decades.

After:
> For over three decades, the leading biological account of depression has been the monoamine hypothesis.

`[P4]` New technical term *monoamine hypothesis* moved to sentence end, where unfamiliar terms belong when first introduced.

---

**Pair 12** — Technical terms accumulating mid-sentence in a passage

Before:
> The monoamine hypothesis has been the leading biological account of depression for over three decades. According to this hypothesis, deficits in monoamines including dopamine, epinephrine, norepinephrine, and serotonin are associated with depression. Monoamine concentrations in neural synapses are regulated in different ways by different types of antidepressants.

After:
> For over three decades, the leading biological account of depression has been the monoamine hypothesis. According to this hypothesis, depression is associated with deficits in neurotransmitters called monoamines, including dopamine, epinephrine, norepinephrine, and serotonin. Different types of antidepressants work in different ways to regulate concentrations of monoamines in neural synapses.

`[P4]` Technical terms in each sentence now appear at the end. `[P3]` Each sentence opens with familiar information carried from the previous sentence.

---

## Principle 4: Complexity Last — Paragraph Opening Sentences

**Pair 13** — Topic of paragraph buried in sentence opening

Before:
> The political situation changed, because disputes over succession to the throne plagued seven of the eight reigns of the Romanov line after Peter the Great.

After:
> The political situation changed, because after Peter the Great seven of the eight reigns of the Romanov line were plagued by turmoil over disputed succession to the throne.

`[P4]` The key term *disputed succession to the throne* is moved to the sentence's end. This is the paragraph's topic; the sentences that follow develop it. The "before" version buries that key term in the middle, weakening the paragraph's opening frame.

---

## Principle 1 + 2 Combined: Abstract Main Character, Excessive Abstraction

**Pair 14** — Abstract main character surrounded by additional nominalizations

Before (abstract subject + pileup of additional nominalizations):
> Our expectation is that greater institutionalization in the political sphere will be of benefit to older democracies. Although definition of political institutionalization is difficult, there seems to be general consensus that functional differentiation, regularization (and hence predictable), professionalization (including meritocratic methods of recruitment and promotion), rationalization (explicable, rule based, and nonarbitrary), and the infusion of value are characteristic of procedures in a well-institutionalized polity. This description is a fit for most long-standing democracies.

After (abstract subject maintained, but surrounding nominalizations removed):
> We expect that older democracies will benefit from greater institutionalization in the political sphere. Although political institutionalization is difficult to define, there seems to be general consensus that procedures in a well-institutionalized polity are functionally differentiated, regularized (and hence predictable), professionalized (including meritocratic methods of recruitment and promotion), rationalized (explicable, rule based, and non-arbitrary), and infused with value. Most long-standing democracies fit this description.

`[P2]` *expectation* → *expect*, *definition* → *define*, *benefit* → *benefit (kept, already a verb)*, adjective nominalizations → predicate adjectives. Note: *institutionalization* is kept throughout because it is the text's main concept. `[P1]` *Our expectation* → *We*.

---

## Two-Style Comparison (Full Paragraphs)

**Pair 15** — Three style levels: opaque academic, oversimplified, and clear collegial

Style A (opaque — dense nominalizations, abstract subjects):
> Conventional management practice assumes that interaction and collaboration enhance organizational performance by improving employee creativity and productivity. But unless collaboration is punctuated by isolation, and unless workspace configurations provide isolation opportunities, erosion rather than enhancement of organizational effectiveness may result.

Style B (oversimplified — short sentences, but childlike):
> Managers want the people who work for them to interact and collaborate. When they do this, they become more creative and productive. The organization then performs better. But people also need opportunities to work alone, and workplaces need to provide these opportunities. Otherwise, the organization may become less effective.

Style C (clear and collegial — correct balance):
> Managers conventionally assume that when employees interact and collaborate, they become more creative and productive, thus leading the whole organization to perform better. But unless employees also have opportunities to work alone, and unless workspaces are configured to provide them, the organization may become less rather than more effective.

Style C succeeds because: `[P1]` *Managers* and *employees* are the subjects. `[P2]` *interaction/collaboration* → *interact and collaborate*, *enhancement* → *more effective*. `[P3]` Each sentence begins with familiar context. Style B is also clear but sounds simpleminded for adult professional writing — clarity does not mean short, choppy sentences.

---
name: data-visualization-selector
description: "Select the correct graphic type (table, bar chart, line graph) for a dataset and rhetorical goal, then design and frame it to communicate evidence clearly and honestly. Use this skill whenever the user needs to present quantitative data in a research paper, report, thesis, presentation, or professional document and asks: which chart should I use, how should I visualize this data, how do I make this graphic clearer, is my chart misleading, how do I label or title a table or figure, or how do I introduce a graphic in text. Also triggers on: \"my advisor said the table is confusing,\" \"should I use a bar chart or line graph,\" \"how do I make readers see my point in this figure,\" \"is this graph ethical,\" \"my chart looks amateurish,\" \"the scale on my graph looks off,\" or any request to improve the visual communication of numerical evidence. Covers the full workflow: verbal-vs-visual decision → graphic type selection based on rhetorical effect → design simplification → framing with title, intro sentence, and labels → ethical integrity checks."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/data-visualization-selector
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [15]
tags: [research-methodology, data-visualization, academic-writing, evidence-presentation, tables, charts, graphs, research-ethics]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: data
      description: "The dataset, table, or graphic the user wants to present or improve — pasted as text, described verbally, or provided as a file"
    - type: claim
      description: "The argument or point the graphic is meant to support — what readers should conclude from it"
  tools-required: []
  tools-optional: [Read, Write]
  mcps-required: []
  environment: "Any agent environment. Works on described data, pasted tables, or referenced files."
---

# Data Visualization Selector

## When to Use

The user has quantitative data and needs to present it visually in a research paper, report, thesis, or presentation. Typical triggers:

- User has a dataset and asks "should I use a table or a chart?"
- User shares an existing graphic and asks why it looks confusing or amateurish
- User asks how to title, label, or introduce a graphic
- An advisor or reviewer said a figure was misleading, hard to read, or poorly framed
- User has multiple data points they are struggling to fit into prose
- User suspects their chart may be exaggerating or misrepresenting their findings

**Scope:** This skill covers the selection, design, and ethical framing of tables, bar charts, and line graphs — the three graphic types essential for basic quantitative research. For scatter plots, heat maps, box plots, or advanced statistical graphics, use this skill's decision logic as a foundation and consult field-specific visualization resources.

**Do not use this skill to select graphics for decorative purposes.** Every graphic must directly support a specific argument or claim.

## Context

### Required (must be present before proceeding)

- **The data:** What values are being presented? How many data points? What variables?
- **The argument:** What claim should readers draw from the graphic? What should they notice?

### Useful but not required

- **Audience and venue:** Specialist readers in your field vs. general audience? Print vs. screen? Black-and-white vs. color?
- **Existing graphic (if revising):** Share it as text, a description, or a file so this skill can diagnose specific problems.

---

## Stage 1 — Verbal or Visual?

Before selecting a graphic type, decide whether visualization is necessary at all.

**Use prose (no graphic) when:**
- You have 3 or fewer numbers to present. Readers can hold a small set of values in a sentence as easily as in a table.
- Example: *"In 2013, men earned $50,033 a year and women $39,157, a difference of $10,876."* This sentence communicates the same information as a 3-row table — and takes less space.

**Use a graphic when:**
- You have more than a few numbers and prose forces readers to track more values than working memory allows.
- The relationships or patterns in the data are the point — readers need to see structure, not just remember numbers.
- The visual comparison is itself the argument (e.g., a dramatic difference in magnitude).

**Why this decision matters:** Graphics carry an implicit claim of complexity. A table with two rows signals to readers that the data required a table — if it did not, the graphic looks inflated and the writer looks like they are padding the paper with visual bulk.

---

## Stage 2 — Select the Graphic Type

The three primary graphic types each produce a distinct rhetorical effect. Choose the effect that matches your argument.

### Table

**Rhetorical effect:** Precise, objective, neutral. Presents discrete numbers and lets readers draw their own comparisons.

**Use when:**
- Exact values matter and readers will need to look up or compare specific numbers
- You have many variables across multiple categories and readers may need to extract different values depending on their question
- The argument requires readers to see all the data, not just a highlighted trend

**Example trigger:** "Here are the unemployment rates for nine countries across two time periods" — readers may want to check any specific country, so exact numbers in a table serve them better than a simplified visual.

**Trade-off:** A table requires readers to infer relationships themselves (unless you state them in an introductory sentence). A table makes your data available; it does not automatically make your argument visible.

---

### Bar Chart

**Rhetorical effect:** Visual contrast. Emphasizes differences in magnitude among discrete items.

**Use when:**
- The argument is a comparison — one group or item is larger, smaller, or different from others
- Bars arranged to match your point (e.g., sorted by size, grouped by category) directly support the claim
- You want readers to see relative sizes at a glance rather than read exact numbers

**Example trigger:** "Most of the world's deserts are concentrated in North Africa and the Middle East." A bar chart with deserts grouped by region (rather than listed alphabetically) makes that geographic concentration visible as a visual pattern — readers do not have to read and calculate. An alphabetical listing of the same values produces no coherent image.

**Trade-off:** Bar charts communicate less precisely than tables. Readers see approximate magnitudes, not exact values. Add numbers to bars when exact values also matter.

**Variants:**
- *Stacked bars:* Use only when you want readers to compare total bar heights across groups; do not use when the argument depends on comparing individual segment sizes (readers cannot judge segment proportions accurately by eye).
- *Clustered bars:* Use when segment sizes are as important as totals. Group and order logically; label group totals.
- *Pie charts:* Avoid for quantitative data — readers cannot accurately judge proportional areas. Acceptable only for qualitative impressions (e.g., "one segment is disproportionately larger than all others combined"). Use bar charts instead for quantitative comparisons.

---

### Line Graph

**Rhetorical effect:** Continuous change over time. Suggests trend, trajectory, and movement.

**Use when:**
- Time is one axis and you want to argue that something changed, increased, or declined
- The argument is about direction and rate of change, not specific values at points in time
- Multiple trends need to be compared visually over the same period

**Example trigger:** "Two-parent households declined steadily from 1970 to 2010 while mother-headed households rose." The line graph conveys this crossing trajectory in a single image; a table of the same values requires readers to read across rows and do arithmetic to see the trend.

**Trade-off:** A line graph implies continuous change between plotted points. If your data are categorical, intermittent, or collected at uneven intervals, a line graph misleads by implying smooth progression. Use a bar chart for discrete items, even if time is involved.

**Variable direction note:** Choose the variable that makes the line travel in the direction that supports your argument. If a reduction is good news, you may represent the same data as an increase in the opposing variable (e.g., retention rate instead of dropout rate) to produce a rising line that carries positive visual force.

---

## Stage 3 — Design the Graphic

Once the type is selected, apply these design principles to produce a clear, readable graphic. Resist default software settings — most produce graphics that look complex without communicating clearly.

### 3.1 Keep It Simple

**Include only data relevant to your argument.** If you have additional data to report for completeness, label it separately and move it to an appendix.

**Why:** Readers assume every element of a graphic is there to support your point. Irrelevant data confuses them about what they should conclude.

**For tables:**
- Use horizontal lines only (top border, bottom border, header separator). Do not use a grid of horizontal and vertical dark lines — it creates visual noise.
- For long tables, shade every fifth row lightly to help readers track rows across columns.
- Do not add background color or shading.

**For charts and graphs:**
- Box a graphic only if you are grouping two or more figures together.
- Do not color or shade the background.
- Use background grid lines only if the graphic is complex or readers need to read precise values from it. Make grid lines light gray.
- Use color or shading only to show a contrast. Verify the graphic is still readable in black-and-white (avoid distinguishing elements by color alone if the document may be photocopied).
- Never use iconic bars (e.g., images of cars to represent automobile production) — they distort how readers judge values and look amateurish.
- Do not add a third dimension (3D effect) for visual appeal. Three-dimensional bars and pie charts make proportions harder to judge, not easier.

### 3.2 Use Clear Labels

**Label everything:** All rows and columns in tables; both axes in charts and graphs; all lines in line graphs.

**For graphs and charts:**
- Use tick marks and labels to indicate intervals on the vertical axis.
- Label lines and bar segments directly on the graphic image when possible. Use a legend only when direct labeling would make the graphic too cluttered to read.
- When exact values matter, add them to bars (above or inside) or to dots on line graphs.

**Why:** A graphic that requires readers to scan between image and legend interrupts reading and increases the chance of misidentifying elements. Direct labeling keeps interpretation in the image.

### 3.3 Order Data to Match Your Argument

**Tables:** Order rows and columns by the principle that lets readers quickly find the comparison you want them to make. Do not default to alphabetical order unless the argument has no inherent ordering.

**Bar charts:** Arrange bars to create a visual image that matches your claim. Bars in no particular order imply no particular point.

**Example:** A bar chart of desert sizes listed alphabetically makes no argument. The same data sorted by region (North Africa, Middle East, Australia, South Africa, North America) immediately supports the claim that deserts cluster in certain geographic zones.

**Tables — rounding:** Round to the precision that makes differences visible. If differences smaller than 1,000 do not matter to your argument, then 2,123,499 is misleadingly precise — use 2,123,000 or 2.1 million. Place totals at the bottom of columns or the end of rows, not at the top.

---

## Stage 4 — Frame the Graphic in Text

A graphic does not speak for itself. You must frame it with three elements that tell readers what to see and why it matters.

### 4.1 Provide a Descriptive Title

**Tables:** Title goes above the table, flush left.
**Figures (charts, graphs):** Legend goes below the figure, flush left.

**Rules for titles and legends:**
- Describe the data specifically enough to distinguish this graphic from every other in the document.
- Do not use a general topic as a title: **Not** *Heads of households* — **Use** *Changes in one- and two-parent heads of households, 1970–2010*
- Do not editorialize in the title — do not characterize what the data imply: **Not** *Weaker effects of counseling on depressed children before professionalization of staff, 1995–2004* — **Use** *Effect of counseling on depressed children, 1995–2004*
- Make the title descriptive, not argumentative. The argument belongs in your introductory sentence (see 4.2).

**Why:** Titles serve as identifiers when readers scan, search for, or cite a specific figure. An argumentative title permanently embeds your interpretation into the graphic's identity and prevents a reader from drawing an independent conclusion from the data.

### 4.2 Write an Introductory Sentence

Before every graphic, write one sentence that tells readers what to see in it and how it supports your argument. Do not assume readers will draw the right conclusion.

**Structure:** State the claim you want the graphic to support, then refer readers to the figure.

**Without framing:** *Table 15.3. Gasoline consumption.* (Readers must figure out why it is here.)

**With framing:** *Gasoline consumption has not grown as predicted. Though Americans drove 23 percent more miles in 2000 than in 1970, they used 32 percent less fuel. [See Table 15.3.]*

**Why:** Readers approach a graphic already looking for whatever the preceding text suggested they would find. Without a framing sentence, readers do not know which relationship in the graphic to focus on — they look everywhere and conclude nothing specific. A good framing sentence makes the reader's eye go directly to the key number, trend, or relationship.

### 4.3 Add Internal Emphasis When the Data Is Dense

If a graphic contains many values but only a few are directly relevant to your argument, highlight those values internally — for tables, shade the relevant rows or columns; for graphs, annotate key inflection points.

**Example:** A table showing per-capita mileage and fuel consumption across four decades becomes readable when the percentage-change column is shaded and labeled, directing attention to the relevant comparison. Without shading, readers scan all 16 cells equally.

---

## Stage 5 — Check for Ethical Integrity

A graphic that is technically accurate can still mislead. Before finalizing, run these four checks:

**Check 1 — Scale manipulation.** Does the vertical axis begin at zero? A truncated scale (starting at a non-zero value) compresses the visible range and exaggerates contrasts. A bar chart where bars appear to show a 50% reduction may actually represent a drop from 94 to 90 on a scale starting at 80. The data are accurate; the visual impression is not.

- Rule: Do not manipulate scale to magnify or reduce a contrast beyond what the data warrant.

**Check 2 — Dual-axis false correlation.** Does your line graph use two different vertical scales on left and right axes? Two independent variables measured on different scales will appear to move together if their scales are calibrated to overlap visually. This creates the impression of correlation or causation where neither has been established.

- Rule: Do not use dual-axis graphs to suggest a relationship between variables unless you have established the relationship through analysis, not visual calibration.

**Check 3 — Figure type distorting values.** Does the figure's geometry accurately represent the data? Stacked area charts, 3D graphics, and certain pie chart configurations can make the same data appear to show different patterns depending on which elements are placed at the top or bottom.

- Rule: Do not use a figure whose image distorts the values it displays.

**Check 4 — Missing argument statement.** If the graphic supports a specific point, have you stated that point? Presenting data without a claim allows readers to draw unintended conclusions.

- Rule: If the table or figure supports a point, state it. Silence is not neutrality — it is an invitation to misinterpret.

---

## Examples

### Example 1: Verbal vs. Visual Decision

**Data:** Average annual salaries for men ($50,033) and women ($39,157) in 2013, with a difference of $10,876.

**Decision:** Three numbers. Prose handles this without a graphic — *"In 2013, men earned $50,033 a year on average and women $39,157, a difference of $10,876."* A table here adds visual bulk without adding comprehension.

**But:** Add four time periods and two additional family types across 40 years (the family structure data), and the same logic flips — 20+ numbers exceeds what readers can hold in working memory from prose. A table or chart is now necessary.

---

### Example 2: Bar Chart vs. Line Graph for the Same Data

**Data:** Percentages of U.S. families by type (two-parent, mother-headed, father-headed, no-adult-headed) across five decades (1970–2010).

**If the argument is:** "Family structure changed dramatically over 40 years, with two-parent households declining steadily while mother-headed households tripled."
- Use a **line graph**: it visualizes continuous directional change and allows all four trends to be compared over time on a single image.

**If the argument is:** "In any given year, different family types had dramatically different prevalence rates."
- Use a **bar chart**: it emphasizes magnitude contrasts at a given point in time (contrast among discrete items within each year's cluster).

**Design note (bar chart):** Do not sort bars alphabetically or by family-type name. Group them by year, then order the bars within each group by size — largest first — to make the dominant category immediately visible.

---

### Example 3: Detecting and Correcting an Ethical Violation

**Scenario:** A bar chart shows a city's pollution index declining from 101 to 90 over twelve years, presented as evidence of significant environmental improvement.

**The problem:** The vertical axis starts at 80 (not 0). The resulting bars appear to shrink by roughly half their height, creating the visual impression of a ~50% reduction. The actual decline is 101 to 90 — about 11% on the full scale.

**Correction:** Start the vertical axis at 0. The bars will now show a modest, accurate decline. If the decline is real and meaningful, state it in the introductory sentence: *"Pollution fell 11 percent over twelve years, from 101 to 90."* Do not rely on a distorted visual to inflate the apparent magnitude.

---

## Quick-Reference: Graphic Type Selection

| Your argument | Data characteristics | Best type |
|---|---|---|
| Present all values for reference or lookup | Many exact numbers | Table |
| Show difference in magnitude between discrete items | Categories, no time axis | Bar chart |
| Show trend or change over time | Continuous time axis | Line graph |
| Show parts of a whole (totals matter most) | Whole + proportional parts | Stacked bar |
| Show proportion is disproportionately large/small | One segment dominates | Pie (qualitative only) |

---

See `references/graphic-type-rhetorical-effects.md` for an extended table of rhetorical uses for additional graphic types (scatter plots, frequency distributions, box plots) used in advanced research contexts.

See `references/ethical-visualization-checklist.md` for a printable checklist of all five ethical integrity checks.

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

This skill is standalone. Browse more BookForge skills: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/ethical-visualization-checklist.md
# Ethical Visualization Checklist

Use this checklist before finalizing any graphic in a research paper, report, or thesis. A graphic can contain accurate data and still mislead through design choices. These checks catch the most common forms of visual misrepresentation.

Source: The Craft of Research, 4th Ed., §15.5 — Communicating Data Ethically.

---

## The Four Rules

1. Do not manipulate a scale to magnify or reduce a contrast.
2. Do not use a figure whose image distorts values.
3. Do not make a table or figure unnecessarily complex or misleadingly simple.
4. If the table or figure supports a point, state it.

---

## Check 1 — Axis Scale (Truncation)

**Question:** Does the vertical (value) axis begin at zero?

- [ ] If yes: proceed.
- [ ] If no: is there a legitimate field convention for starting at a non-zero value (e.g., temperature scales, index benchmarks)? If yes, label the truncation explicitly.
- [ ] If no legitimate reason: reset axis to zero. Recheck whether the argument still holds. If the argument depended on the compressed visual, state the actual magnitude of change in the introductory sentence.

**The problem:** A bar chart with a y-axis starting at 85 instead of 0 can make a drop from 101 to 90 appear to represent a 50% reduction. The data are accurate; the visual impression is false.

---

## Check 2 — Dual-Axis Lines (False Correlation)

**Question:** Does your graph use two different vertical scales — one on the left axis, one on the right — for two different variables?

- [ ] If no dual axis: proceed.
- [ ] If dual axis present: have you established the relationship between these two variables through analysis (regression, correlation coefficient, causal mechanism)? If yes, label both axes clearly and note the relationship in the text.
- [ ] If no established relationship: do not use dual-axis display. Two variables that merely "trend together" on a dual-axis graph imply causation by visual proximity, not by evidence.

**The problem:** A dual-axis graph showing union membership (left axis, 13–17%) and unemployment rate (right axis, 4–8%) over the same period can be calibrated so both lines trace nearly identical paths. This implies they move together causally. But the scales are different — any two variables that decline over the same period will appear to correlate on a dual-axis graph if you calibrate the axes to match.

---

## Check 3 — Figure Geometry Distorting Values

**Questions to check:**

- [ ] **3D graphics:** Are you using a three-dimensional rendering of a bar, pie, or area chart? 3D rendering creates depth illusions that make front bars appear larger than back bars of equal value. Remove the 3D effect.
- [ ] **Stacked area chart ordering:** In a stacked area chart, have you placed the most volatile band at the bottom? If a bottom band rises sharply, all bands above it will appear to rise as well, even if their individual values are stable. Place the most stable bands at the bottom.
- [ ] **Iconic bars:** Are bars replaced with images (e.g., stacked coins, car icons)? Iconic bars distort magnitude perception. Replace with plain bars.
- [ ] **Pie chart precision:** Are you using a pie chart to communicate precise proportions between segments of similar size? Switch to a bar chart.

---

## Check 4 — Missing Argument Statement

**Question:** For every table and figure in the document, is there an introductory sentence that states the claim the graphic supports?

- [ ] Every graphic has an introductory sentence before it.
- [ ] The introductory sentence states the argument — what readers should conclude — not just "Table 3 shows..."
- [ ] The introductory sentence highlights the specific values or relationships readers should look for.

**Why silence misleads:** A graphic without an introductory argument statement invites readers to draw any conclusion the data could support — including incorrect ones. The author's silence does not signal neutrality; it signals that the point was too obvious to state, or that there is no point, or (worst) that the author does not know what the data mean.

---

## Check 5 — Complexity and Simplicity Balance

**Question:** Is the graphic as simple as its content allows, but not simpler?

- [ ] Does the graphic include only data relevant to the argument? (Extraneous data: label separately and move to appendix.)
- [ ] Does the graphic display enough data that readers can verify the claim? (A graphic with only two data points selected to support a trend claim that exists across twenty data points is misleadingly simple.)
- [ ] Are all labels, axes, and titles present and specific? (Missing labels force readers to guess what they are looking at.)

---

## Quick Self-Test

Before submitting, ask:

1. If a hostile reader examined only my graphic — not my text — could they use a design choice in it to argue I was exaggerating or suppressing the contrast? If yes, fix the design choice and explain the actual magnitude in text.

2. If I described what this graphic looks like to someone who cannot see it, would my description match what the data actually say? If not, the figure's image is distorting the values.

3. Have I stated in prose what I want readers to conclude from this graphic? If not, add an introductory sentence.

FILE:references/graphic-type-rhetorical-effects.md
# Graphic Type Rhetorical Effects

Extended reference for choosing among graphic types beyond the three core types (table, bar chart, line graph). Use the core types when possible. Add types from this table only when your field expects them, your readers are familiar with them, or the data cannot be displayed any other way.

Source: The Craft of Research, 4th Ed., Chapter 15.

---

## Core Types (Covered in Main Skill)

| Type | Primary Rhetorical Effect | When to Use | When to Avoid |
|---|---|---|---|
| Table | Precise, objective; emphasizes discrete numbers | Exact values matter; many variables; readers need lookup access | Argument is about trend or magnitude contrast |
| Bar chart | Visual contrast among discrete items | Comparing magnitudes across categories; argument depends on visible size differences | Data are continuous over time; argument is about trend |
| Line graph | Continuous change; trend and trajectory | Time-series data; directional change is the argument | Data are categorical or intermittent; implying continuity would mislead |

---

## Extended Types

| Type | Primary Rhetorical Effect | When to Use |
|---|---|---|
| Stacked bar | Total values plus proportional composition | Readers need to compare whole-bar totals AND see composition; use only when total matters as much as segments |
| Clustered bar | Segment-by-segment comparison across groups | Segment sizes are as important as totals; comparisons across groups are the argument |
| Scatter plot | Correlation or distribution between two continuous variables | Showing a relationship between two quantitative variables; plotting individual data points for pattern detection |
| Frequency distribution / histogram | Distribution shape of a single variable | Showing how values cluster, spread, or concentrate; argument is about distributional properties (skew, outliers, modes) |
| Box plot | Median, spread, and outlier structure | Comparing distributions across groups; arguing about variability as well as central tendency; familiar to statisticians and quantitative social scientists |
| Pie chart | Qualitative impression of proportional dominance | Showing that one segment is disproportionately larger, or that data breaks into many small parts; NOT for quantitative comparison — readers cannot judge segment proportions accurately |
| Area chart (stacked) | Cumulative totals over time with composition visible | Showing how a total changes over time while displaying how its composition shifts; place the most stable bands at the bottom to avoid visual distortion |
| Heat map | Intensity across two dimensions | Showing patterns in a matrix where color encodes magnitude; fields: genomics, geography, behavioral data |
| Dot plot | Discrete comparisons with less visual weight than bars | Same use as bar chart; preferred in some scientific fields for avoiding the ink-to-data ratio problem of filled bars |

---

## Notes on Pie Charts

Pie charts are widely used in journalism, annual reports, and business presentations. They have two legitimate uses in research:

1. **Dominance impression:** Showing qualitatively that one segment is so large it dwarfs all others combined.
2. **Fragmentation impression:** Showing that data is divided into many small segments with no dominant category.

For any other use — comparing proportions, showing change, presenting exact values — use a bar chart. Readers cannot accurately judge the comparative sizes of circular segments, especially when segments are similar in size.

---

## Checklist: Selecting Among Extended Types

1. Is your data continuous over time? → Line graph.
2. Are you comparing discrete magnitudes? → Bar chart.
3. Do readers need exact numbers? → Table.
4. Is composition within a total important? → Stacked or clustered bar.
5. Are you showing a relationship between two continuous variables? → Scatter plot.
6. Is the argument about how a variable distributes? → Histogram or box plot.
7. Nothing else works and your readers are familiar with it? → Extended type from the table above.

---

## Common Default-Software Traps

Software defaults are optimized for visual appeal, not rhetorical clarity. Override these defaults:

| Default Setting | Problem | Override |
|---|---|---|
| Alphabetical axis ordering | Implies no argument; bars in no logical sequence | Order by the principle that shows your point (size, category, region, etc.) |
| 3D chart rendering | Distorts proportions; makes values harder to judge | Use flat (2D) rendering |
| Colored background | Visual noise that distracts from data | Set to white/transparent |
| Thin gridlines in full gray or black | Creates visual clutter | Remove or set to light gray, only if needed |
| Automatic axis starting point | May truncate at non-zero value | Check and set to zero unless the field convention requires otherwise |
| Legend instead of direct labels | Forces eye to travel away from image to identify elements | Label directly on the image when possible |
| Iconic bars (images instead of bars) | Distorts magnitude judgment; appears amateurish | Use plain bars with explicit values |

---
name: counterargument-handler
description: Anticipate, acknowledge, and respond to reader objections and alternative views in a research argument. Use this skill when the user has a draft argument or storyboard and needs to identify which objections readers will predictably raise, wants to decide which objections to acknowledge and which to set aside, needs vocabulary and sentence templates for introducing and responding to counterarguments without weakening their position, has discovered a flaw in their argument and does not know how to handle it honestly, is building a cause-and-effect argument and needs to address competing causes, has made claims with counterexamples that readers will invoke, uses key terms that readers may define differently and needs to address definitional scope, or wants to avoid either ignoring objections (seeming ignorant) or acknowledging too many (losing argumentative focus). This skill is the detailed companion to research-argument-builder — use it after assembling the argument's core structure (claim, reasons, evidence) and before drafting, to map every acknowledgment slot with a calibrated response strategy.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/counterargument-handler
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [10]
tags: [research-methodology, academic-writing, argumentation, counterarguments, critical-thinking]
depends-on: [research-argument-builder]
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "User's research argument storyboard, draft, or claim+reasons+evidence summary from research-argument-builder, plus the research question the argument answers"
  tools-required: [Read, Write]
  tools-optional: []
  mcps-required: []
  environment: "Any agent environment; user should have a core argument assembled — at minimum a claim and at least two reasons with supporting evidence"
discovery:
  goal: "Map every predictable objection to the user's argument, decide which to acknowledge and how, then produce calibrated acknowledgment-response pairs using precise vocabulary — so that the final draft shows readers the argument has already been through its own wringer"
  tasks:
    - "Run the objection anticipation protocol against each element of the argument (problem definition, claim, solution type, evidence)"
    - "Identify which of the three predictable disagreement types apply: competing causes, counterexamples, or definitional scope disputes"
    - "Apply the Goldilocks filter to decide which objections to acknowledge, which to pass, and which to set aside"
    - "Select an appropriate response strategy for each acknowledged objection: rebut, qualify, or concede-and-reframe"
    - "Choose acknowledgment and response vocabulary calibrated to the weight given the objection"
    - "Produce an acknowledgment map with draft sentences ready for insertion into the argument"
  audience: "Researchers, graduate students, analysts, and professionals who have assembled a research argument and need to address counterarguments before drafting or revising"
  triggers:
    - "User has a structured argument but has not considered what skeptical readers will object to"
    - "User's draft ignores competing views or alternative interpretations"
    - "User knows readers will push back but does not know what to say or how to say it"
    - "User argues from a single cause but worries readers will cite other causes"
    - "User makes a broad claim and needs to handle counterexamples without abandoning the claim"
    - "User's argument depends on a term that readers may define differently"
    - "User has found a genuine flaw in their argument and does not know whether to hide it, concede it, or reframe it"
    - "User's draft acknowledges so many objections that the main argument gets lost"
---

# Counterargument Handler

## When to Use

You have assembled the core of your argument — a claim backed by reasons and evidence. Now you need to do what separates a thin argument from a persuasive one: anticipate what skeptical readers will say, decide which objections to address, and respond in a way that is honest, precise, and targeted.

An argument that presents only its own case looks one-sided. Readers who hold different views will feel dismissed, not persuaded. But an argument that acknowledges every possible objection loses focus and exhausts readers. The goal is calibrated engagement: acknowledge what readers care about, respond in proportion to the weight of the objection, and direct criticism at ideas rather than people.

This skill covers the full acknowledgment-response workflow:

1. **Anticipate** — question your argument as your most skeptical readers will
2. **Identify** — recognize the three types of disagreements that are almost always predictable
3. **Filter** — decide what to acknowledge (Goldilocks: not too many, not too few)
4. **Choose a strategy** — rebut, qualify, or concede-and-reframe
5. **Write it** — use vocabulary calibrated to the weight you give each objection

**Preconditions to verify:**

- Does the user have a core argument (claim + reasons + evidence)? If not, invoke `research-argument-builder` first.
- Can the user state their main claim in one or two sentences? If not, that is a sign the argument itself is not yet assembled — return to `research-argument-builder`.

## Context and Input Gathering

### Required (ask if missing)

- **The main claim:** What the argument asserts. One or two sentences.
  - Check prompt for: thesis statement, main conclusion, storyboard from `research-argument-builder`
  - If missing, ask: "What is the main point you want readers to accept?"

- **The reasons and any evidence summary:** The logical structure of the argument.
  - Check prompt for: storyboard output, outline, or any ordered set of assertions and sources
  - If missing, ask: "What are the two or three main reasons your claim is true?"

### Useful (gather from environment if available)

- **Sources or bibliography:** Other researchers' positions reveal what alternatives and objections exist in the field
  - Look for files like `sources.md`, `bibliography.md`, `notes.md` in the working directory
  - If found, scan for positions that differ from the user's claim before asking the user to generate objections from memory

- **Research question and intended audience:** Determines which objections are most salient
  - If not stated, ask: "Who are the readers — what is their field, level, and likely stake in a different outcome?"

## Process

### Step 1 — Run the Objection Anticipation Protocol

**WHY:** You know your argument too well to see its weaknesses clearly. You believe in it too much to challenge it seriously. The only way out is to deliberately adopt the perspective of someone who wants you to be wrong — a skeptical reader with a stake in a different outcome. Doing this systematically before readers actually raise objections gives you time to fix real problems and to prepare responses for apparent ones.

Work through each element of the argument in order. For each, ask the questions below and note every objection you cannot immediately dismiss:

**On the problem definition:**
- "Why do you think there's a problem at all? The costs or consequences don't seem that significant."
- "Have you defined the problem correctly? Maybe it involves a different issue than the one you raise."

**On the claim:**
- "Have you stated this too strongly? I can think of exceptions and limitations."
- "Is your solution practical or conceptual — and does it match the type of problem you've identified?" (Practical problems need practical solutions; conceptual problems need conceptual ones.)

**On why your solution is better than alternatives:**
- For a practical claim: "What you propose will cost too much and create new problems."
- For a conceptual claim: "This doesn't fit with other well-established knowledge."

**On the evidence:**
- Nature: "I want a different sort of evidence — hard numbers, not anecdotes (or lived experience, not cold numbers)."
- Accuracy: "The numbers don't add up."
- Precision: "What do you mean by 'many'? You need to be more specific."
- Currency: "There is newer research on this."
- Representativeness: "You didn't get data on all the relevant groups."
- Authority: "Your source is not an expert on this."
- Sufficiency: "One data point is not enough to support this."

After this pass, sort your objections into two lists:
- **Fixable before you draft:** Genuine weaknesses in the argument — fix them now, do not just acknowledge them.
- **Apparent but not real, or unfixable:** Objections you can address with acknowledgment and response.

For any fixable weakness: go back and fix it. Acknowledging a flaw you could have fixed signals incompetence; hiding it signals dishonesty. Neither serves you.

### Step 2 — Identify the Three Predictable Disagreements

**WHY:** Three kinds of alternatives appear in almost every research argument, regardless of field. Recognizing them by type lets you address them systematically rather than being surprised when readers raise them. If you argue about causes, expect competing causes. If you make broad claims, expect counterexamples. If your argument turns on a contested term, expect definitional challenges.

**Disagreement Type 1 — Competing causes**

Applies when: your argument is about cause and effect.

No effect has a single cause. If you argue that X causes Y, every reader will think of other causes. You do not need to defeat every competing cause — but you need to acknowledge the most prominent ones and explain either why they are less important than your cause or why your cause is the focus of this argument.

Example: Arguing that pesticide use is causing honeybee colony collapse. Predictable competing causes: habitat loss, disease, parasites, genetically modified crops. Acknowledge the most prominent and explain your scope.

**Disagreement Type 2 — Counterexamples**

Applies when: you make any generalization.

Readers will think of exceptions they believe undermine your claim. You must think of them first, acknowledge the more plausible and vivid ones, and explain why you do not consider them as damaging as readers might. Be especially vigilant when your claim covers a phenomenon with wide natural variation — readers who do not understand statistical reasoning will focus on aberrant cases.

Example: Arguing about climate warming trends. A reader might cite a cold summer in one location. Acknowledge that local variation exists; explain why it does not invalidate a trend across decades and geographies.

**Disagreement Type 3 — Definitional scope disputes**

Applies when: your argument depends on a key term that has more than one plausible meaning.

Readers must accept your definition to accept your claim. They will often redefine terms to suit their own views. When your argument hinges on a term, define it with a supporting sub-argument — not a dictionary citation — and acknowledge plausible alternative definitions. Explain why you have adopted your definition.

Example: Arguing that social media is "addictive." Readers may define addiction as requiring physical dependence; you may mean behavioral compulsion. Acknowledge the definitional dispute; argue for the definition your evidence supports; explain why alternative definitions do not fit your research context.

Do NOT cite a dictionary as your definitional authority. Dictionaries record usage; they do not resolve contested meaning in specialized arguments.

### Step 3 — Apply the Goldilocks Filter

**WHY:** Too many acknowledgments lose the argument. Too few make you seem ignorant of or dismissive toward your readers' views. The filter is not about being nice to every opposing idea — it is about strategic, credibility-building engagement with the objections readers actually care about.

Prioritize acknowledging the following — in roughly this order of importance:

1. **Plausible charges of weakness you can rebut.** These are objections that seem serious at first but are not — acknowledging and rebutting them actively strengthens your case.
2. **Alternative lines of argument that are important in your field.** Failing to engage with the major alternative positions in your field signals that you have not done the research.
3. **Alternative conclusions that readers want to be true.** Readers with a stake in a different outcome need to see you have considered their preferred position and can explain why yours is better.
4. **Alternative evidence readers know about.** If readers know of a study or data set that points in a different direction, ignoring it looks deliberate.
5. **Important counterexamples you must address.** Vivid, memorable exceptions that readers will immediately think of.

**What to set aside:** Objections that are obviously frivolous, easily dismissed without argument, or so minor that noting them at all inflates their importance. Better to ignore what readers like than to disparage it — disparaging it weakens your credibility; ignoring something minor wastes no one's time.

**Important constraint on tone:** Do not denigrate those you disagree with in the acknowledgment. Label the work, not the person: write "the evidence here is thin" not "naive researchers have claimed." Save critical characterization for the response portion, and direct it at the argument rather than the author.

### Step 4 — Choose a Response Strategy

**WHY:** The response is not just a counter-claim. Every response is itself a mini-argument — it needs at least one reason, and for substantial objections, it needs its own evidence. Choosing the right strategy before drafting prevents you from either conceding too much (weakening the original argument) or asserting without reasoning (which is not a response, it is just a louder version of the original claim).

Three response strategies, from most to least accommodating:

**Strategy A — Rebut**

Use when: the objection is based on a misunderstanding of your scope, a factual error, or reasoning that does not hold under scrutiny.

Structure: Acknowledge the objection → give the reason it does not apply or is incorrect → provide supporting evidence if readers do not already know the basis for your response.

Minimum form: `[Acknowledgment phrase] + [reason the objection fails or misses the point]`
Full form: `[Acknowledgment phrase] + [reason] + [sub-reason or evidence backing the reason]`

**Strategy B — Qualify**

Use when: the objection identifies a genuine limitation on your claim, but the claim is still valid within a defined scope. Do not fight a legitimate narrowing — concede it, restate the bounded claim, and move on. A scoped claim that holds is worth more than an overbroad claim that does not.

Structure: Acknowledge that the objection applies in certain cases → restate your claim with an explicit qualifier → show that the bounded claim is still significant.

Example: "We acknowledge this pattern does not hold in the earliest periods of industrialization; our claim applies to the period after 1850, when the relevant conditions stabilized."

**Strategy C — Concede-and-reframe**

Use when: the objection identifies a real flaw in your argument that you cannot fix and cannot rebut. This is the hardest case but also the most honest one.

Three moves are available:
- The rest of your argument more than balances the flaw — state this explicitly and show why.
- While the flaw is serious, more research would show a way around it — frame this as a limitation and a future direction.
- While the flaw makes it impossible to accept your claim fully, your argument offers important insight into the question and suggests what a better answer would require — reframe the contribution.

**What you cannot do:** Ignore a flaw that readers will notice. If readers see it and you have not mentioned it, they conclude you lack competence. If they think you saw it and hid it, they conclude you lack honesty. Either outcome is worse than the concession.

### Step 5 — Write the Acknowledgment-Response Pairs

**WHY:** The vocabulary of acknowledgment does work that the logic alone cannot do — it signals to readers exactly how much weight you give an objection. Using the wrong register (too dismissive or too deferential) misrepresents your actual position. Precise vocabulary controls the reader's interpretation of your credibility as an honest, informed interlocutor.

**Acknowledgment vocabulary — five levels, from most to most deferential:**

Level 1 — Downplay: signals you consider the objection minor but worth noting.
- Introduce with: *despite*, *regardless of*, *notwithstanding*
- Or as a subordinate clause: *although*, *while*, *even though*
- Example: "Although some researchers have attributed this pattern to seasonal variation, our analysis of five-year trends..."

Level 2 — Hedged acknowledgment: signals partial validity without full concession.
- Introduce with: *seem*, *appear*, *may*, *could*; adverbs like *plausibly*, *justifiably*, *reasonably*, *surprisingly*
- Example: "This proposal may have some merit, but the evidence for scalability..."

Level 3 — Unnamed source acknowledgment: gives the objection some weight without elevating a specific source.
- "It is easy to think / imagine / claim that..." followed by your response
- "Some evidence might suggest / indicate / point to..." followed by your counter-evidence
- Example: "It is easy to imagine that shorter deadlines improve output quality, but the data on revision depth..."

Level 4 — Named community acknowledgment: gives the objection significant weight by attributing it to researchers or critics in the field.
- "Some / many / a few researchers / critics / scholars have argued that..." followed by response
- "Although some researchers have argued that..., our findings show..."
- Caution: If you attribute a position to a named person or group, do not characterize them negatively before your response. Write "Smith's evidence is important" not "the occasionally careless Smith." Save critique for the response and direct it at the work.

Level 5 — First-person concession: your own voice conceding validity; highest weight given to the objection.
- "I / we understand / know / realize that..., but..."
- "It must / should be admitted / acknowledged / conceded that..., nevertheless..."
- "Granted / Certainly / To be sure / Of course, X has argued that..., however..."
- Example: "Admittedly, our sample does not cover the Southern states, where conditions differ substantially; nevertheless..."

**Response vocabulary:**

Begin the response with a term or phrase that signals the pivot: *but*, *however*, *on the other hand*.

Response registers, from tactful to blunt:

- Tactful: "But I do not quite understand how / I find it difficult to see how / It is not clear to me how X can claim that..."
- Noting unsettled issues: "But there are other issues here... / But there remains the problem of..."
- Claiming irrelevance: "But as insightful as that may be, it ignores / is irrelevant to / does not bear on the issue at hand."
- Challenging the reasoning or evidence: "But the evidence is unreliable / shaky / thin." "But the argument is untenable / weak / confused / simplistic." "But the argument overlooks / ignores / misses key factors."
- Noting incomplete analysis: "Smith's evidence is important, but we must look at all the available evidence." "That explains some of the problem, but it is too complex for a single explanation." "That principle holds in many cases, but not in all."

See `references/acknowledgment-response-vocabulary.md` for the complete vocabulary list organized by register.

### Step 6 — Produce the Acknowledgment Map

**WHY:** The acknowledgment map externalizes the full set of decisions made in Steps 1-5 into a format that can be inserted directly into the argument storyboard or draft plan. Without it, acknowledgment-response decisions made in the planning phase get lost or inconsistently applied in drafting.

Produce one entry per acknowledged objection, using this format:

```
OBJECTION: [state the objection in the reader's voice — one sentence]
Type: [competing causes / counterexample / definitional scope / evidence quality / other]
Weight given: [low / medium / high]
Response strategy: [rebut / qualify / concede-and-reframe]
Acknowledgment phrase: [draft opening clause with vocabulary from Step 5]
Response: [one to three sentences: the reason the objection fails, the qualifier, or the concession + reframe]
Insert point: [where in the argument this acknowledgment should appear — reason #, section name, etc.]
```

After producing the full map, count the acknowledgments. If there are more than four or five for a medium-length argument, apply the Goldilocks filter again — remove entries in the "low weight / easily dismissed" category unless they are known to be prominent in the field.

## Examples

### Example 1 — Competing causes objection in a cause argument

**Argument:** School lunch nutrition policies reduce childhood obesity rates in low-income districts.

**Predictable objection (Type 1 — competing causes):** "Physical activity levels, home food environment, and socioeconomic status all affect obesity rates. You can't isolate the effect of school lunch policy."

**Goldilocks filter:** This is a legitimate alternative and important in the public health field. Acknowledge.

**Strategy:** Qualify — the argument does not claim school lunch policy is the only cause, but that it is one with measurable, independent effects.

**Acknowledgment map entry:**

```
OBJECTION: Physical activity, home food environment, and income level all contribute to
childhood obesity, making it impossible to attribute rate changes to school lunch policy alone.
Type: competing causes
Weight given: high
Response strategy: qualify
Acknowledgment phrase: "Although home food environment and physical activity independently
affect obesity rates..."
Response: "...our analysis controls for both variables across matched district pairs, isolating
the policy effect. We do not claim school lunch policy is the sole cause — our claim is that
it is a statistically significant, independently actionable factor."
Insert point: Immediately after presenting the main causal evidence (Reason 2)
```

---

### Example 2 — Definitional scope dispute

**Argument:** Social media use among adolescents is addictive.

**Predictable objection (Type 3 — definitional scope):** "Addiction requires physiological dependence and withdrawal symptoms. Social media doesn't qualify — you're misusing the term."

**Goldilocks filter:** Definitional objection that, if not addressed, allows readers to dismiss the entire argument. Must acknowledge.

**Strategy:** Rebut the definitional framing — define addiction as behavioral compulsion that persists despite negative consequences, cite behavioral addiction research, explain why physiological dependence is not the defining criterion in current clinical literature.

**Acknowledgment map entry:**

```
OBJECTION: Addiction requires physiological dependence; applying the term to social media
misuses clinical vocabulary and overstates the case.
Type: definitional scope
Weight given: high
Response strategy: rebut
Acknowledgment phrase: "Some researchers maintain that addiction requires physiological
dependence and formal withdrawal symptoms..."
Response: "...however, behavioral addiction — characterized by compulsive use despite
negative consequences and inability to reduce use voluntarily — is now recognized as a
clinical category (American Psychiatric Association, 2013). We use 'addictive' in this
behavioral sense, which social media use has been shown to satisfy across multiple
validated scales."
Insert point: Within the claim definition, before presenting the supporting evidence
```

---

### Example 3 — Genuine flaw requiring concede-and-reframe

**Argument:** Remote work policies increase individual productivity in knowledge-work firms.

**Discovered flaw:** The three datasets used were all from technology companies. The claim cannot be generalized to all knowledge-work firms.

**Strategy:** Concede the scope limitation honestly; reframe the contribution as establishing the pattern in one sector and identifying conditions that future research should test in others.

**Acknowledgment map entry:**

```
OBJECTION: All three studies draw from technology companies. The claim cannot extend to
law firms, financial services, or other knowledge-work sectors.
Type: evidence — representativeness
Weight given: high
Response strategy: concede-and-reframe
Acknowledgment phrase: "It must be acknowledged that our evidence comes entirely from
the technology sector, which may have atypically high rates of autonomous task structures..."
Response: "...this limits the generalizability of our findings. We do not claim this pattern
holds universally; our contribution is to establish the effect under known conditions and
to identify the task-autonomy variable as the most likely mechanism — which future
research can test across other sectors."
Insert point: In the limitations section; also flag in the claim with a qualifier ("in
technology-sector knowledge work firms")
```

---

## Output

Produce the acknowledgment map (one entry per acknowledged objection using the Step 6 format), then provide:

1. **Objection audit summary:** List every objection surfaced in Step 1, noting which are fixable (user should fix before drafting) and which are addressed in the map.
2. **Three predictable disagreement types:** State which of the three types apply to this argument and what the specific objection is in each case.
3. **Goldilocks verdict:** How many objections the map includes, and why those were selected over the ones set aside.
4. **Draft acknowledgment-response sentences:** For each entry in the map, provide a draft sentence or two using vocabulary from Step 5, ready to be inserted into the draft.

If a fixable weakness was discovered in Step 1: flag it clearly before the acknowledgment map and state what needs to be fixed in the argument before drafting proceeds.

## Anti-Pattern Quick Reference

| Anti-pattern | Signal | Fix |
|---|---|---|
| Ignoring obvious alternatives | No acknowledgments in the draft; all competing views absent | Run the objection anticipation protocol systematically; check your sources for disagreements |
| Acknowledging without responding | "Some argue X. As we can see, Y..." — no reason why X is wrong | Every acknowledgment needs at least one reason; for substantial objections, build a sub-argument |
| Asserting instead of responding | "But this is clearly wrong." — no reason given | Add the reason the objection fails, then add evidence if readers need the basis for the response |
| Denigrating the source in the acknowledgment | "Naive researchers have claimed..." "The careless historian X says..." | Label the work, not the person; save criticism for the response portion and direct it at the argument |
| Over-acknowledging | Four or more acknowledgments in a short paper | Apply the Goldilocks filter; remove low-weight entries that are not prominent in the field |
| Hiding a genuine flaw | Known limitation not mentioned; argument overstated | Concede, then reframe: state the limitation, then explain the contribution within those limits |
| Dictionary definition as authority | "According to Merriam-Webster, 'addiction' means..." | Build a sub-argument for your definition; acknowledge alternative definitions; explain your choice |
| Competing causes ignored | Cause argument with no acknowledgment of other causes | Name the most prominent competing causes; explain why your cause is the focus |

## References

- `references/acknowledgment-response-vocabulary.md` — Complete vocabulary list for acknowledgments (5 levels) and responses (4 registers), with sentence templates
- `references/three-predictable-disagreements.md` — Extended guidance on competing causes, counterexample, and definitional scope disputes with field-specific examples

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-argument-builder`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/acknowledgment-response-vocabulary.md
# Acknowledgment and Response Vocabulary

Complete vocabulary for introducing objections and alternatives at five credence levels, and for responding at four registers. Use these templates by selecting the level and register appropriate to the weight you give the objection and the forcefulness of your response.

Source: Booth, Colomb, Williams, Bizup & FitzGerald, *The Craft of Research*, 4th ed., Chapter 10.

---

## Acknowledgment Vocabulary (5 Levels)

### Level 1 — Downplay

Signals: the objection is minor but worth noting; you give it little weight.

**Introduce at the start of a clause:**

- Despite [the objection], our analysis shows...
- Regardless of [the alternative], the evidence points to...
- Notwithstanding [the competing view], we find...

**As subordinate clauses (same register):**

- Although [acknowledgment], [response]...
- While [acknowledgment], [response]...
- Even though [acknowledgment], [response]...

**Examples:**

- "Although Hong Kong is experiencing economic problems, Southeast Asia remains a strong destination for investment."
- "Notwithstanding Congress's claims that it wants to cut taxes, the latest budget proposals suggest otherwise."

---

### Level 2 — Hedged acknowledgment

Signals: the objection may have some validity; you acknowledge possibility without concession.

**Modal verbs:**
- seem, appear, may, might, could, would

**Adverbs:**
- plausibly, justifiably, reasonably, surprisingly, certainly

**Examples:**

- "In his letters, Lincoln expresses what seems to be depression, but those who observed him..."
- "This proposal may have some merit, but we find that..."
- "Plausibly, the shorter deadline improves focus, yet the data on output quality show..."

---

### Level 3 — Unnamed source acknowledgment

Signals: the objection exists in the field or is a reasonable inference; you give it some weight but do not name a specific source.

**Introduce with:**
- It is easy to [think / imagine / say / claim / argue] that...
- Some evidence [might / may / can / could / does] [suggest / indicate / point to / lead some to think] that...
- There is [another / an alternative / a possible] [explanation / line of argument / account / possibility]...

**Examples:**

- "It is easy to claim that taxes should be cut; but there is another possible line of argument..."
- "Some evidence might suggest that the correlation is coincidental, but our experimental controls show..."

---

### Level 4 — Named community acknowledgment

Signals: the objection is held by identifiable researchers, critics, or scholars in your field; you give it significant weight.

**Introduce with a general subject:**
- There are [some / many / a few] who [might / may / could / would] [say / think / argue / claim / charge / object] that...
- Although [some researchers / critics / scholars] have argued that..., our research shows...
- [Some researchers / critics / scholars] have [said / argued / claimed / charged / objected] that...

**Examples:**

- "There are many who would argue that Cuba is not a communist state in the classical sense, but in fact..."
- "Although some researchers have argued that formal writing instruction hinders natural development, our study of revision processes shows..."

**Caution on naming specific people:** If you name a person or group, do not characterize them negatively before your response. Write the characterization of their work in the response, not in the acknowledgment. Compare:

- Weakens your case: "Some **naive** researchers have claimed that..." "The **occasionally careless** historian Smith has even argued..."
- Correct: "Smith has argued that... [acknowledgment]. But the evidence Smith relies on is thin [response]."

Direct critical characterization at the work, not the person.

---

### Level 5 — First-person concession

Signals: you concede the validity of the objection in your own voice; highest weight given to the alternative view.

**First-person:**
- I [understand / know / realize] that..., but...
- We [would / could / can / might / may] [say / argue / claim / think] that..., but these effects are outweighed by...

**Passive with concessive phrase:**
- It is [true / possible / likely] that...
- It [must / should / can] be [admitted / acknowledged / noted / conceded] that..., nevertheless...

**Fronted concessive:**
- [Granted / Certainly / Admittedly / True / To be sure / Of course], X has argued that..., however...

**Examples:**

- "Admittedly, our sample does not cover the Southern states, where conditions differ substantially; nevertheless, the pattern we identify is robust within the studied region."
- "It must be acknowledged that no good evidence proves that needle exchanges reduce drug initiation rates. Nevertheless, the evidence that they reduce disease transmission is..."
- "Granted, Adams has claimed that the program increases dependency; however, our longitudinal data shows the opposite effect after 18 months."

---

## Response Vocabulary (4 Registers)

Begin the response with a pivot term: **but**, **however**, **on the other hand**, **yet**, **still**.

### Register 1 — Tactful (softest)

Use when: the objection is from a respected source or you genuinely find it difficult to see the basis for the claim, without wanting to be blunt about it.

- But [I do not quite understand how / I find it difficult to see how / It is not clear to me how] X can claim that, when...

---

### Register 2 — Noting unsettled issues

Use when: the objection raises genuine complexity that your response addresses but does not fully resolve.

- But there are other issues here... / But there remains the problem of...
- But we must look at all the available evidence.
- That explains some of the problem, but it is too complex for a single explanation.
- That principle holds in many cases, but not in all.

---

### Register 3 — Claiming irrelevance or unreliability (moderate force)

Use when: the objection, though interesting, does not bear on the specific claim at issue, or the evidence/reasoning behind it is weak.

**Irrelevance:**
- But as insightful as that may be, it [ignores / is irrelevant to / does not bear on] the issue at hand.
- But the argument [overlooks / ignores / misses] key factors.

**Unreliable reasoning or evidence:**
- But the [evidence / reasoning] is [unreliable / shaky / thin].
- But the argument is [untenable / weak / confused / simplistic].

---

### Register 4 — Direct challenge (bluntest)

Use when: the objection is clearly wrong, based on bad evidence, or logically defective — and you can show this.

- But the evidence does not support this. [cite counter-evidence]
- But this analysis ignores [the full dataset / the relevant controls / the conflicting findings].
- But we must look at all the available evidence, not just the studies that support this view.

**Important:** Even in the bluntest register, direct criticism at the work, not the person. "Smith's evidence is important, but we must look at all the available evidence" is more credible than "Smith carelessly ignored conflicting data."

---

## Full Sentence Templates

These templates combine an acknowledgment phrase and a response pivot. The bracketed text is a fill-in slot.

**Low-weight rebut:**
Although [some have argued that X], [our evidence shows Y] because [reason].

**Medium-weight qualify:**
Although [some researchers / critics] have argued that [X], our findings suggest this holds only [in condition Z]; within [our scope], we find [Y].

**High-weight concede-and-reframe:**
It must be acknowledged that [flaw]. Nevertheless, [the rest of the argument more than balances this because...] / [more research would show...] / [our contribution is to establish... and to suggest what a better answer would require].

**Definitional dispute:**
Some researchers maintain that [term X means Y]. We use [term X] in the sense of [your definition], which [reason for adoption]; we acknowledge that [alternative definition] is used in [contexts], but [reason it does not fit this research problem].

**Competing causes:**
Although [competing cause C] also contributes to [effect E], [our argument focuses on / our analysis controls for C, isolating the effect of] [your cause] because [reason].

**Counterexample:**
Although [counterexample CE] might seem to undermine this claim, [it falls within the normal range of variation / it is an outlier in a dataset of N / it does not apply to the period / population / conditions we study] because [reason].

FILE:references/three-predictable-disagreements.md
# Three Predictable Disagreements

No matter what field you work in, three kinds of alternatives appear in almost every research argument. Knowing these types in advance lets you address them deliberately rather than being caught off guard.

Source: Booth, Colomb, Williams, Bizup & FitzGerald, *The Craft of Research*, 4th ed., Chapter 10 (Quick Tip: Three Predictable Disagreements).

---

## Type 1 — Competing Causes

**When it applies:** Any argument that makes a causal claim — "X causes Y," "X leads to Y," "X is responsible for Y."

**The core problem:** No effect has a single cause. If you argue that X causes Y, every reader will think of other causes. Readers do not expect you to prove your cause is the *only* cause, but they do expect you to acknowledge that other causes exist, explain why you are focusing on yours, and address any competing cause that is significantly more prominent than your own in the field.

**What readers will say:**
- "There are other causes you haven't mentioned."
- "The cause you name is less important than the one I'm thinking of."
- "You've focused on one factor while ignoring others."

**What you must do:**
1. Name the most prominent competing causes your readers will know about.
2. Either explain why your cause is the focus of this argument (scope), or show why it is the most important (relative weight), or show that your analysis controls for the competing causes (methodological isolation).
3. If you feel readers might think another cause deserves more attention than you give it, acknowledge that view and explain your choice to deemphasize it.

**Example — Honeybee colony collapse:**
If you argue that pesticide use is causing colony collapse, acknowledge that readers can also list habitat loss, disease, genetically modified crops, and parasites. Do not ignore them; state your scope or provide controls.

**Example — Student performance decline:**
If you argue that phone use in class causes lower test scores, readers will raise: sleep deprivation, economic stress, pre-existing skill gaps, teaching quality. Acknowledge the most prominent; explain your methodological controls or your argument's scope.

---

## Type 2 — Counterexamples

**When it applies:** Any argument that makes a generalization — claims covering a class of things, a trend, a pattern, a typical case.

**The core problem:** No matter how rich your evidence, readers will think of exceptions. Some exceptions are real and damaging; some are aberrant cases that fall within normal variation. You must address the former and explain the latter. If you do not, readers who know the counterexample will dismiss your argument.

**What readers will say:**
- "But what about [vivid exception]? That undermines your claim."
- "You're ignoring cases where this doesn't hold."
- "One cold summer disproves your warming claim." (statistical reasoning failure)

**What you must do:**
1. Think of the counterexamples yourself before readers raise them — especially vivid, memorable, or well-known ones.
2. Acknowledge the more plausible ones, especially if they are vivid.
3. Explain why you do not consider them as damaging as readers might — either because they fall within normal variation, because they apply to conditions outside your scope, or because your claim is a trend claim and not a universal.

**Special warning — Statistical variation:** Be particularly careful with claims about phenomena that have wide natural variation (climate, test scores, economic indicators). Readers who do not understand statistical reasoning will focus on an aberrant case as if it refutes a trend. Acknowledge this explicitly: "A cold Fourth of July in Florida does not disprove a claim about global warming, any more than a warm New England Christmas proves it." The counterexample only damages the claim if it represents a systematic pattern, not an individual aberration.

**Example — Remote work productivity:**
Claiming remote work increases output in knowledge firms. Predictable counterexample: "My team's productivity collapsed after going remote." Acknowledge that individual team variation exists; explain that your evidence is about aggregate patterns across controlled comparisons, not individual cases.

---

## Type 3 — Definitional Scope Disputes

**When it applies:** Any argument whose claim depends on a term that has more than one plausible meaning in common or scholarly use.

**The core problem:** Readers must accept your definition to accept your claim. Readers often redefine terms to suit their own views — sometimes in good faith, sometimes strategically. If you do not anticipate this, readers can dismiss your entire argument by rejecting your key term rather than engaging with your evidence.

**What readers will say:**
- "I don't define X the way you do. By my definition, your claim is false."
- "That's a technical term and you're using it colloquially."
- "You're using a technical definition that most readers don't accept."

**What you must do:**
1. When your argument hinges on a contested term, define it with a supporting sub-argument — explain why your definition is the right one for this research context.
2. Never treat a dictionary definition as authoritative in a specialized argument. Dictionaries record usage; they do not resolve contested meaning. ("According to Merriam-Webster..." is not a legitimate definitional move in academic argumentation.)
3. Acknowledge plausible alternative definitions.
4. If you use a technical term that also has a common meaning (like *social class*, *theory*, *culture*), acknowledge the common meaning and explain why you have adopted the technical one.
5. Conversely, if expert readers expect you to use a technical term and you do not, acknowledge that and explain your choice.

**Example — Addiction:**
Arguing that social media is "addictive." Physiological dependence definition vs. behavioral compulsion definition. Each leads to a different verdict on social media. Acknowledge the physiological definition, then argue for the behavioral one and show why your evidence meets those criteria.

**Example — Theory:**
In everyday speech, "theory" means a guess. In science, "theory" means a well-tested explanatory framework. If you use "theory" in the scientific sense for a general audience, acknowledge the common meaning and explain your usage.

**Example — Social class:**
Different disciplines (economics, sociology, cultural studies) use "social class" with different criteria. Acknowledge the plurality; define your usage precisely; explain why it fits your research context.

---

## Using the Three Types Together

Most arguments face more than one type of predictable disagreement. A climate change argument, for example, may face all three simultaneously:

- **Competing causes:** Other factors (volcanic activity, solar cycles) also influence temperature.
- **Counterexamples:** Record cold events in specific regions.
- **Definitional scope:** What counts as "climate" vs. "weather"; how "significant warming" is measured.

Check your argument against all three types before deciding which objections to prioritize in the Goldilocks filter (Step 3 of `counterargument-handler`).

---
name: argument-organization-reviser
description: "Revise the structural organization of a research paper draft by applying a four-level top-down procedure — Frame (intro/conclusion alignment), Argument (section reasons + evidence ratios), Paper Organization (key-term threading + section signals), and Paragraphs (topic sentence placement + length). Use this skill whenever the user has a complete draft and asks to revise, reorganize, or strengthen its structure — not its prose style. Triggers include: user shares a draft paper and asks for structural feedback; user says sections feel disconnected or the argument is hard to follow; user's introduction and conclusion seem to contradict or not reinforce each other; user suspects their sections lack clear points or bury them in the middle; user cannot tell whether their evidence-to-reasoning ratio is balanced; user's paragraphs open with evidence rather than claims; user is preparing to submit and wants a final organizational pass. Also triggers on: \"revise my structure,\" \"does my argument hold together,\" \"my advisor said the organization is unclear,\" \"do my sections flow,\" \"I need to check the coherence.\" This skill applies structural revision only — it does NOT revise prose style or sentence clarity (use prose-clarity-reviser for that), and does NOT rebuild an argument from scratch (use research-argument-builder for that)."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-craft-of-research/skills/argument-organization-reviser
metadata: {"openclaw":{"emoji":"📚","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-craft-of-research
    title: "The Craft of Research, 4th Edition"
    authors: ["Wayne C. Booth", "Gregory G. Colomb", "Joseph M. Williams", "Joseph Bizup", "William T. FitzGerald"]
    chapters: [13]
tags: [research-methodology, academic-writing, revision, organization, argumentation]
depends-on: [research-paper-planner]
execution:
  tier: 1
  mode: full
  inputs:
    - type: document
      description: "A complete or near-complete draft of a research paper, report, or analytical essay — any length, any field"
  tools-required: [Read]
  tools-optional: [Write]
  mcps-required: []
  environment: "Any agent environment. User supplies draft as pasted text or a file path."
discovery:
  goal: "Apply the four-level top-down revision procedure to a draft, producing annotated diagnoses and specific repair prescriptions at each level before moving to the next"
  tasks:
    - "Level 1: Check frame — verify introduction and conclusion boundaries are visible, main point appears at/near end of introduction, conclusion point is specific and contestable"
    - "Level 2: Check argument — verify each major section has its own point, evidence ratio >= 1/3 per section, acknowledgments present, warrants surfaced"
    - "Level 3: Check paper organization — verify key terms thread through sections, section joints are signaled, each section is clearly tied to the whole"
    - "Level 4: Check paragraphs — verify topic sentences open or close paragraphs (not buried), paragraph lengths are appropriate, breaks signal transitions"
    - "Cool-and-paraphrase test — skim structure only and paraphrase; check whether the paraphrase matches the intended argument"
    - "Abstract diagnostic — if field requires one, assess whether the abstract matches the three-component structure"
  audience: "Researchers, students, analysts, and professionals with a complete draft who need structural revision before submission"
  triggers:
    - "User shares a draft paper and asks for structural feedback or revision"
    - "Introduction and conclusion feel disconnected or contradictory"
    - "Sections feel like independent essays rather than parts of one argument"
    - "User cannot locate the main point in the introduction"
    - "Evidence vastly outnumbers reasoning or vice versa"
    - "Paragraphs open with evidence or quotations rather than claims"
    - "User preparing to submit and wants a final structural pass"
---

# Argument Organization Reviser

## When to Use

You have a complete draft — or near-complete draft — of a research paper, report, or analytical essay. The argument is already built. What you need now is structural revision: making the organization of your draft as clear to readers as the argument is to you.

Structural revision is different from prose revision. This skill addresses:
- Whether your introduction and conclusion frame the same argument
- Whether each section carries its own distinct reason
- Whether the evidence-to-reasoning ratio is balanced
- Whether key concepts thread visibly through the paper
- Whether paragraphs are properly anchored

It does **not** address sentence clarity, nominalization, or information flow. For those concerns, use `prose-clarity-reviser` after completing this skill.

**A note on order:** Revise top-down, from large structure to small. Do not fine-tune paragraph sentences before confirming that the sections those paragraphs belong to are correctly placed. Time spent polishing a section that later gets cut or moved is wasted.

**Preconditions:**
- A draft exists. If the user has only a plan or outline, redirect to `research-paper-planner`.
- If the argument itself is missing (no main claim, no reasons), redirect to `research-argument-builder`.

---

## Input Gathering

### Required

- **The draft:** Full text, pasted or provided as a file path. If only a section is provided, note that the diagnosis will be partial.
  -> Ask: "Please share the draft you want to revise structurally — paste it here or give me the file path."

### Useful

- **Field and genre:** Knowing whether this is a lab report, a humanities seminar paper, a policy brief, or a thesis chapter affects what conventions govern heading use, abstract format, and where the main point should appear.
- **Target audience:** Expert readers vs. generalists affects how much background and how many key-term definitions are needed.
- **Submission deadline and purpose:** If submission is imminent, prioritize Levels 1 and 2 only.

---

## Execution

Work strictly top-down. Complete each level's diagnosis and repair before descending to the next. This order matters because: a repositioned section changes every paragraph in it; a repositioned main point changes the meaning of both introduction and conclusion. Fixing sentences before fixing structure is like painting walls before fixing the foundation.

---

### Level 1 — Revise the Frame

**What this level checks:** Do readers know instantly where your introduction ends, where your conclusion begins, and where your main point is stated?

Readers need to recognize three things instantly and unambiguously:
1. Where the introduction ends
2. Where the conclusion begins
3. Which sentence in one or both states the main point

**Why this matters:** Readers do not read word by word, adding up detail. They begin with a sense of the whole — where the paper is going and why. If the frame is not visible, readers spend cognitive effort searching for orientation instead of engaging with the argument.

**Diagnosis — check these three things:**

**1a. Are the introduction and conclusion boundaries visible?**

Look for one of these markers:
- An extra blank line (white space) separating the introduction from the body
- A heading marking where the body or first major section begins
- A heading marking the conclusion

If neither exists, readers may not know when your introduction ends.

-> If boundaries are missing: add white space or headings at those joints. (Check whether your field permits or expects headings — if not, use extra space.)

**1b. Is your main point stated at or near the end of your introduction?**

Find the last two or three sentences of your introduction. Does one of them state a specific, contestable claim — what the whole paper argues?

Watch for the weak form: an introduction that ends with a topic announcement ("In this paper I will discuss...") rather than a claim ("This paper argues that..."). A topic announcement tells readers what you will write about; a claim tells them what you are arguing. Readers need the claim.

> Example of a topic announcement (weak): "In this paper I will discuss the reasons for the Crusades."
> Example of a claim (strong): "The Crusades were driven not only by religious zeal but by shrewd political moves to unify the Roman and Greek churches in the face of internal forces threatening to tear them apart."

-> If the main point is absent from the introduction: draft a point sentence (one or two sentences, specific and contestable) and place it at the end of the introduction. It should name the key concepts the body will develop.

-> If you are writing in a field where the convention is to reserve the main point for the conclusion (some humanities traditions): your introduction still needs a "launching point" — a sentence that introduces the key concepts the paper will explore, even if it does not state the full conclusion.

**1c. Do your introduction point and conclusion point align?**

Read the point sentence in your introduction and the point sentence in your conclusion side by side.

They should:
- Not contradict each other
- Not be identical — the conclusion's version should be more specific, more developed, and more directly address the "so what?" question

> Good alignment example: Introduction claims "political concerns drove the Crusades as much as religious ones." Conclusion claims "the political calculations of Urban II and Gregory VII — specifically their need to prevent Byzantine schism — shaped the Crusades more decisively than theological unity arguments did."

-> If they contradict: one of the two is wrong; decide which version of the argument you actually made in the body, then revise the other to match.
-> If they are identical word for word: develop the conclusion's version; it should add something the introduction could not say before the argument was made.

---

### Level 2 — Revise the Argument

**What this level checks:** Does the structure of your paper match the structure of your argument? Is the evidence-to-reasoning ratio sound?

Do this only after Level 1 is resolved. A frame you cannot trust will mislead your assessment of the argument.

**Why this matters:** Your argument consists of a main claim supported by reasons, each backed by evidence. If your sections do not map onto that structure — if sections mix multiple reasons, or contain only evidence with no stated reason, or contain stated reasons with no evidence — readers cannot follow the argument even if the frame is clear.

**2a. Does each major section carry exactly one reason supporting the main claim?**

Go through each major section. For each one, ask: *What is the one reason this section asserts?* Write it in a sentence at the top of a blank page.

- If you cannot state a reason for a section, the section may be context, background, or evidence — not a structural section of the argument. Determine whether it belongs as a subsection inside another section or as background before the argument begins.
- If a section contains two distinct reasons, split it into two sections.
- If two sections contain the same reason expressed differently, merge them.

> Example: A section titled "Evidence from Historical Documents" contains both the reason "documentary evidence shows political motivation" and separate evidence for a second reason about ecclesiastical authority. These need to be separated.

**2b. Does each section strike a sound balance between reasons and evidence?**

In each section, identify every item that counts as evidence: summaries, paraphrases, quotations, facts, figures, graphs, tables, reported cases — anything you took from a source.

Apply this test: **If evidence and its explanation account for less than roughly one-third of a section, you probably do not have enough evidence for that reason.** If you have extensive evidence but few or no stated reasons, you may have a data dump — information without analytical claim.

-> Too little evidence: find additional data, or narrow the section's claim to match what your evidence actually supports.
-> Data dump (evidence without reasoning): write a sentence stating what all this evidence adds up to — the reason it supports. Place that sentence at the top of the section.

**2c. Is your evidence reliable? Have you acknowledged objections and expressed warrants?**

- Check data and quotations against your notes. Errors of fact undermine even a structurally sound argument.
- Make sure every quotation or data item is explicitly connected to the reason it supports. Do not assume the connection is obvious.
- For each section, identify the most likely objection a skeptical but fair reader would raise. Have you acknowledged it? Have you responded to it?
  -> Ask yourself: *Why do I believe this? Am I really making as strong a point as I think? What about...?* Answer the objections that matter.
- For each section, identify the unstated warrant — the principle that connects the evidence to the reason. Ask: would a reader in this field accept this warrant without argument? If not, state it and support it before the reason.

---

### Level 3 — Revise the Organization of the Paper

**What this level checks:** Does the paper cohere? Do readers see it as one argument rather than separate sections that happen to share a cover page?

Do this only after Levels 1 and 2 are resolved.

**Why this matters:** Coherence is not a feeling — it has a structural cause. Readers experience a paper as coherent when key terms from the main claim recur visibly throughout the paper, when section beginnings signal how each section relates to what came before, and when each section clearly contributes to the whole.

**3a. Do key terms thread through the whole paper?**

Locate your main point sentence in the introduction. Circle the words that name the central concepts your argument depends on — the terms that define your specific question, not the general vocabulary of your field.

> Example from a paper on the Crusades: circle *political concerns*, *unity*, *internal forces*, *dividing*. These are not general medieval history terms — they are the specific concepts your argument turns on.

Now go through the body:
- Circle those same terms wherever they appear in the body.
- Underline related words — synonyms, associated concepts, terms in the same semantic family.

**Test:** Does at least one of these key terms appear in most paragraphs? If a passage goes several paragraphs without any key term, readers may conclude the paper has wandered.

-> If a passage lacks key terms entirely: try adding a few. If that feels forced, the passage may be off-topic — consider cutting or rewriting it so it connects to the argument.

**3b. Is the beginning of each major section clearly marked?**

Read through the paper and ask: could you quickly and confidently insert section headings to mark where each major section begins?

- If yes: your organization is probably visible to readers. (Insert the headings, or confirm existing ones are accurate.)
- If you struggle to identify where one section ends and another begins: readers will struggle too. Consider adding white space, transition sentences, or headings.

-> If your field prohibits headings: add an extra blank line at the major joints and open each new section with a transition sentence that signals the shift.

**3c. Does each section signal how it relates to the preceding section?**

The first words of each major section should tell readers why it comes where it does. Readers need to understand not just where sections begin but why they appear in this order.

Useful signal phrases:
- "More important..." (escalation)
- "The other side of this issue is..." (counterargument or contrast)
- "Some have objected that..." (acknowledgment)
- "One complication is..." (qualification)
- "First... Second..." (explicit enumeration)
- "As a result..." (consequence)

-> Review the opening sentence of each section. Does it signal the logical relationship to the previous section? If not, add a brief transition phrase or sentence.

**3d. Is the point of each section clearly stated?**

For each section, locate the sentence that states the section's point — what it argues, not merely what it discusses.

- Prefer stating the section point at the end of the section's introductory paragraph, not buried mid-section. Readers need to know early what a section is arguing.
- For sections longer than four to five pages, consider also restating the point and summarizing the argument at the section's close.

**3e. Do the key terms of each section uniquely distinguish it from the others?**

For each section, repeat the key-term exercise: circle the terms in that section's point sentence (not the whole-paper terms, but the terms unique to this section). Check whether those terms run through the section.

- If you find no unique key terms, the section may not be contributing a new idea — it may be repeating another section. Consider merging them.
- If the same unique key terms run through two sections, those sections may overlap. Consider merging or sharpening the distinction.

---

### Level 4 — Check the Paragraphs

**What this level checks:** Are paragraphs properly anchored by topic sentences? Are they the right length?

Do this only after Levels 1–3 are resolved. Moving sections in Level 3 changes which paragraphs exist and where they sit.

**Why this matters:** Paragraphs are the unit at which readers test whether they understand what they are reading. Each paragraph must signal its key concept at or near its opening — otherwise readers cannot connect what they are reading to the section's point.

**4a. Does each paragraph signal its key concept near the opening?**

Open each paragraph with one or two sentences that name its key concept and orient readers to what follows. This is not the same as stating a formal "topic sentence" — the goal is to give readers an anchor before they encounter the detail.

- If the paragraph's point is not stated in the opening, it should appear in the paragraph's final sentence. Never bury the point in the middle.
- If neither the first nor the last sentence states the point, revise: move the point sentence to the opening or the close.

> Example: A paragraph opening with "In 1095, Urban II gave a speech at the Council of Clermont..." buries the analytical point. Better: "Urban II's Council of Clermont speech reveals more political calculation than religious urgency. In 1095..."

**4b. Are paragraphs the right length?**

There is no universal rule, but:
- Short paragraphs of two or three sentences suggest underdeveloped points. If a series of paragraphs are this short, either merge the ones making related points or develop each point further.
- Long paragraphs exceeding one page may mean the point is not sharply focused, or that the paragraph is doing the work of two. Consider splitting: give each idea its own paragraph.
- Use short paragraphs deliberately to emphasize a transition or highlight a key point — the visual pause draws attention.

**4c. Do paragraph breaks signal transitions effectively?**

Paragraph breaks function like pauses in a conversation: they give readers a moment to process a point before moving on. Use a break:
- After making a strong or complex point (let it land)
- When transitioning to a new but related idea
- When you shift from evidence to analysis or vice versa

---

### Level 5 — Cool-and-Paraphrase Test

**What this test does:** Checks whether the structure you have revised is legible at the skimming level — the level at which readers first encounter a paper and decide whether to engage with it.

**Why this matters:** You know your argument too well to test it as your readers will encounter it. This test forces you to read like a stranger.

**How to run it:**

1. Set the draft aside. Do not read it for at least a day if possible. (If time is short, skip directly to step 2.)
2. Return to the draft and skim only its structure: the introduction, the first paragraph of each major section, and the conclusion. Do not re-read the full body.
3. Based only on what you skimmed, paraphrase the paper to yourself (or to someone else): What is the problem this paper addresses? What does it argue? What are the main reasons? What is the significance?
4. Does the paraphrase match the argument you intended? If yes, the structure is working. If not — if parts are missing, confused, or inverted — the gaps are where the structure fails.

-> If you can find someone else to do the skimming and paraphrasing: their paraphrase will predict with reasonable accuracy how well your final readers will understand the argument. Consider their version seriously, even if you disagree with specific suggestions.

---

## Abstract Diagnostic (if applicable)

If your field requires an abstract, use the abstract as an additional diagnostic after the cool-and-paraphrase test.

**Why:** An abstract must compress the whole paper's argument into a paragraph. If you cannot write a coherent abstract, the paper's structure is probably not coherent.

An abstract should do three things (the same three things an introduction does, compressed):
1. State the research problem or context
2. Announce the key themes or approach
3. State the main point or a launching point that anticipates it

Three common patterns:

| Pattern | Structure | When to use |
|---|---|---|
| **Context + Problem + Main Point** | 1–2 sentences of context, 1–2 sentences of problem, 1 sentence of result | When you have a specific result to state upfront |
| **Context + Problem + Launching Point** | Same, but states the nature of the investigation rather than the result | When the result is complex or when disciplinary convention defers the conclusion |
| **Summary** | Context + problem + compressed argument (methods, evidence, and procedures) + main point | Required in many scientific and social-scientific journals |

-> If you cannot state your main point in one sentence for the abstract, return to Level 1 and revise the introduction's point sentence first.

-> For discoverability: put your most important search keywords in both the title and the first sentence of the abstract. A future researcher searching for your work will find it through those terms.

---

## Distinguishing This Skill from Adjacent Skills

| Concern | Use this skill | Use another skill |
|---|---|---|
| Sentences are dense or unclear | No | `prose-clarity-reviser` |
| Main claim and reasons need to be built | No | `research-argument-builder` |
| Draft does not exist yet; planning a structure | No | `research-paper-planner` |
| Introduction and conclusion need full writing | No | See references/introduction-conclusion-revision.md |
| Objections and counterarguments need developing | Partially (Level 2) | `counterargument-handler` |

---

## Examples

### Example A: Introduction/Conclusion Misalignment (Level 1)

**Introduction point (end of intro):** "This paper will explore the political and religious motivations behind the First Crusade."

**Conclusion point:** "The efforts of Urban II and Gregory VII were shrewd political moves to unify the Roman and Greek churches and prevent the breakup of the empire from internal forces threatening to tear it apart."

**Diagnosis:** The introduction states a topic ("will explore motivations") while the conclusion states a specific, contestable claim. These do not align — the introduction promises an exploration, the conclusion delivers an argument.

**Repair:** Replace the introduction's final sentence with a claim that introduces the paper's key concepts: "In a series of documents, the popes proposed their Crusades to restore Jerusalem to Christendom, but their words suggest other issues involving political concerns about European and Christian unity in the face of internal forces that were dividing them."

---

### Example B: Data Dump (Level 2)

A section titled "Economic Conditions" contains three pages of statistics, tables, and historical data about medieval European trade, population, and agricultural output, with no sentence stating what these data collectively prove.

**Diagnosis:** Evidence < 1/3 of section is not the problem here — evidence > 2/3 is. There is no stated reason. This is a data dump.

**Repair:** Write the section's reason statement and place it at the top: "The economic instability of late eleventh-century Europe gave feudal lords strong incentives to redirect their restless knight class outward — the Crusades offered an outlet that a stable economy would not have required."

---

### Example C: Paragraph Point Buried in Middle (Level 4)

**Before:**
> "In 1095, Pope Urban II gave a speech at the Council of Clermont in which he appealed to knights to join a Crusade. He described the suffering of Eastern Christians and the loss of Jerusalem. His words, however, reveal something more calculated: Urban framed the Crusade as a solution to the growing threat of Byzantine schism and the erosion of papal authority in the West. By 1096 the First Crusade had been launched."

**After (point moved to opening):**
> "Urban II's Council of Clermont speech reveals more political calculation than religious urgency. In 1095, Urban appealed to knights to join a Crusade, describing the suffering of Eastern Christians and the loss of Jerusalem. But his words frame the expedition as a solution to a more urgent problem: the growing threat of Byzantine schism and the erosion of papal authority in the West. By 1096 the First Crusade had launched."

---

## References

- `references/introduction-conclusion-revision.md` — Full procedure for revising the introduction and conclusion as polished final text (not just structural diagnosis)
- `references/key-term-threading.md` — Step-by-step guide to identifying key concepts and tracing them through a multi-section paper
- `references/abstract-patterns.md` — Annotated examples of all three abstract patterns across disciplines

## License

This skill is licensed under [CC-BY-SA-4.0](https://creativecommons.org/licenses/by-sa/4.0/).
Source: [BookForge](https://github.com/bookforge-ai/bookforge-skills) — The Craft of Research, 4th Edition by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams, Joseph Bizup, William T. FitzGerald.

## Related BookForge Skills

Install related skills from ClawhHub:
- `clawhub install bookforge-research-paper-planner`

Or install the full book set from GitHub: [bookforge-skills](https://github.com/bookforge-ai/bookforge-skills)

FILE:references/abstract-patterns.md
# Abstract Patterns

Reference for `argument-organization-reviser` — annotated examples of the three abstract patterns across disciplines, plus guidelines for writing and evaluating abstracts.

Source: *The Craft of Research, 4th Edition*, Booth, Colomb, Williams, Bizup, FitzGerald — Chapter 13.6 (Quick Tip: Abstracts).

---

## What an Abstract Must Do

An abstract is a paragraph that tells readers what they will find in a paper, article, or report. It is shorter than an introduction but does the same three things:

1. **State the research problem** — why this paper exists
2. **Announce key themes** — the concepts the argument depends on
3. **State the main point or a launching point** that anticipates the main point

An abstract that does only (1) is incomplete. An abstract that does only (3) is disorienting without (1) and (2).

---

## Pattern 1: Context + Problem + Main Point

This is an abbreviated introduction. It begins with one or two sentences to establish the context of previous research, continues with one or two sentences to state the problem, and concludes with the main result of the research.

**Use when:** You have a specific, reportable result that can be stated directly.

**Example (political science):**

> Scholars have long assumed that democracy improves the quality of life for its citizens. [*context*] But recent research has called this orthodoxy into question, suggesting that there is little or no relationship between a country's regime type and its level of human development. [*problem*] In this article, we argue that democracy can be shown to advance human development, but only when considered as a historical phenomenon. [*main point*]

**Analysis:**
- Context sentence establishes what the field has assumed (one clear claim, not a literature survey)
- Problem sentence states the specific challenge to that assumption
- Main point sentence is specific (not "we discuss democracy") and contestable (the restriction "only when considered as a historical phenomenon" is arguable)
- All three key concepts appear: *democracy*, *human development*, *historical phenomenon*

---

## Pattern 2: Context + Problem + Launching Point

This pattern is the same as Pattern 1, except that instead of stating the specific result, it states the nature of the investigation — what the paper does, rather than what it finds.

**Use when:** The result is complex or qualified enough that a one-sentence statement would mislead; or when disciplinary convention defers the main conclusion to the paper's end; or when the paper's contribution is methodological rather than conclusive.

**Example (same paper, launching point version):**

> Scholars have long assumed that democracy improves the quality of life for its citizens, [*context*] but recent research has called this orthodoxy into question, suggesting that there is little or no relationship between a country's regime type and its level of human development. [*problem*] In this article, we review this body of work, develop a series of causal pathways through which democracy might improve social welfare, and test two hypotheses: (a) that a country's level of democracy in a given year affects its level of human development and (b) that its stock of democracy over the past century affects its level of human development. [*launching point*]

**Analysis:**
- The launching point states what the paper does (reviews, develops, tests) rather than what it concludes
- This gives readers a map of the inquiry without foreclosing the result
- The two hypotheses are specific enough to anchor expectations

---

## Pattern 3: Summary

A summary abstract states the context and problem, then compresses the argument — including the methods or evidence — before stating the main point.

**Use when:** Required by journals or disciplines that want readers to assess the full argument from the abstract alone (common in quantitative social sciences, medicine, and engineering).

**Example (same paper, as published):**

> Does democracy improve the quality of life for its citizens? Scholars have long assumed that it does, [*context*] but recent research has called this orthodoxy into question. [*problem*] This article reviews this body of work, develops a series of causal pathways through which democracy might improve social welfare, and tests two hypotheses: (a) that a country's level of democracy in a given year affects its level of human development and (b) that its stock of democracy over the past century affects its level of human development. Using infant mortality rates as a core measure of human development, we conduct a series of time-series cross-national statistical tests of these two hypotheses. We find only slight evidence for the first proposition, but substantial support for the second. [*summary*] Thus, we argue that the best way to think about the relationship between democracy and development is as a time-dependent, historical phenomenon. [*main point*]

**Analysis:**
- Opens with a rhetorical question that will be upended — a stylistic option that the compressed form of the abstract can accommodate
- Summary compresses both the approach (statistical time-series tests) and the finding (slight evidence for first, substantial for second)
- Main point appears last — as it does in the paper itself
- Note: the opening rhetorical question allowed the authors to abbreviate the context statement; the problem is stated more concisely as a result

---

## Choosing the Right Pattern

| If... | Use Pattern |
|---|---|
| You have a specific result you can state in one sentence | 1 (Context + Problem + Main Point) |
| Your finding is complex or your field defers the conclusion | 2 (Context + Problem + Launching Point) |
| Your journal requires readers to evaluate the full argument from the abstract | 3 (Summary) |
| You are unsure | Check three recent abstracts in your target journal and match their pattern |

---

## Abstract as Revision Diagnostic

**The abstract is not only a product — it is a test.**

If you cannot write a coherent abstract, the paper's structure is probably not coherent. Specifically:

- If you cannot write the problem sentence: your research question may not be formulated as a gap in existing knowledge. Return to `research-problem-framer`.
- If you cannot write the main point sentence: your introduction probably lacks a specific, contestable point sentence. Return to Level 1 of `argument-organization-reviser`.
- If the abstract's main point contradicts your conclusion: your introduction and conclusion are misaligned. Return to Level 1.
- If the abstract mentions concepts that do not appear in the introduction: the key terms in your introduction may not match the key terms your argument actually uses. Return to Level 3.

---

## Discoverability: Keywords in the Abstract

If your paper will be published or indexed:

- Identify the two or three most important search terms a future researcher would use to find your paper.
- Put those terms in the paper's title.
- Put those same terms in the first sentence of the abstract.

Search engines and database indexers weight title and abstract opening heavily. An abstract that buries its key terms in the third sentence is harder to find than one that opens with them.

> Example: A paper on youth concussion rates should open its abstract with "youth athletes" and "concussion" in the first sentence — not in the third sentence after context has been established. The context can accommodate the key terms: "Youth athlete concussion rates have become a major focus of sports medicine research over the past decade."

FILE:references/introduction-conclusion-revision.md
# Introduction and Conclusion Revision

Reference for `argument-organization-reviser` — covers revision of the introduction and conclusion as final polished text, beyond the structural diagnosis in Level 1.

Source: *The Craft of Research, 4th Edition*, Booth, Colomb, Williams, Bizup, FitzGerald — Chapter 13 (structural revision), Chapter 16 (introductions and conclusions).

---

## The Three Things Readers Need from the Frame

Readers must recognize three things instantly and unambiguously:

1. **Where the introduction ends** — a visual signal (white space, heading) that separates orientation from argument
2. **Where the conclusion begins** — same principle: a visible boundary
3. **Which sentence states the main point** — ideally in both the introduction and conclusion, positioned where readers expect it (end of introduction; beginning or near-beginning of conclusion)

If any of these three markers are missing, readers will spend cognitive effort searching for orientation instead of following the argument.

---

## Revising the Introduction

A polished introduction does four things in sequence:

### 1. Establish Context

Briefly summarize the prior work or conventional understanding that your paper engages. Order sources by significance to your argument, not by chronology or the order you read them. Omit sources that are only loosely related — excess context diffuses focus.

> The context establishes the conversation your paper joins. Readers need to know what is already on the table before they can understand what you are adding or challenging.

### 2. State the Problem (Gap)

Transform your research question into a statement of what the existing literature fails to explain, leaves contested, or gets wrong.

Pattern: "Few of these researchers, however, have explained [what your question addresses]."

The gap is the reason your paper exists. If there is no gap — if everything is already well understood — there is no reason for readers to read further.

### 3. State the Consequence (So What?)

Answer: "Until we understand [your gap], we cannot know [a more important thing]." The consequence turns a private curiosity into a shared concern. Without it, readers may feel the paper addresses only a specialist puzzle.

### 4. State the Main Point

At or near the end of the introduction, state the specific, contestable claim your paper will defend. This sentence should name the key concepts that the body of the paper will develop.

**What makes a main point sentence strong:**
- Specific — not "this paper will discuss" but "this paper argues that"
- Contestable — a reasonable reader could initially disagree
- Anchoring — names the terms (concepts, variables, categories) that the argument depends on

**Exception:** If your field or genre reserves the conclusion for the main point, your introduction should still end with a "launching sentence" — a sentence that introduces the key concepts the paper will explore, even without stating the full result.

---

## Revising the Conclusion

The conclusion is not a summary of what you said. It is the place where the argument, now made, allows you to say something you could not say at the start.

### The Conclusion Does Three Things

1. **Restate the main point** — not word for word, but developed in light of what the argument has established. The conclusion's version of the main point should be more specific and more nuanced than the introduction's version.

2. **Articulate significance** — return to the "so what?" question. Now that the argument has been made:
   - What does a reader understand about a larger issue that they could not before?
   - What should they do, think, or investigate differently?
   - What new questions does the argument open?

3. **Avoid ending on a list** — a conclusion that merely recaps the argument point by point closes on what readers already know. Close on what they now understand that they did not before.

### Testing Introduction/Conclusion Alignment

Place both point sentences side by side. They should:
- Agree on the claim (no contradiction)
- Not be word-for-word identical
- Show progression: the conclusion's version is more developed, more nuanced, or more broadly significant

If they contradict: decide which version the body actually supports, and revise the other to match.

If they are identical: the conclusion is a missed opportunity. Develop it.

---

## Common Introduction Failures and Repairs

| Failure | Symptom | Repair |
|---|---|---|
| Topic announcement | "In this paper I will discuss X" | Replace with a claim: "This paper argues that X because Y" |
| Gap without consequence | States what is unknown but not why it matters | Add the "so what?" chain: "Until we understand X, we cannot know Y" |
| Main point too vague | "Policy needs to improve" | Make it contestable and specific: "Current monitoring policy fails female youth athletes for three structural reasons that equipment regulation alone cannot address" |
| Key concepts absent from point sentence | Introduction names the topic but not the specific terms the argument turns on | Revise point sentence to name the key concepts that will recur throughout the body |
| Introduction ends with a question | Leaves readers uncertain | Follow the question immediately with the paper's answer |

---

## Heading and White Space Conventions

| Field | Convention |
|---|---|
| Natural sciences, social sciences | Section headings expected; use them to mark introduction end and conclusion start |
| Humanities (literary criticism, history) | Headings often absent; use extra blank line at frame boundaries |
| Engineering, technical writing | Headings required; number them if the document is long |
| Thesis chapters | Each chapter has its own mini-introduction and mini-conclusion; apply the same frame logic |

When in doubt about your field's convention: look at three recent articles in your target journal or publication venue and observe how the top researchers handle frame visibility.

FILE:references/key-term-threading.md
# Key Term Threading

Reference for `argument-organization-reviser` — step-by-step procedure for identifying and tracing key concepts through a multi-section paper to verify coherence.

Source: *The Craft of Research, 4th Edition*, Booth, Colomb, Williams, Bizup, FitzGerald — Chapter 13.4.

---

## What Key Terms Are (and Are Not)

**Key terms** are the specific words or phrases that name the central concepts your argument depends on. They are not the general vocabulary of your field, nor every important word in the paper.

> In a paper about the Crusades, key terms are *political concerns*, *unity*, *internal forces*, *dividing* — not *pope*, *medieval*, *religion*, or *Europe*. The latter are background vocabulary; the former are the specific concepts your argument turns on.

A well-organized paper threads three to six key terms through most of its paragraphs. When a paragraph lacks any key term, readers may feel the paper has drifted — even if the paragraph is internally coherent.

---

## Step 1: Extract Key Terms from the Main Point

Locate the main point sentence in your introduction (or conclusion if your field reserves it there).

Read it carefully. Circle the words that:
- Name what is specifically contested or argued (not just the topic)
- Name the categories, variables, mechanisms, or concepts your analysis depends on
- Would appear in the most specific, accurate search query a future reader would use to find your paper

Aim for three to six terms. If you circle more than eight, you are probably circling topic vocabulary rather than argument vocabulary — narrow the selection.

**Exercise:** Test each candidate term by asking: "If I removed this term from the paper and replaced it with a vague substitute, would my specific argument be lost?" If yes, it is a key term. If no, it may be general vocabulary.

---

## Step 2: Thread Key Terms Through the Body

With your key terms circled in the introduction, go through the body paragraph by paragraph.

For each paragraph:
1. Circle any direct use of a key term (exact word or phrase)
2. Underline any synonym, related concept, or semantically connected word (words in the same conceptual family)
3. Mark paragraphs where neither appears as potentially drifting

**Test:** Look down the page at your circles and underlines. Do at least one or two key terms appear in most paragraphs? If a stretch of three or more consecutive paragraphs lacks any key term, those paragraphs may not be contributing to the argument.

---

## Step 3: Handle Paragraphs That Lack Key Terms

For each paragraph where no key term appears:

**Option A: Add the key term.** Sometimes the paragraph makes a point that clearly belongs to the argument, but the wording has drifted. Add or substitute the key term without changing the meaning.

**Option B: Identify the connection.** Sometimes the paragraph provides evidence or context that is genuinely connected to the argument, but the connection is not explicit. Add a sentence that explicitly names how this paragraph's content relates to the key concept.

**Option C: Cut or relocate.** If you cannot add the key term or articulate the connection without forcing it, the paragraph may be off-topic. Consider cutting it or moving it to a different section where it naturally connects.

> A passage that lacks key terms usually indicates one of two things: (1) the passage has drifted from the argument, or (2) the passage belongs to a different section than the one it is currently in. Both problems are structural.

---

## Step 4: Section-Level Key Term Check

Each major section should have its own specific key terms that distinguish it from all other sections.

For each section:
1. Find the section's point sentence (what this section argues).
2. Circle the two or three terms in that sentence that are **unique to this section** — not the whole-paper key terms, but the terms that mark what *this* section contributes.
3. Check whether those section-specific terms run through the section's paragraphs.
4. Check whether those same terms appear in other sections.

**Test results:**
- Section terms run through the section: good — the section is focused.
- Section terms do not run through the section: the section's paragraphs may not actually support the section's point.
- The same section terms appear in another section: the two sections may overlap or repeat each other — consider merging or sharpening the distinction.
- A section has no unique terms (only the whole-paper key terms): the section may not be contributing a new idea to the argument. Ask: what does this section add that is not already in the other sections? If there is no answer, consider cutting it.

---

## Quick Visual Protocol

Use this to run the key-term check quickly on a printed or PDF draft:

1. Print the draft (or open it in a PDF annotator).
2. Use one color (e.g., yellow) to highlight the key terms in the introduction point sentence.
3. Use the same color to highlight every appearance of those terms in the body.
4. Use a different color (e.g., blue) to underline related/synonymous terms.
5. Scan the page: are the yellow and blue marks distributed fairly evenly? Are there sections with no yellow or blue at all?

This visual scan often reveals coherence gaps in thirty seconds that a re-reading would miss entirely — because when we re-read, we supply the connections ourselves.

---

## Example

**Paper:** On the political motivations behind the First Crusade.

**Main point sentence:** "In a series of documents, the popes proposed their Crusades to restore Jerusalem to Christendom, but their words suggest other issues involving **political concerns** about European and Christian **unity** in the face of **internal forces** that were **dividing** them."

**Key terms identified:** *political concerns*, *unity*, *internal forces*, *dividing*

**Section 1 check:** Section point — "Gregory VII's 1074 letter reveals concern about Byzantine fragmentation threatening papal authority." Unique terms: *Byzantine fragmentation*, *papal authority*. These run through Section 1. Good.

**Problem found:** A three-paragraph passage in Section 2 discusses agricultural and trade conditions in eleventh-century Europe with no mention of political concerns, unity, internal forces, or their synonyms. Decision: add a sentence explicitly connecting economic instability to internal forces dividing the Western church, or cut the passage if the connection cannot be made.

---
name: tailor-pitch-by-stakeholder
description: |
  Tailor a Commercial Teaching pitch to each stakeholder role using Functional Bias Cards and route the pitch to avoid the C-suite elevation trap.

  Trigger this skill when you need to:
  - Tailor a sales pitch to different roles in a buying committee
  - Build stakeholder-specific messaging for a multi-stakeholder sale
  - Create a different message for VP Engineering vs CFO vs procurement
  - Use a Functional Bias Card to map what each role cares about
  - Build a stakeholder map for an enterprise deal
  - Understand who to pitch to first and in what order
  - Adapt a pitch for different audience roles without losing the core insight
  - Tailor for resonance with a buying committee
  - Create per-role pitch variants from a single base pitch script
  - Build consensus across a buying group before reaching the decision-maker

  NOT for: building the base pitch (use author-commercial-teaching-pitch first),
  building the underlying commercial insight (use build-commercial-insight),
  or diagnosing an existing pitch's structure (use diagnose-pitch-for-commercial-teaching-fit).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/tailor-pitch-by-stakeholder
metadata:
  openclaw:
    emoji: "🧭"
    homepage: "https://github.com/bookforge-ai/bookforge-skills"
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors:
      - Matthew Dixon
      - Brent Adamson
    chapters:
      - 6
tags:
  - sales
  - b2b-sales
  - challenger-sale
  - stakeholder-mapping
  - pitch-tailoring
  - sales-messaging
depends-on:
  - author-commercial-teaching-pitch
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "pitch-script.md (from author-commercial-teaching-pitch) and stakeholder-map.md with the roles and titles present in this buying group"
  tools-required:
    - Read
    - Write
    - AskUserQuestion
  works-offline: true
discovery:
  goal: "Tailor a Commercial Teaching pitch to each stakeholder role using Functional Bias Cards and route the pitch to avoid the C-suite elevation trap. Outputs tailored-pitch-variants.md with per-role Functional Bias Cards, per-role pitch variants, and a routing plan."
  triggers:
    - "Tailor my pitch to different stakeholders"
    - "I need different messaging for the CFO vs the VP Engineering"
    - "Who should I pitch to first in a consensus sale?"
    - "Build a Functional Bias Card for my buyer roles"
    - "How do I adapt my commercial teaching pitch for multiple roles?"
    - "Create per-role pitch variants for my buying committee"
    - "Help me build stakeholder-specific messaging"
  audience:
    - Sales reps and account executives managing multi-stakeholder enterprise deals
    - Sales enablement teams building role-based message libraries
    - Product marketing managers translating pitch scripts into role-specific assets
    - Sales managers coaching reps on stakeholder engagement strategy
  prerequisites:
    - "Run author-commercial-teaching-pitch first to produce pitch-script.md — this skill adapts that artifact, not a generic pitch"
  not_for:
    - Building the base pitch — use author-commercial-teaching-pitch
    - Building the commercial insight — use build-commercial-insight
    - Diagnosing pitch structure — use diagnose-pitch-for-commercial-teaching-fit
---

# tailor-pitch-by-stakeholder

Adapt a completed Commercial Teaching pitch to each role in a buying committee. The base pitch from `author-commercial-teaching-pitch` contains the right insight — this skill changes the language, data points, concerns, and framing for each person who needs to hear it. Outputs `tailored-pitch-variants.md` with per-role Functional Bias Cards, adapted pitch variants, and a routing plan that builds bottom-up consensus before reaching the decision-maker.

---

## When to Use

Use this skill after `author-commercial-teaching-pitch` has produced `pitch-script.md`. The base pitch was built for a target segment, not for every individual in a buying group. In complex B2B sales, multiple people participate in a purchase but each one evaluates the deal through a different lens — their own role, metrics, concerns, and career objectives.

A pitch that lands with a VP of Operations will confuse a CFO. The same Reframe can either energize or alienate depending on whether it maps to what that person measures. This skill ensures the core insight reaches every stakeholder in their own language.

---

## Context and Input Gathering

### Load the prerequisite artifacts

Read `pitch-script.md` from `author-commercial-teaching-pitch`. Extract:
- The Reframe headline (the core insight that must remain consistent across all variants)
- The unique strength anchor (what only your solution can deliver)
- The target segment and the Rational Drowning data already in the pitch
- The New Way framing (the organizational change the customer agrees to in Step 5)

If `pitch-script.md` does not exist, ask the user to run `author-commercial-teaching-pitch` first. Do not proceed without it — this skill adapts an existing pitch; it does not build one from scratch.

### Load or elicit the stakeholder map

If `stakeholder-map.md` exists, read it. Extract:
- Each unique role / title present in the buying group
- Whether each person is a decision-maker (signs the agreement) or an influencer/end-user (participates but does not sign)
- Any known priorities, concerns, or history already gathered through discovery

If `stakeholder-map.md` does not exist, use `AskUserQuestion` to collect:
1. What roles are involved in this purchase? (e.g., "CFO, VP Engineering, Head of IT, procurement manager")
2. Which of these people ultimately signs the agreement?
3. What industry and company context applies?
4. What is already known about each person's priorities or concerns?

---

## Process

### Step 1 — Apply the Four-Layer Tailoring Frame

Before adapting language per role, establish the shared outer layers that apply to all stakeholders.

**Layer 1 — Industry context:**  
What trends, competitive events, or regulatory pressures are active in this customer's sector? Name 2-3 that are undeniable. These provide the backdrop all stakeholders share.

**Layer 2 — Company context:**  
What does this specific company's recent history show? Recent earnings themes, strategic priorities, announced initiatives, or market position shifts. Pull from public sources: press releases, earnings calls, investor day materials.

**Layer 3 — Role context (Functional Bias):**  
For each unique role, build a Functional Bias Card (see Step 2 below).

**Layer 4 — Individual context:**  
Anything known specifically about this person: career stage, past decisions, internal political position, stated priorities, any history with your company. Fill in from discovery notes where available; leave as a gap to fill during live conversation where not.

**Why:** The vast majority of sales messaging is not contextualized at any level. Industry and company layers are table stakes — they establish that the rep has done their homework. Role-level and individual-level tailoring are what differentiate Challenger reps from core performers and are the layers most likely to produce the reaction "you understand my world."

---

### Step 2 — Build a Functional Bias Card for Each Unique Role

For each role in the stakeholder map, complete the following four-component card. This is the operational unit of tailoring.

**The Functional Bias Card pattern comes from the Solae case** — the food-ingredients supplier that operationalized role-level tailoring across their entire sales force using pre-built cards per buyer role. The skill's output MUST cite the Solae pattern by name so the user understands this is a proven, scaled approach — not an improvised artifact. Include a one-line attribution ("Functional Bias Cards as implemented at Solae") in the `tailored-pitch-variants.md` output.

**Card format:**

```
ROLE: [Title]
STAKEHOLDER TYPE: [Decision-maker | Influencer/End-user]

(a) DECISION CRITERIA — What does this role define as success?
    What are they measured on? What outcomes make this person look good?
    Format outcomes as: "[verb] [metric] by [magnitude] in [context]"

(b) FOCUS — What metrics does this person monitor daily or weekly?
    These are the numbers they pull up, the dashboards they check.
    Use these metrics when you cite data in the pitch for this role.

(c) KEY CONCERNS — What does this person worry about?
    What questions do they ask themselves day-to-day?
    What risks can keep them from saying yes?

(d) POTENTIAL VALUE AREAS — What levers can this person pull?
    What actions, investments, or decisions are within their authority?
    This is the vocabulary to use when presenting your solution to this person.
```

**Why each field matters:** Decision criteria tell you which outcomes anchor the pitch. Focus fields tell you which data will be credible. Key concerns allow you to open with empathy — you already know what they're worried about, so you do not need to ask. Potential value areas are the language layer: they tell you how to translate your capability into this person's world.

**Scale note:** Role-level outcomes are predictable (a CFO's top concerns at one company predict a CFO's concerns at another), stable over time, and finite (3-5 outcomes per role). A central marketing or enablement team can build these cards once and deploy them across the entire sales force.

See `references/functional-bias-card.md` for the full template, the Solae case walkthrough, and the outcome statement format.

---

### Step 3 — Map the Pitch to Each Role's Functional Bias Card

Return to `pitch-script.md` and work through each step of the pitch for each role:

For each role's Functional Bias Card, identify:

- **Warmer adaptation:** Which of the 2-3 challenge hypotheses are most relevant to this role's key concerns? Swap out challenges that do not match this person's world.
- **Reframe adaptation:** The core insight stays the same. What changes is the language — restate the Reframe using this role's decision criteria vocabulary. A CFO hears the Reframe in revenue/cost terms; a VP Engineering hears it in system reliability terms; a procurement officer hears it in supplier risk terms.
- **Rational Drowning adaptation:** Which data points from the pitch match this role's focus metrics? Replace generic data with the numbers this person monitors. A CFO cares about margin impact; an operations director cares about throughput and cycle time.
- **Emotional Impact adaptation:** Does the story in the pitch feature a character in a similar role? If not, adapt the scenario protagonist to this role. The story must feature someone this stakeholder can recognize as "like me."
- **New Way adaptation:** Which capabilities in the New Way directly map to this role's potential value areas? Lead with those. The CFO hears about the financial control capability; the head of operations hears about workflow efficiency.
- **Your Solution adaptation:** Which of your unique capabilities map to this role's specific value areas? Do not present every feature — present only the features that resolve this person's specific concerns and outcomes.

Document the changes per role. Not every step requires a different script — sometimes only the data points change, sometimes only the story protagonist changes. But every role should have at least one adapted element that would not have landed without the Functional Bias Card.

---

### Step 4 — Apply the Decision-Maker vs. Influencer Routing Rule

Classify each stakeholder:

**Decision-makers** (people who sign the agreement — typically senior executives or procurement):
- What they care about most: "widespread support for this supplier across my organization"
- What drives their loyalty: the overall sales experience of the buying group, not individual rep performance
- They think of themselves as buying from an organization, not from a person
- **Pitch goal:** Show that the team believes in this. Build a case for consensus, not a polished solo performance.
- **Pitch content:** Emphasize organizational outcomes, ease of doing business, ability to collaborate across functions.

**Influencers and end-users** (people who participate but do not sign):
- What they care about most: the individual rep's professionalism, unique perspectives, and ability to teach them something new
- They buy from people, not organizations
- The top drivers of their loyalty are insight and education — discovering something they did not know about how to be more effective in their role
- **Pitch goal:** Teach them something genuinely new. The Reframe lands hardest here — if they have never considered this angle before, they become your most effective internal advocates.
- **Pitch content:** Emphasize the insight, the fresh perspective, the specific outcome for their role.

**Why this split matters:** The research underlying this framework found that for decision-makers, organizational sales experience attributes are nearly twice as important as individual rep attributes — the reverse is true for influencers. The same pitch strategy cannot optimize for both simultaneously. Sequence and framing must differ.

---

### Step 5 — Anti-Pattern Check: C-Suite Elevation Trap

Before finalizing the routing plan, apply this check:

**The trap:** Conventional sales training emphasizes getting direct access to the senior decision-maker as quickly as possible. The logic: "If we can just get in that door, that's going to help us close the deal."

**Why it fails:** Decision-makers' top loyalty driver is "widespread support for the supplier across my organization." They are not willing to go out on a limb for a supplier on a large purchase without team consensus behind it. A polished top-down pitch from a sales rep carries far less weight with the decision-maker than confirmation from their own team.

**The stronger path:** Build genuine value with influencers and end-users first. Teach them something they did not know. When they become convinced, they evangelize upward — and their word carries more weight than anything the rep says directly to the executive.

**Flag if the current plan includes any of the following:**
- Prioritizing an executive meeting before any team-level conversations
- Using team interactions primarily to gather intelligence rather than deliver insight
- Planning to present a "finely tuned pitch" to the senior decision-maker using information pulled from stakeholders — rather than using those interactions to genuinely teach the stakeholders something valuable

**If flagged:** Reorder the routing plan so influencer/end-user conversations precede the decision-maker conversation. The goal is not to avoid the decision-maker — it is to arrive at that meeting with documented consensus already in place.

---

### Step 6 — Build the Value Planning Tool (Optional — Recommended for Late-Stage Deals)

Once stakeholder conversations are progressing, use this stage-gate template to document per-stakeholder buy-in before the final close conversation:

**For each stakeholder:**

| Stakeholder Role | Specific Outcome Addressed | Strongest Concern / Objection | Capability Response | Sign-Off |
|-----------------|---------------------------|------------------------------|-------------------|----------|
| [Role] | [Outcome in their vocabulary] | [Their concern, stated as they would say it] | [Specific thing you will do or deliver] | [Optional signature] |

**How to use this with the decision-maker:**  
When the rep sits down to close, place this document on the table. Every stakeholder's specific outcome, concern, and agreed response is visible in one view. The decision-maker does not need to take the rep's word for team support — it is documented on one page.

**Optional best practice:** Ask each stakeholder to initial their column as conversations progress. This is not a formal contract. It is a working alignment tool that creates a micro-commitment, making reversal harder when the full deal reaches the decision-maker.

---

### Step 7 — Write tailored-pitch-variants.md

Create the output artifact with the following structure:

```markdown
# Tailored Pitch Variants
**Base pitch:** [link or reference to pitch-script.md]
**Deal/account:** [Customer name or deal identifier]
**Date:** [Date]

---

## Layer 1: Industry Context (all stakeholders)
[2-3 industry trends or events that apply to everyone in this buying group]

## Layer 2: Company Context (all stakeholders)
[2-3 company-specific facts: strategic priorities, recent announcements, market position]

---

## Stakeholder 1: [Title]
**Type:** Decision-maker | Influencer/End-user

### Functional Bias Card
**(a) Decision Criteria:**
- [Outcome 1]
- [Outcome 2]

**(b) Focus (monitored metrics):**
- [Metric 1]
- [Metric 2]

**(c) Key Concerns:**
- "[Worry 1]"
- "[Worry 2]"

**(d) Potential Value Areas:**
- [Lever 1]
- [Lever 2]

### Pitch Adaptations (what changes from the base)
- **Warmer:** [Challenge hypotheses most relevant to this role]
- **Reframe:** [How the core insight is re-stated in this role's vocabulary]
- **Rational Drowning:** [Data points that match this role's focus metrics]
- **Emotional Impact:** [Scenario protagonist adjusted to this role]
- **New Way:** [Capabilities emphasized that match this role's value areas]
- **Your Solution:** [Features to lead with for this stakeholder]

### Routing Priority: [Early / Mid / Final]
[1-2 sentences on when and why to engage this person in the sequence]

---

[Repeat per stakeholder]

---

## Routing Plan
**Sequence:**
1. [Stakeholder A] — [why first]
2. [Stakeholder B] — [why second]
3. [Decision-maker] — [deliver after documented consensus is in place]

**C-suite elevation check:** [PASSED / FLAG — describe if flagged]

---

## Value Planning Tool (when deal is in late stage)
| Role | Outcome Addressed | Concern | Capability Response | Sign-Off |
|------|------------------|---------|-------------------|----------|
| [Role] | | | | |
```

---

## Inputs

| Input | Required | Description |
|-------|----------|-------------|
| `pitch-script.md` | Yes | Output of `author-commercial-teaching-pitch` — base six-step pitch with Reframe, Rational Drowning data, and unique strength anchor |
| `stakeholder-map.md` | Yes (or elicit) | Roles, titles, and decision-maker / influencer classification for this buying group |
| Discovery notes | Optional | Known priorities, concerns, and history for specific individuals — used to populate Layer 4 (individual context) |

---

## Outputs

| Output | Path | Description |
|--------|------|-------------|
| `tailored-pitch-variants.md` | `./tailored-pitch-variants.md` | Per-role Functional Bias Cards, per-role pitch adaptations, routing plan, and Value Planning Tool template |

---

## Key Principles

**The Reframe stays constant; the language changes.** The core commercial insight must remain consistent across all role variants — if the insight is correct, it applies to the whole organization. What changes is the vocabulary, the metrics cited, the story protagonist, and the capabilities emphasized. Never dilute the Reframe to make it less challenging for a specific role.

**Decision-makers buy from organizations; influencers buy from people.** These are not personality types — they are structural differences in how these two groups evaluate a purchase. Organizational-level attributes (widespread support, ease of doing business) dominate for decision-makers; individual rep insight and education dominate for influencers. Pitch strategy must reflect this split.

**The path to the C-suite runs through the team, not directly to the executive.** Reps who bypass stakeholders to reach the decision-maker faster undermine themselves. The decision-maker's top priority is team consensus — which the rep cannot provide but which the team, once taught something valuable, absolutely can.

**Functional Bias Cards are role-level, not person-level.** A card for "VP of Operations" at one company predicts with reasonable accuracy what a VP of Operations at a comparable company cares about. This is what makes tailoring scalable — a central team builds the cards once; every rep uses them across the portfolio.

**Teaching influencers is how you build the consensus decision-makers require.** Influencers advocate most powerfully for suppliers who taught them something genuinely new. Mining them for intelligence makes them vendors; teaching them something valuable makes them advocates.

---

## References

- `references/functional-bias-card.md` — Full template, four-component walkthrough, Solae case study (head of manufacturing card + Value Planning Tool), and outcome statement format

**Source:** *The Challenger Sale* by Matthew Dixon and Brent Adamson. Chapter 6 (Tailoring for Resonance). Solae case study: functional bias card and value planning tool implementation. Loyalty driver data from CEB Customer Loyalty Survey (2011).

---

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). You are free to use, adapt, and redistribute with attribution.

Source book content is copyrighted by the authors. This skill contains no verbatim passages — all content is paraphrased, structured, and extended for agent execution.

---

## Related BookForge Skills

| Skill | Relationship |
|-------|-------------|
| `author-commercial-teaching-pitch` | Run before this skill — produces the pitch-script.md that this skill adapts |
| `build-commercial-insight` | Run before author-commercial-teaching-pitch — validates the Reframe this pitch chain is built on |
| `diagnose-pitch-for-commercial-teaching-fit` | Alternative entry — audit an existing pitch before building or tailoring |
| `plan-negotiation-with-constructive-tension` | Run after tailoring — manages the negotiation dynamic once stakeholders are engaged |

FILE:references/functional-bias-card.md
# Functional Bias Card — Template and Solae Case Walkthrough

**Source:** *The Challenger Sale*, Chapter 6 (Tailoring for Resonance)  
**Usage:** Reference for the `tailor-pitch-by-stakeholder` skill. Complete one card per unique stakeholder role before adapting the pitch.

---

## The Four-Component Card Template

For each stakeholder role, populate the following fields:

```
ROLE: [Title / Function]

(a) DECISION CRITERIA — Business outcomes this role is accountable for
    • [Outcome 1] — format: "[verb] [metric] by [magnitude] in [context]"
    • [Outcome 2]
    • [Outcome 3]

(b) FOCUS — Metrics this person monitors daily/weekly to track progress
    • [Metric 1]
    • [Metric 2]

(c) KEY CONCERNS — Daily questions and worries this role carries
    • "[Question or worry 1]"
    • "[Question or worry 2]"

(d) POTENTIAL VALUE AREAS — Levers this role can pull to improve performance
    • [Lever 1] — how your solution maps to this lever
    • [Lever 2]
```

**Why each field matters for pitch adaptation:**

- **Decision criteria** tell you which outcomes to anchor your pitch to for this role. If you don't cite outcomes in this person's vocabulary, they respond "so what?"
- **Focus** tells you which data points and metrics will land as credible. Use their metrics, not your metrics, in the Rational Drowning step.
- **Key concerns** let you open the conversation with empathy without having to ask "What's keeping you up at night?" — you already know, which builds immediate credibility.
- **Potential value areas** provide the language to translate your capabilities into this person's world. A head of manufacturing does not buy "AI-powered quality analytics." They buy "reduce scrap rate on the high-capacity line."

---

## Properties of Role-Based Outcomes

Role-level outcomes are the right level of abstraction for scale:

| Property | Implication |
|----------|-------------|
| **Predictable** | The top outcomes for a VP of Supply Chain at Company A predict those for the same role at Company B |
| **Stable** | When a person leaves and is replaced, the new hire typically inherits similar outcome targets |
| **Finite** | Each role has 3-5 dominant outcomes — not infinite |
| **Scalable** | A marketing or enablement team can build these cards centrally and deploy to all reps |

This means Functional Bias Cards can be built once by a central team and used across the entire sales force — they are not improvised on a per-deal basis.

---

## Solae Case: Head of Manufacturing Card (Paraphrased)

The following illustrates what a completed Functional Bias Card looks like, paraphrased from the Solae implementation described in the book.

```
ROLE: Head of Manufacturing / VP Manufacturing

(a) DECISION CRITERIA
    • Maintain or improve production throughput
    • Reduce input costs per unit
    • Maintain product quality and consistency standards
    • Minimize unplanned downtime and line stoppages

(b) FOCUS
    • Daily/shift yield rates
    • Reject and scrap rates by line
    • Equipment utilization and downtime logs
    • Input material cost per batch

(c) KEY CONCERNS
    • "Will a formula change disrupt my throughput or quality consistency?"
    • "Can I cut input costs without my customer noticing a difference?"
    • "What happens if this change creates a recall risk?"
    • "How much line time will this require to test and validate?"

(d) POTENTIAL VALUE AREAS
    • Optimize input ingredient ratios without quality impact
    • Reduce ingredient costs that are fluctuating in commodity markets
    • Improve yield consistency by reducing batch-to-batch variation
    • Streamline new ingredient validation process
```

**How Solae used this card in a pitch:**  
Rather than opening with technical specifications of their soy protein product, a Solae rep could open with: "When we work with heads of manufacturing in your sector, the top concern we hear is protecting throughput when reformulating — even a small batch-to-batch variation can disrupt the line. Is that something you're managing right now?"

This is the Functional Bias Card doing its job: enabling the rep to open with a hypothesis derived from the card, not a question asking the customer to educate them.

---

## Solae Case: Value Planning Tool (Paraphrased)

Once a deal is in proposal stage, Solae used a per-stakeholder table to document and secure individual buy-in before the final close.

**Structure:**

| Stakeholder Role | Outcome Addressed | Strongest Concern / Objection | Solae Capability Response | Agreed (sign-off) |
|-----------------|-------------------|------------------------------|--------------------------|-------------------|
| Head of Manufacturing | Maintain throughput during reformulation | Line disruption risk from new ingredient | Pilot trial protocol + QA support during validation period | [signature line] |
| VP Marketing / CMO | Maintain or improve consumer taste perception | Reformulation risks brand quality perception | Blind consumer testing data showing taste parity | [signature line] |
| Procurement | Reduce input material costs by X% | Price volatility exposure | Long-term supply contract with indexed pricing | [signature line] |

**How this was used at close:**  
When the Solae rep sat down with the senior decision-maker to close, they placed this document on the table. Every stakeholder's specific outcome, concern, and agreed response was visible in one view. The decision-maker did not need to take the rep's word for team support — it was documented. This is the "widespread support" that decision-makers require before committing.

**Optional best practice:** Ask each stakeholder to initial their column as conversations progress. This is not a formal contract — it is a working alignment tool. The act of signing creates a micro-commitment that makes reversal harder when the full deal reaches the decision-maker.

---

## Outcome Statement Format

When documenting stakeholder outcomes, use this structure for precision:

`[direction] [metric] by [magnitude] in [context]`

| Component | Example |
|-----------|---------|
| Direction | Decrease, increase, maintain, eliminate |
| Metric | Reject rate, conversion rate, cycle time, cost per unit |
| Magnitude | 5%, $2M, 3 days, 20 basis points |
| Context | On the high-capacity line, in North America, during Q3 close |

**Full example:** "Decrease reject rate by 5% on our high-capacity production line by Q2."

This format makes the outcome concrete enough to anchor a specific pitch argument, and clear enough that the stakeholder can confirm or correct it.

---

**Source:** *The Challenger Sale* by Matthew Dixon and Brent Adamson, Chapter 6. Paraphrased for agent execution — no verbatim text reproduced. Licensed CC-BY-SA 4.0.

---
name: plan-negotiation-with-constructive-tension
description: |
  Build a pre-call negotiation plan using DuPont's four-step framework and sequence concessions to avoid the proactive-discount and ultimatum traps.

  Trigger this skill when you need to:
  - Plan a negotiation for a B2B sales deal before a pricing or concession conversation
  - Respond to a customer discount request without immediately caving on price
  - Counter a price reduction demand by broadening the negotiation beyond price
  - Structure your concession sequence so you trade low-value items first and protect margin
  - Avoid proactive discounting, escalating concession patterns, or ultimatum traps
  - Apply the DuPont four-step negotiation framework (Acknowledge & Defer → Deepen & Broaden → Explore & Compare → Concede According to Plan)
  - Use the Situational Sales Negotiation (SSN) pre-call planning template to score concessions before the call
  - Build scripted deferral language to buy time without threatening the deal
  - Prepare for a challenger negotiation with constructive tension
  - Draft a pre-call negotiation worksheet for an upcoming customer conversation

  NOT for: diagnosing whether you have a taking-control problem in the first place — run diagnose-taking-control-gaps first to produce the control diagnosis this skill consumes.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/plan-negotiation-with-constructive-tension
metadata:
  openclaw:
    emoji: "⚖️"
    homepage: "https://github.com/bookforge-ai/bookforge-skills"
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors:
      - Matthew Dixon
      - Brent Adamson
    chapters:
      - 7
tags:
  - sales
  - b2b-sales
  - challenger-sale
  - negotiation
  - concession-management
  - sales-control
depends-on:
  - diagnose-taking-control-gaps
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "deal-brief.md + taking-control-diagnosis.md + counter-proposal.md (or equivalent deal context including the customer's specific ask, deal size, and available solution elements)"
  tools-required:
    - Read
    - Write
    - AskUserQuestion
  works-offline: true
discovery:
  goal: "Build a pre-call negotiation plan using DuPont's four-step framework and sequence concessions to avoid the proactive-discount and ultimatum traps."
  audience:
    - B2B sales representatives preparing for a pricing or concession conversation
    - Account executives managing enterprise deals with procurement-driven pressure
    - Sales managers coaching reps on negotiation discipline and concession sequencing
  prerequisites:
    - "diagnose-taking-control-gaps produces the taking-control-diagnosis.md this skill consumes"
  not_for:
    - Posture diagnosis and foil RFP detection → use diagnose-taking-control-gaps
    - Building a commercial teaching pitch → use build-commercial-insight
---

# Plan Negotiation with Constructive Tension

## Overview

Most reps lose value in negotiations not because customers are more powerful, but because reps seek premature closure. When a customer asks for a discount, the rep's instinct is to resolve the tension immediately — by conceding. The DuPont four-step framework, built on BayGroup International's Situational Sales Negotiation (SSN) methodology, is a pre-call planning and in-call navigation system that equips any rep to hold position assertively without crossing into aggression.

The framework works in two phases:

1. **Before the call** — complete the SSN pre-call worksheet to score every possible concession by cost-to-supplier and value-to-customer. This converts the call from an improvised reaction to a planned negotiation.
2. **During the call** — navigate four named steps in sequence. Skipping steps is the primary cause of premature concessions.

This skill produces a `negotiation-plan.md` file containing the completed worksheet and scripted language for each step.

---

## When to Use

Use this skill when:
- A customer has made a specific demand (price reduction, extended terms, added scope) and you have an upcoming conversation to respond
- You have completed `diagnose-taking-control-gaps` and have a `taking-control-diagnosis.md` indicating passive posture or concession risk
- A deal is at or near commercial terms and you need a structured plan for the negotiation conversation
- You want to ensure concessions are sequenced (low-value first) rather than improvised under pressure

**Run `diagnose-taking-control-gaps` first** if you have not yet established whether you have a control gap. This skill assumes the diagnosis is complete and the rep has decided to plan a negotiation, not diagnose a deal.

---

## Step 1 — Load Deal Artifacts

Read all available deal documents from the working directory.

**Required (at least one):**
- `deal-brief.md` — account name, deal size, deal stage, what the customer is asking for (the demand), what solution elements are on the table
- `taking-control-diagnosis.md` — output of `diagnose-taking-control-gaps`; tells you the rep's behavioral posture and whether active misconceptions are present

**Optional but high-value:**
- `counter-proposal.md` — the customer's formal counter-proposal or written demand
- `discovery-notes.md` — customer's underlying business objectives, key stakeholders, what they care about beyond price
- `solution-capabilities.md` — full inventory of what can be traded (service tiers, terms, delivery options, support packages)

**If no documents exist**, ask the user:

1. What is the customer asking for? (specific demand — e.g., "20% price reduction before signing")
2. What is the deal size and current stage?
3. What solution elements are on the table beyond price? (service plans, terms, implementation support, delivery timelines)
4. What does the customer care about most beyond price, if known?
5. What is the supplier's walkaway position — the floor below which the deal is not worth doing?

Why this matters: the SSN worksheet in Step 2 cannot be scored without knowing what is on the table. The scripted language in Steps 3–6 cannot be calibrated without knowing the specific customer demand.

---

## Step 2 — Complete the SSN Pre-Call Worksheet

Complete the six-section SSN worksheet before drafting any call language. This is the planning foundation that enables assertive behavior in the call. Challengers do this intuitively; this worksheet makes it explicit for all reps.

See `references/ssn-template.md` for the full template with scoring tables.

**Section 1 — Power Positions**

List the supplier's strengths and weaknesses across five dimensions: products/capabilities, brand/reputation, pricing/value ratio, service/support, customer relationships. Score each 1–10.

Purpose: most reps underestimate their leverage. Writing strengths down shifts the psychological default away from immediate concession. A rep who knows they have a 9/10 in technical implementation support is harder to move off price than one who hasn't thought about it.

**Section 2 — Information to Get from Customer**

List the questions to ask in the Deepen and Broaden step. The most important: "What are you looking to achieve with a [X]% price reduction?" — this surfaces whether the demand is driven by budget constraint or a specific business outcome that can be satisfied another way.

**Section 3 — Information to Provide or Protect**

Decide in advance what information to share with the customer and what to withhold. Deciding this during a live negotiation under pressure leads to disclosing too much or too little.

**Section 4 — Anticipated Objections and Planned Responses**

Write out 3–5 likely objections and draft specific responses. Improvised responses under pressure almost always result in unnecessary concessions.

**Section 5 — Supplier Goals and Customer Hypotheses**

State the supplier's non-negotiable terms (scope items, price floor, delivery requirements). Then write at least two hypotheses about the customer's underlying need — the business outcome behind the stated demand. These hypotheses drive the Deepen and Broaden conversation.

**Section 6 — Concession Inventory with Scoring**

This is the operational core of the SSN template. For each potential concession:
- Score cost to supplier (1 = minimal cost, 10 = very expensive)
- Score value to customer (1 = low value, 10 = high value)

Identify the concessions that are high-value to the customer but low-cost to the supplier — offer these first. Price concessions typically score high on cost to supplier and low-to-moderate on customer value, which is why they should come last, not first.

Also list concessions to request from the customer in exchange for each offer. Concessions should be traded, not donated.

---

## Step 3 — Draft Step 1 Language: Acknowledge and Defer

Write the scripted language the rep will use to defer the customer's demand without dismissing it.

**The key principle:** Deferring without customer permission is aggressive. Winning permission first makes the same deferral assertive. The script must do three things: acknowledge the demand, promise eventual closure, and ask permission to defer.

**Template (adapt to deal context):**

> "I understand that [price / the discount / the terms change] is something we need to address, and I want to make sure we do. Before we get there, I'd like to take a few minutes to make sure I fully understand what you're trying to achieve — so we can look at everything we can do to make this deal as valuable as possible for you. Is that all right?"

**What makes this work:**
- "I understand" — acknowledges the demand without committing to it
- "I want to make sure we do" — promises the customer they will get a resolution
- "Is that all right?" — seeks explicit permission; without this, the deferral reads as avoidance

Write the deal-specific version of this language in `negotiation-plan.md`.

---

## Step 4 — Draft Step 2 Language: Deepen and Broaden

Write the questions and prompts the rep will use to expand the negotiation beyond the single stated demand.

**Purpose:** Price is rarely the only thing the customer cares about. Deepen and Broaden surfaces what else matters — warranty coverage, service response time, implementation timeline, payment terms, future roadmap commitments — so that price is no longer the only negotiable.

**Technique — start with what the customer already values:**

Begin by getting the customer to restate what they like about the offering. This activates positive framing before returning to the demand. Then introduce the broadening questions.

**Core broadening questions (adapt to deal):**

1. "What are you looking to achieve with a [X]% price reduction?" — surfaces the business rationale; the answer often reveals a non-price solution
2. "Beyond price, what other elements of this deal are most important to you in terms of making this work?" — expands the negotiable surface
3. "If we were able to [address the underlying need another way], how would that compare to the price reduction you're looking for?"

**Write 2–3 deal-specific questions** in `negotiation-plan.md` based on what is known about the customer's business situation from the discovery notes.

---

## Step 5 — Draft Step 3 Language: Explore and Compare

Using the concession scoring from Section 6 of the SSN worksheet, draft the trade offers the rep will put on the table during the Explore and Compare conversation.

**Purpose:** Convert the expanded negotiable universe (identified in Step 2) into concrete trade proposals. The SSN scoring determines which concessions to offer first.

**Trade structure:**

Prioritize concessions where cost-to-supplier is low and value-to-customer is high. Present these as genuine value moves, not consolation prizes. Frame each as a trade with a reciprocal ask:

> "If we [extend the service plan by six months / move the delivery date up / include implementation support], would that address the concern you're trying to solve with the price reduction?"

> "If I can [do X], I'd want [ask for Y in return] — does that work for you?"

**Concession sequencing rules:**
1. Never skip to price while non-price options remain on the table
2. Always frame concessions as trades — what the rep gives must come with something the customer gives
3. Follow the SSN scoring order — highest customer value / lowest supplier cost comes first
4. Hold price concessions as the final lever, not the first

Write the ordered list of trade proposals in `negotiation-plan.md`.

---

## Step 6 — Draft Step 4 Language: Concede According to Plan

Write the concession sequence the rep will follow if Steps 2–3 do not resolve the customer's demand and a price concession is required.

**The core rule:** Start with a meaningful concession, then offer progressively smaller concessions as negotiations continue. This is the inverse of the escalating pattern.

**Why this works:** The customer perceives movement and good faith from the initial meaningful concession. The decreasing size of subsequent concessions signals that the rep is approaching their limit, which motivates the customer to close rather than push further. Both parties end the negotiation feeling they won.

**Concession plan structure (write this in negotiation-plan.md):**

| Round | Concession | Amount / Scope | Trigger | Reciprocal Ask |
|-------|------------|----------------|---------|----------------|
| 1 | (First offer — meaningful) | | If they reject the non-price trades from Step 5 | (what to ask in return) |
| 2 | (Second — smaller) | | If they reject Round 1 | (what to ask in return) |
| 3 | (Third — smaller still) | | If they reject Round 2 | (what to ask in return) |
| Walkaway | — | — | If they reject Round 3 | Disengage protocol |

**Before executing any concession in this plan:** verify the concession is in the pre-planned sequence. If the customer is pushing for a concession not in the plan, pause, reference Section 5 (supplier goals and non-negotiables), and respond deliberately.

---

## Step 7 — Anti-Pattern Check

Before finalizing `negotiation-plan.md`, review the plan against the three named concession anti-patterns. See `references/concession-antipatterns.md` for full descriptions.

**Check 1 — Escalating concessions:**
Is each concession in your plan smaller than the previous one? If Round 2 is larger than Round 1, reorder the sequence.

**Check 2 — Proactive discounting:**
Does the plan include any concession offered before the customer has made a specific ask? Remove it. Concessions are given in response to demands, not in anticipation of them.

**Check 3 — Ultimatums:**
Does the plan use language like "this is our final offer" or "take it or leave it" before the walkaway point has been reached? Replace with the Acknowledge and Defer language from Step 3. Reserve the walkaway statement for the genuinely planned walkaway — the position identified in Section 5 of the SSN template.

---

## Step 8 — Write negotiation-plan.md

Produce `negotiation-plan.md` in the working directory with the following structure:

```
# Negotiation Plan — [Deal Name / Account]
Date: [date]
Customer demand: [specific ask]
Deal context: [deal size, stage, key stakeholders]

## SSN Pre-Call Worksheet Summary
[Power position highlights — top 3 supplier strengths]
[Key information to gather — top 3 questions]
[Concession inventory — scored table]
[Walkaway terms]

## Step 1 — Acknowledge and Defer Script
[Deal-specific scripted language]

## Step 2 — Deepen and Broaden Questions
[3–5 deal-specific questions to expand the negotiation]

## Step 3 — Explore and Compare: Trade Offers
[Ordered list of trade proposals with framing language]

## Step 4 — Concede According to Plan
[Concession sequence table with triggers and reciprocal asks]

## Anti-Pattern Status
[ ] Escalating concessions — [CLEAR / FLAG: describe]
[ ] Proactive discounting — [CLEAR / FLAG: describe]
[ ] Ultimatums — [CLEAR / FLAG: describe]
```

---

## Self-Check Before Delivering the Plan

Before writing `negotiation-plan.md`, verify:

- [ ] SSN concession scoring is complete — every potential concession has a cost-to-supplier AND a value-to-customer score
- [ ] Acknowledge and Defer script explicitly seeks customer permission ("Is that all right?" or equivalent)
- [ ] Deepen and Broaden includes the rationale question: "What are you looking to achieve with [X]%?"
- [ ] Explore and Compare trade offers follow the SSN scoring order (high-value/low-cost first)
- [ ] Concession sequence gets smaller each round, not larger
- [ ] No concession in the plan is offered before the customer makes a specific ask
- [ ] Walkaway terms are defined in the SSN worksheet, not invented during the call
- [ ] Anti-pattern check passed for all three patterns
- [ ] No verbatim book passages appear in the output artifact

---

## Reference Materials

- `references/ssn-template.md` — Full SSN pre-call planning template with scoring tables
- `references/concession-antipatterns.md` — Detailed descriptions and detection signals for all three named anti-patterns
- Prerequisite skill: `diagnose-taking-control-gaps` — posture diagnosis that produces the `taking-control-diagnosis.md` this skill consumes

---

## Example Test Scenarios

**Scenario A — Enterprise software deal, CFO discount request:**
> "We have a $2M enterprise deal. The CFO called our AE yesterday and said they need a 25% discount to get board approval. We have discovery notes showing their main concern is implementation risk, not budget. The call is tomorrow."

Expected outputs: SSN worksheet completed with implementation support and extended warranty as high-value/low-cost concessions; Acknowledge and Defer script that buys time; Deepen and Broaden questions probing the board's actual concern (budget vs. implementation risk); trade offer: extended implementation support package in exchange for full contract term; concession plan with 10% / 5% / final offer structure; all three anti-pattern checks clear.

**Scenario B — Manufacturing contract renewal, procurement escalation:**
> "A $500K annual contract renewal. Procurement is demanding a 15% price reduction or they will 'go to market.' They haven't asked for anything else. Our account manager always caves on price. We have solution-capabilities.md listing 8 service elements we can trade."

Expected outputs: SSN scoring of 8 service elements (2–3 will be high-value/low-cost); proactive discounting anti-pattern flagged as risk based on account manager history; Acknowledge and Defer language; Deepen and Broaden questions to determine whether "going to market" is a real threat or an ultimatum; concession sequence starting with service element trades, price last.

**Scenario C — Late-stage deal, customer issued an ultimatum:**
> "Customer sent an email: 'Our final position is 30% off or we're done.' We've already given 10% in the last two rounds. The AE wants to give them 20% more just to close."

Expected outputs: escalating concession anti-pattern flagged (10% already given, AE wants to give 20% — escalating); ultimatum anti-pattern in customer behavior noted (but rep should not mirror it); SSN worksheet check — has the rep exhausted non-price concessions? If not, Deepen and Broaden before any further price concession; if yes, assess whether the deal meets walkaway terms; Acknowledge and Defer script to re-open the conversation rather than respond to the ultimatum directly.

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill depends on:

- `diagnose-taking-control-gaps` — produces the `taking-control-diagnosis.md` this skill consumes as input. Run it first to confirm the deal posture before negotiating.

FILE:references/concession-antipatterns.md
# Concession Anti-Patterns

**Source:** The Challenger Sale, Ch. 7 (DuPont SSN case study). Paraphrased.

These are the three named concession patterns that Challengers are trained to avoid. DuPont teaches reps to recognize them by experiencing them from the customer's side in role-play workshops.

---

## Anti-Pattern 1 — Escalating Concessions

**What it looks like:**
The rep opens with a small concession ("I can give you 5%") and then, when the customer pushes back, offers a larger one ("Okay, 12%"), then larger still ("Fine, 20% if we close today").

**Why it fails:**
- It signals that the rep has not reached their floor — the customer learns to keep pushing
- It trains the customer to associate pressure with getting more
- It destroys the rep's credibility as a confident value partner
- The customer ends the negotiation wondering whether they could have gotten even more

**The correct inverse pattern:**
Open with a meaningful concession to establish good faith, then offer progressively smaller concessions as the negotiation continues. The customer perceives momentum and fairness. The seller preserves margin. Both parties feel they won.

**Detection signal in your own behavior:**
If each new concession is larger than the last, you are in the escalating pattern. Stop. Reference your SSN concession inventory and hold to the next planned trade, not an improvised escalation.

---

## Anti-Pattern 2 — Proactive Discounting

**What it looks like:**
The rep offers a price reduction before the customer has asked for one. Common in situations where the rep fears the customer might ask and wants to preempt the discomfort.

**Why it fails:**
- Explicitly described as passive behavior — the defining characteristic of the Relationship Builder, not the Challenger
- Immediately anchors the conversation at a lower price, destroying negotiating room that did not need to be surrendered
- Signals that the rep does not believe in their own pricing or value delivery
- Teaches customers to wait for unsolicited discounts before engaging seriously

**The correct behavior:**
Do not offer a concession until the customer makes a specific ask. When they do, proceed through the four-step framework (Acknowledge and Defer → Deepen and Broaden → Explore and Compare → Concede According to Plan) rather than jumping directly to a price move.

**Detection signal:**
If you find yourself thinking "I should offer a discount to keep them interested" before the customer has asked for one, this is the proactive discounting impulse. Resist it. Challengers hold position because they are confident in the value they deliver — value the customer has not yet fully articulated.

---

## Anti-Pattern 3 — Ultimatums ("Take It or Leave It")

**What it looks like:**
The rep (or supplier organization) presents a final offer as non-negotiable: "This is our best price. Take it or leave it."

**Why it fails:**
- Described as "risky" — it can collapse a deal that would have closed with one more round of negotiation
- Leaves the customer feeling they were not treated fairly, damaging long-term loyalty even if the deal closes
- Removes the customer's ability to feel they "won" — which is a psychological requirement for most buyers to feel good about the outcome
- Is often interpreted as aggressive rather than assertive, crossing the spectrum line that Challengers avoid

**When it is sometimes justified:**
There are genuine walkaway terms — price floors, non-negotiable scope items, or contractual conditions the supplier cannot breach. These should be identified in Section 5 (Supplier Goals) of the SSN template before the call, not announced as an ultimatum in the heat of negotiation. A planned walkaway is assertive; an ultimatum issued under pressure is aggressive.

**Detection signal:**
If you are announcing a final offer without having worked through Steps 2–3 of the four-step framework, you are using an ultimatum as a shortcut. Go back to Deepen and Broaden.

---

## Summary Table

| Anti-Pattern | Core Failure | Detection Signal | Correct Alternative |
|--------------|-------------|-----------------|---------------------|
| Escalating concessions | Signals no floor; trains more pressure | Each new concession larger than last | Start meaningful, get smaller each round |
| Proactive discounting | Destroys value before demand is made | Offering discount before customer asks | Wait for specific ask; use 4-step framework |
| Ultimatums | Risks deal; leaves customer feeling cheated | Issuing "take it or leave it" under pressure | Use pre-planned walkaway terms from SSN template |

---

## The Underlying Cause

All three anti-patterns share a root cause: the rep's **desire for premature closure**. Ambiguity in negotiation creates psychological discomfort, and the fastest way to end that discomfort is to concede. Challengers are distinguished not by lacking this discomfort, but by having the pre-call preparation and scripted language that allows them to hold position despite it.

The SSN template (`references/ssn-template.md`) is the structural tool that makes this possible for non-Challenger reps.

FILE:references/ssn-template.md
# SSN Pre-Call Planning Template

**Source:** BayGroup International's Situational Sales Negotiation (SSN) methodology, as deployed by DuPont. Paraphrased; SSN Negotiation Planner is © 2009 BayGroup International, Inc.

This template is completed before any negotiation call where a concession request is expected. It is the foundation of Steps 3 and 4 in the DuPont four-step framework.

---

## Section 1 — Power Positions

Map supplier strengths and weaknesses across each dimension:

| Dimension | Supplier Strength (1–10) | Supplier Weakness | Notes |
|-----------|--------------------------|-------------------|-------|
| Products / capabilities | | | |
| Brand / reputation | | | |
| Pricing / value ratio | | | |
| Service / support | | | |
| Relationships / access | | | |

**Why:** Getting supplier strengths on paper before the call builds the rep's confidence to hold firm. Challengers know their leverage; this template makes it explicit for non-Challengers.

---

## Section 2 — Information to Gather from Customer

List the specific questions the rep will ask during the call to understand the customer's real position:

| Question | What it uncovers |
|----------|-----------------|
| | |
| | |
| | |

**Key question for Step 2 (Deepen and Broaden):** "What are you looking to achieve with [X]% price reduction?" — uncovers whether the demand is driven by economic constraint or a specific business outcome that can be addressed differently.

---

## Section 3 — Information to Provide or Protect

| Information customer will ask for | Provide / Protect / Redirect |
|-----------------------------------|------------------------------|
| | |
| | |

**Note:** Decide in advance, not in the moment. Improvising disclosure during negotiation pressure is a primary cause of unnecessary concessions.

---

## Section 4 — Objections and Planned Responses

| Anticipated objection or difficult question | Planned response |
|---------------------------------------------|-----------------|
| | |
| | |
| | |

**Why:** Preparing responses in advance prevents the rep from caving to avoid discomfort. The primary driver of premature closure is not customer power — it is rep discomfort with ambiguity.

---

## Section 5 — Supplier Goals and Customer Hypotheses

**What the supplier is looking for in this deal:**
- (List specific terms, conditions, scope items the supplier wants to preserve or gain)

**Hypotheses about what the customer actually needs (vs. what they demand):**
- (Stated demand: ____; Underlying need hypothesis: ____)
- (Stated demand: ____; Underlying need hypothesis: ____)

---

## Section 6 — Concession Inventory with Scoring

Score each possible concession on two dimensions independently:
- **Cost to supplier** (1 = minimal cost, 10 = very expensive)
- **Value to customer** (1 = low value to them, 10 = very high value to them)

**Concessions to offer:**

| Concession | Cost to Supplier (1–10) | Value to Customer (1–10) | Priority (offer first if low cost, high value) |
|------------|------------------------|--------------------------|------------------------------------------------|
| | | | |
| | | | |
| | | | |

**Concessions to request (in exchange):**

| Concession to ask for | Why it matters to supplier | When to introduce |
|-----------------------|---------------------------|-------------------|
| | | |
| | | |

**Walkaway terms:**
- Price floor: ____
- Non-negotiable scope items: ____
- Conditions under which to walk: ____

---

## How to Use This in the Call

1. Complete all sections before the call.
2. Keep the completed template visible during the negotiation.
3. In Step 2 (Deepen and Broaden): reference Section 2 for questions to ask.
4. In Step 3 (Explore and Compare): reference Section 6 to identify low-cost/high-value concession offers.
5. In Step 4 (Concede According to Plan): follow the priority order in Section 6 — offer the highest-value/lowest-cost concessions first, work toward price last.

**If all non-price concessions are exhausted and customer still demands a price cut:** Reference Section 1 (power positions) and Section 5 (walkaway terms) before responding. The rep who has done this prep is equipped to hold position or make a deliberate, bounded price concession — not an improvised capitulation.

---
name: plan-challenger-model-rollout
description: "Plan a full Challenger model rollout for a sales organization. Use when someone asks: 'implement Challenger model', 'roll out new sales methodology', 'sales transformation plan', 'change management sales', 'pilot Challenger', 'sales methodology adoption', 'implementation roadmap', 'enablement plan', 'sales training rollout', 'sales change management', 'how do I roll out Challenger', 'Challenger implementation plan', 'sales force transformation', 'how to scale Challenger across the team'. Applies Grainger's four-question pilot framework, star/core/laggard adoption sequencing, 80% adoption target, 20–30% attrition planning, and a four-track parallel workstream design (training, tools, coaching, manager enablement). Produces a rollout-plan.md with pilot scope, adoption sequence, 12-month milestone schedule, and attrition/backfill plan."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/plan-challenger-model-rollout
metadata: {"openclaw":{"emoji":"🚀","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors: ["Matthew Dixon", "Brent Adamson"]
    chapters: [9]
tags: [sales, sales-enablement, challenger-sale, implementation, change-management, sales-transformation]
depends-on: [classify-rep-profile, diagnose-manager-effectiveness]
execution:
  tier: 1
  mode: plan-only
  inputs:
    - type: document
      description: "Team composition context: rep-profile distribution from classify-rep-profile + manager effectiveness diagnoses from diagnose-manager-effectiveness + organizational context (team size, current methodology, business unit scope)"
  tools-required: [Read, Write, AskUserQuestion]
  tools-optional: []
  mcps-required: []
  environment: "Any working directory. No codebase required."
discovery:
  goal: "Plan a staged rollout of the Challenger model using Grainger's four-question pilot framework, adoption sequencing, and a parallel-track workstream design"
  tasks:
    - "Gather organizational context: team size, current methodology, rep profile distribution, manager effectiveness baseline"
    - "Run anti-pattern check: verify Challengers were identified by diagnostic, not manager nomination"
    - "Design pilot scope using Grainger's four-question framework"
    - "Sequence adoption waves by star/core/laggard segments"
    - "Set 80% adoption target and 20–30% attrition expectation with backfill plan"
    - "Design four parallel-track workstreams: training, tools, coaching, manager enablement"
    - "Apply training effectiveness triad: pre-training buzz, safe practice, behavioral certification"
    - "Produce rollout-plan.md with 12-month milestones"
  audience:
    roles: ["Sales enablement leaders", "Sales VPs", "Chief Revenue Officers", "Change managers", "Sales operations"]
    experience: senior
  when_to_use:
    triggers:
      - "Sales leader wants to implement or scale the Challenger Selling Model across a team or organization"
      - "Enablement team needs a structured rollout plan after completing rep profile classification"
      - "Organization is designing a sales transformation program and needs a sequenced adoption strategy"
      - "Leadership wants to move beyond ad hoc Challenger training to a systematic implementation"
    prerequisites:
      - "classify-rep-profile run across the rep population to establish baseline profile distribution"
      - "diagnose-manager-effectiveness run across frontline managers to establish coaching baseline"
    not_for:
      - "Hiring plans — Appendix C hiring guide is out of scope; plan backfill headcount but not the hiring process itself"
      - "Individual rep coaching sessions (use coach-rep-with-pause-framework)"
      - "Diagnosing individual rep profiles (use classify-rep-profile)"
      - "Diagnosing individual manager effectiveness (use diagnose-manager-effectiveness)"
  environment:
    codebase_required: false
    codebase_helpful: false
    works_offline: true
  quality:
    scores:
      density: 4
      tier: 1
      mode: plan-only
---

# plan-challenger-model-rollout

Plan a staged implementation of the Challenger Selling Model using Grainger's four-question pilot framework, star/core/laggard adoption sequencing, and a four-track parallel workstream design. Produces a rollout-plan.md artifact.

## When to Use

Use this skill when a sales leader, enablement team, or CRO needs a structured implementation plan for rolling out the Challenger model at organizational scale. This skill synthesizes inputs from `classify-rep-profile` (team profile distribution) and `diagnose-manager-effectiveness` (manager coaching baseline) into an executable rollout plan.

Do not use this skill for individual rep coaching, individual manager diagnosis, or hiring design.

---

## Step 1 — Gather Organizational Context

Ask the following questions if not already provided in the input document:

1. **Team scope:** How many reps and frontline managers are in scope? Single region, BU, or global?
2. **Current methodology:** What is the current sales methodology (if any)? How long has it been in place?
3. **Rep profile baseline:** Has `classify-rep-profile` been run? What is the current distribution across the five profiles (Challenger, Hard Worker, Relationship Builder, Lone Wolf, Reactive Problem Solver)?
4. **Manager baseline:** Has `diagnose-manager-effectiveness` been run? How many managers are strong on Coaching/Innovating drivers vs. weak?
5. **Existing Challenger content:** Does a Commercial Teaching message library exist? Is it marketing-owned or field-generated?
6. **Timeline pressure:** Is there a board or QBR milestone driving timeline? What is the desired 12-month outcome?
7. **Performance tier distribution:** What share of reps are currently in top 20% / middle 60% / bottom 20% by quota attainment?

**Why:** The rollout plan is calibrated to your actual team composition. A team with 35% existing Challengers needs a different pilot scope than a team with 10%. Managers with weak coaching capacity need enablement before they can sustain behavior change in reps.

---

## Step 2 — Anti-Pattern Check: Verify Challenger Identification Method

Before designing any rollout, confirm how Challengers were identified.

**The anti-pattern:** Asking managers to nominate their best Challengers. Managers reliably nominate their high performers regardless of actual selling profile. Roughly 40% of high performers are true Challengers — the remainder are high-performing Lone Wolves, Hard Workers, or Relationship Builders.

**The consequence:** If the rollout is designed to replicate the behaviors of a high-performing Lone Wolf (or Relationship Builder) rather than a true Challenger, the entire model being scaled is wrong.

**The fix:**
- Confirm that `classify-rep-profile` was run using the Appendix B diagnostic instrument (44-attribute behavioral assessment), not manager nomination
- If manager nomination was used, flag this explicitly in the rollout plan and re-classify before proceeding
- Verify that the Challengers being studied have all three subscale strengths: Teach (high), Tailor (high), Take-Control (high)
- Check for "inactive Challengers" — reps who have the profile but haven't activated it; these are high-leverage early intervention targets

Document this finding in the rollout plan's readiness section.

---

## Step 3 — Design the Pilot Using Grainger's Four-Question Framework

Before launching to the full organization, design a pilot. W. W. Grainger pilots not just to test tools, but specifically to understand adoption dynamics before broad launch.

Apply the four questions to your pilot design:

**Q1: How big is the early adopter group?**
Estimate how many reps will self-select into the pilot. This tells you when adoption will naturally plateau without active intervention. Typically: reps with existing Challenger traits + high performers eager for a new edge.

**Q2: Who are the early adopters, and how are they different from non-adopters?**
Profile the early adopter group: profile distribution, quota performance, tenure, region. This tells you where the initial lift will come from and what characteristics predict adoption success — essential for building the majority-wave case.

**Q3: What metrics will predict tool and methodology impact?**
Define leading indicators before the pilot begins. Options: conversation-quality scores, deal velocity, win rate on complex accounts, customer pushback frequency. These give you a signal before quota impact shows up in lagging data.

**Q4: What can we learn to improve tool impact and push majority adoption?**
Build a structured learning cadence into the pilot: weekly debrief, mid-pilot adjustment gate, post-pilot review before majority wave launch.

**Pilot scope recommendation:**
- Size: 10–20% of total rep population, or one region/segment
- Duration: 60–90 days (enough time for behavior change to show in leading indicators)
- Comparison: Include a control group (comparable reps not in pilot) if feasible — enables delta measurement
- Champion managers: Assign your two or three strongest-coaching managers to the pilot cohort

---

## Step 4 — Sequence Adoption by Wave

Map Grainger's adoption segments to your performance tiers:

| Wave | Adoption Segment | Performance Tier | Strategy |
|------|------------------|------------------|----------|
| 1 (Pilot) | Early adopters | Stars + activated/inactive Challengers | Self-selection + direct invitation; champion managers; generate success stories |
| 2 (Majority) | Majority | Core performers (middle 60%) | Show early adopter success stories; proximity matters — use peer stories not star stories |
| 3 (Laggards) | Laggards | Resistant core / lower performers | Use success stories from peers in their own segment; manager-led individual conversations |
| 4 (Non-adopters) | Naysayers | Quota-beaters who refuse | Apply "live by the sword" policy; monitor quota performance; no active forcing |

**Proximity rule (critical):** Do not use star performer success to persuade average performers. People adopt when they see people *like themselves* succeeding. Document average-performer transition stories — reps who moved from non-Challenger to Challenger and improved quota attainment — specifically for majority and laggard wave communication.

**Target:** 80% adoption across the full rep population. Do not target 100%. The final 20% is disproportionately costly to attain. Non-adopters who beat quota are treated as new Lone Wolves: acceptable while above goal, required to adopt or transition if performance slips.

---

## Step 5 — Set Attrition Expectation and Backfill Plan

Expect 20–30% of reps to not complete the transition to the Challenger model. This is not a failure of the program — it reflects genuine profile incompatibility with a teaching-and-control-oriented approach.

**Redeployment options (not termination by default):**
- Customer success or account management roles (relationships, not complex selling)
- Marketing or product specialist roles (deep domain knowledge without quota pressure)
- Internal sales or smaller-account segments with lower complexity requirements

**Backfill plan:**
- Estimate headcount gap: if team is 100 reps, plan for 20–30 eventual departures over 18–24 months
- Note: The Challenger Hiring Guide (Appendix C) is the recommended hiring instrument for replacements, but is out of scope for this skill
- Plan headcount addition to start 6–9 months into the rollout as attrition pattern becomes clear
- Inform HR and recruiting of the incoming volume before the majority wave launches

Document the attrition expectation explicitly in the rollout plan — surprises here create panic; expectation-setting creates confidence.

---

## Step 6 — Design the Four Parallel-Track Workstreams

Do not sequence these tracks. All four must run concurrently, with explicit coordination checkpoints.

**Track 1 — Training:** Build Challenger behaviors (teach, tailor, take control) across the rep population. Use experiential safe practice on real accounts. Source facilitators with frontline sales credibility.

**Track 2 — Tools:** Build or source the Commercial Teaching message library. Pull from existing field Challengers first — they are already delivering insights to customers. Pilot tools with the early adopter wave before broad launch.

**Track 3 — Coaching:** Establish behavioral certification (not attendance certification). Managers must observe and certify rep behavior change. Coaching is the primary lever for training stickiness — 87% of training content is forgotten within 30 days without reinforcement.

**Track 4 — Manager Enablement:** Train managers before they are expected to coach. Diagnose the democratic coaching anti-pattern (managers who spread time equally miss high-leverage opportunities). Assign managers to waves based on their coaching driver strength.

Full workstream detail, milestone templates, and adoption-wave mapping: see [references/parallel-track-workstreams.md](references/parallel-track-workstreams.md).

---

## Step 7 — Apply the Training Effectiveness Triad

Training-only programs fail. Three phases are required:

**Pre-training — Generate demand:**
- Create internal buzz before rollout (not a top-down announcement)
- Share early research findings, compelling data on core performer lift (19% average improvement)
- Identify rep champions who can create peer pull-through demand
- Hold preview sessions framed as "here's what some of your peers are already doing"

**During training — Safe practice on real accounts:**
- Use real accounts, real deals, real customer challenges — not fabricated scenarios
- Require reps to apply Challenger behaviors during training to an actual live opportunity
- Design around experiential learning, not presentation delivery

**Post-training — Behavioral certification:**
- Replace "did you attend?" with "can you demonstrate the behavior?"
- Define certification criteria for each subscale: Teach, Tailor, Take-Control
- Build ongoing manager coaching cadence into the certification cycle
- Track leading behavior indicators (not just lagging quota results)

---

## Step 8 — Write rollout-plan.md

Produce the following artifact:

```markdown
# Challenger Model Rollout Plan
## Organization: [Name]
## Scope: [Teams/BUs in scope]
## Date: [Plan date]

## Readiness Assessment
- Rep profile baseline: [distribution from classify-rep-profile]
- Manager baseline: [distribution from diagnose-manager-effectiveness]
- Challenger identification method: [diagnostic / nomination — flag if nomination]
- Inactive Challengers identified: [count and names]
- Lone Wolf risk: [count of Lone Wolves — do not scale these behaviors]

## Pilot Design (Grainger Framework)
- Q1 — Early adopter group size: [estimate]
- Q2 — Early adopter characteristics vs. non-adopters: [profile]
- Q3 — Leading indicator metrics: [3–5 specific metrics]
- Q4 — Learning capture plan: [weekly debrief structure, adjustment gate timing]
- Pilot cohort: [rep count, region/segment, assigned managers]
- Control group: [yes/no, description]
- Pilot duration: [60 / 90 days]

## Adoption Sequence
- Wave 1 Pilot: [dates, cohort, champion managers]
- Wave 2 Majority: [dates, cohort, peer success stories from Wave 1]
- Wave 3 Laggards: [dates, cohort, segment-proximity stories]
- Wave 4 Non-adopters: [monitoring policy — "live by the sword"]

## Adoption Targets
- 80% adoption target by: [date]
- Non-adopter policy: [quota-beating = tolerated; performance slip = adopt or transition]

## Attrition Plan
- Expected attrition: 20–30% of [N] reps = [low estimate]–[high estimate] reps
- Timeline: over [18–24 months]
- Redeployment options: [customer success / marketing specialist / smaller accounts]
- Backfill start: [6–9 months into rollout]
- HR notification: [date]

## Four-Track Workstream Timeline (12 months)
| Month | Training | Tools | Coaching | Manager Enablement |
|-------|----------|-------|----------|--------------------|
| 1–2   | Demand generation; facilitator selection | Challenger ID and field message inventory | Manager diagnostic complete | Manager training begins |
| 3–4   | Pilot cohort trained | Pilot tool set v1 ready | Certification rubric defined | Champion managers briefed |
| 5–6   | Pilot debrief; majority wave prep | Tool improvements from pilot | Pilot cohort certification cycle | All managers trained |
| 7–8   | Majority wave trained | Improved tool set deployed | Majority wave certification | Coaching cadence running |
| 9–10  | Laggard wave prep | Full tool library | Laggard wave support | Anti-pattern corrections |
| 11–12 | Laggard wave trained | Tool adoption at 80% | 80% certified | Manager performance reviews updated |

## 12-Month Milestones
- Month 2: Readiness audit complete; pilot cohort defined
- Month 4: Pilot wave trained and certified; leading metrics baselined
- Month 6: Pilot post-mortem; majority wave launch decision
- Month 8: Majority wave certified; adoption plateau analysis
- Month 10: Laggard wave active; attrition pattern clear; backfill in progress
- Month 12: 80% adoption achieved; non-adopter policy enforced; program review
```

---

## Self-Check Before Delivering

Verify the rollout plan includes:

- [ ] Challenger identification was diagnostic-based, not manager nomination
- [ ] Pilot uses all four Grainger questions with specific answers
- [ ] Adoption waves are sequenced early adopters → majority → laggards → naysayers
- [ ] 80% adoption target is explicit (not 100%)
- [ ] 20–30% attrition expectation is documented with redeployment options and backfill timeline
- [ ] All four workstreams run in parallel (no sequential ordering)
- [ ] Training effectiveness triad is addressed: pre-training buzz, safe practice, behavioral certification
- [ ] Proximity rule applied: peer success stories, not star stories, for majority and laggard waves
- [ ] Lone Wolf warning is addressed: verify the model being scaled is Challenger-profile behaviors, not Lone Wolf behaviors

---

## References

- [Parallel-Track Workstream Detail + Training Effectiveness Triad](references/parallel-track-workstreams.md)
- Prerequisite: `classify-rep-profile` — produces team rep-profile distribution required for Step 1
- Prerequisite: `diagnose-manager-effectiveness` — produces manager coaching baseline required for Track 4
- Related: `coach-rep-with-pause-framework` — for individual rep coaching sessions during the rollout

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill depends on:

- `classify-rep-profile` — produces the rep-profile distribution this skill consumes
- `diagnose-manager-effectiveness` — produces manager diagnoses this skill uses to plan the manager-enablement workstream

Related skill for individual coaching during the rollout: `coach-rep-with-pause-framework`.

FILE:references/parallel-track-workstreams.md
# Parallel-Track Workstreams Reference

**Skill:** plan-challenger-model-rollout  
**Source:** The Challenger Sale, Ch 9 — Implementation Lessons from the Early Adopters

---

## Why Parallel, Not Sequential

Organizations that invest in rep skills without building organizational capability leave reps feeling they lack the tools to execute the model. Organizations that build Commercial Teaching messages and tools without investing in rep skills find that reps reject the new content and revert to familiar approaches.

The only viable path: develop all four workstreams concurrently, with explicit handoffs and milestones between them.

---

## The Four Workstreams

### Track 1: Training

**Goal:** Build foundational Challenger behaviors — teach, tailor, take control — across the rep population.

**Key design principles:**
- Design for experiential "safe practice," not lecture delivery
- Use real accounts in classroom exercises, not fictional scenarios
- Source facilitators with frontline sales credibility (former sales leaders, not L&D generalists)
- Sequence for the adoption wave: train early adopters first, then use their success stories to pull the majority

**Milestones:**
- M1 (pre-launch): Rep demand generation and social buzz campaign complete
- M2 (pilot): First cohort certified on Teach/Tailor/Take-Control subscales
- M3 (majority wave): Second-wave cohort trained using early adopter success stories
- M4 (laggard wave): Peer-proximity stories documented and deployed

---

### Track 2: Tools

**Goal:** Equip reps with organizational Commercial Teaching messages and supporting collateral so they have something to deliver once trained.

**Key design principles:**
- Tools must be built *concurrently* with rep training — not afterward
- Pull insights from existing high-performing Challengers in the field (they are already teaching customers right now)
- Pilot tools with the early adopter group before broad launch using the Grainger 4-question framework
- Measure tool adoption as a leading indicator of behavior change

**Milestones:**
- M1: Challenger identification complete; existing field teaching messages inventoried
- M2: Pilot-tested Commercial Teaching message library v1 ready for early adopter wave
- M3: Adoption plateau analysis complete; tool improvements deployed for majority wave
- M4: Full tool library available to 80% of reps

---

### Track 3: Coaching

**Goal:** Sustain training gains through ongoing behavioral certification and manager-led reinforcement.

**Why training alone fails:** Neil Rackham's research shows 87% of training content is forgotten within 30 days without reinforcement. Coaching is the primary lever for training stickiness.

**Key design principles:**
- Distinguish coaching (known behavior development) from innovating (deal-level obstacle removal) — see `diagnose-manager-effectiveness`
- Managers must be trained on the Pause Framework before they can coach Challenger behaviors (see `coach-rep-with-pause-framework`)
- Behavioral certification replaces "did you complete training" assessments: reps must demonstrate the new behaviors, not just attend

**Three-phase certification process:**
1. **Pre-training:** Manager baseline assessment of rep Teach/Tailor/Take-Control subscale scores
2. **During training:** Observed "safe practice" with real accounts; manager observes one live call per rep per month
3. **Post-training:** Structured behavioral certification criteria; ongoing coaching cadence with documented skill improvement

**Milestones:**
- M1: Manager cohort diagnosed (see `diagnose-manager-effectiveness`); coaching gap identified
- M2: Manager coaching training complete; certification rubric defined
- M3: Behavioral certification cycle running for early adopter wave
- M4: Full certification coverage at 80% rep population

---

### Track 4: Manager Enablement

**Goal:** Build manager capability to coach Challenger behaviors and allocate coaching time to the highest-leverage reps.

**Why this is a separate track:** Managers are the primary implementation lever for large-scale rep behavior change. But managers have their own adoption curve. Underdeveloped managers will undermine training investments regardless of content quality.

**Key design principles:**
- Prioritize managers who exhibit strong Coaching and Innovating driver scores (see `diagnose-manager-effectiveness`)
- Target democratic coaching anti-pattern: managers who spread coaching time evenly miss the highest-leverage opportunities
- Identify manager champions early — high-performing managers who are themselves Challengers or Challenger-adjacent
- Proximity principle applies to managers too: struggling managers need to see peer managers succeed, not just star managers

**Manager coaching time allocation targets (from CEB research):**
| Driver | Target Weight |
|--------|--------------|
| Innovating (deal-level problem-solving) | 29% |
| Coaching (behavior development) | 28% |
| Selling (joint calls, deal strategy) | 27% |
| Resource Allocation (organizational navigation) | 16% |

**Milestones:**
- M1: All managers diagnosed via `diagnose-manager-effectiveness`
- M2: Manager champions identified and briefed; coaching guide distributed
- M3: Manager coaching cadence established; democratic coaching anti-pattern addressed
- M4: Manager performance reviews include coaching effectiveness metrics

---

## Training Effectiveness Triad

This triad applies across all four workstreams and is most directly relevant to Tracks 1 and 3:

| Phase | Activity | Purpose |
|-------|----------|---------|
| **Pre-training** | Social demand generation, peer testimonials, training preview sessions | Create pull demand (avoid top-down mandate perception) |
| **During training** | Experiential learning on real accounts; safe practice; live facilitators | Build skill with realistic friction |
| **Post-training** | Behavioral certification; ongoing coaching cadence; documented results | Sustain change; close the 87% forgetting gap |

The single most common failure mode in Challenger rollouts is investing heavily in classroom training while neglecting pre-training buzz and post-training certification. Both ends of the triad are required.

---

## Adoption Sequencing Within Workstreams

Each of the four workstreams should be phased to match the rep adoption sequence:

| Wave | Rep Segment | Manager Role | Tools Status | Coaching Phase |
|------|-------------|--------------|--------------|----------------|
| Wave 1 (Pilot) | Early adopters (self-select, often stars + activated Challengers) | Champion managers | Pilot tool set | Baseline certification |
| Wave 2 (Majority) | Core performers — middle 60% | Trained coaches | Improved tool set (post-pilot) | Structured reinforcement |
| Wave 3 (Laggards) | Resistant core + lower performers | All managers | Full tool library | Ongoing cadence |
| Wave 4 (Non-adopters) | Quota-beating resistors (new Lone Wolves) | Monitor only | N/A | "Live by the sword" policy |

---
name: diagnose-taking-control-gaps
description: |
  Diagnose a deal's taking-control posture: foil-RFP detection, passive/assertive/aggressive positioning, and misconception analysis.

  Trigger this skill when you need to:
  - Diagnose a stalled deal to find out why it isn't moving
  - Determine whether you're in a foil RFP (a verification exercise, not a real opportunity)
  - Assess whether a rep is too passive, appropriately assertive, or crossing into aggressive territory
  - Identify misconceptions about "taking control" that are limiting a rep's deal behavior
  - Understand why you're losing control of the sale or the customer conversation
  - Apply constructive tension without becoming combative
  - Evaluate whether a customer is verifying price with a competitor already chosen
  - Unblock a stalled pipeline deal by diagnosing the control gap
  - Coach a rep who is too passive or who conflates assertiveness with aggressiveness
  - Determine whether a rep is helping the customer navigate their own buying process (Lead and Simplify)

  NOT for: full negotiation planning — concession sequencing, the DuPont four-step negotiation roadmap, SSN pre-call templates (use plan-negotiation-with-constructive-tension for those).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/diagnose-taking-control-gaps
metadata:
  openclaw:
    emoji: "🎯"
    homepage: "https://github.com/bookforge-ai/bookforge-skills"
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors:
      - Matthew Dixon
      - Brent Adamson
    chapters:
      - 7
tags:
  - sales
  - b2b-sales
  - challenger-sale
  - deal-diagnosis
  - sales-control
  - negotiation-adjacent
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "deal-brief.md (required) — account, deal stage, stakeholder access granted so far, RFP context; call-transcript.md or discovery-notes.md (optional) — last meeting notes to diagnose control behaviors"
  tools-required:
    - Read
    - Write
    - AskUserQuestion
  works-offline: true
discovery:
  goal: "Diagnose a deal's taking-control posture: detect foil RFPs, classify rep behavior on the passive/assertive/aggressive spectrum, identify active misconceptions, and produce a taking-control-diagnosis.md with concrete next-action recommendations."
  triggers:
    - "diagnose stalled deal"
    - "am I in a foil RFP"
    - "losing control of the sale"
    - "passive assertive aggressive sales"
    - "take control of customer conversation"
    - "constructive tension"
    - "customer is verifying price"
    - "stalled pipeline deal"
    - "rep is too passive"
    - "why isn't my deal moving"
  audience:
    - B2B sales representatives diagnosing a live deal
    - Sales managers reviewing a rep's deal behavior
    - Sales enablement teams coaching assertiveness skills
  not_for:
    - Full negotiation planning → use plan-negotiation-with-constructive-tension
    - Building a commercial teaching pitch → use build-commercial-insight
    - Classifying rep profile type → use classify-rep-profile
---

# Diagnose Taking-Control Gaps

## Overview

Taking control of the sale is one of the three core Challenger behaviors — alongside teaching and tailoring. It is also one of the most misunderstood. Most reps either avoid it (passive posture) or confuse it with aggression. This skill diagnoses *where* a deal or rep behavior has a control gap and *why*.

The diagnosis covers four areas:
1. **Foil RFP detection** — is this opportunity a genuine purchase or a verification exercise?
2. **Behavioral spectrum classification** — passive, assertive, or aggressive?
3. **Misconception scan** — which of the three common control misconceptions are shaping behavior?
4. **Lead and Simplify assessment** — is the rep helping the customer navigate their own buying process, or passively following the customer's lead?

Output is a `taking-control-diagnosis.md` file with verdicts, evidence, and recommended next actions.

---

## When to Use

Use this skill when:
- A deal has stalled and you cannot identify why
- You received an inbound RFP and are unsure whether it's a real opportunity
- A rep's transcript or meeting notes show consistent pattern of deferring, over-agreeing, or proactively conceding
- A manager wants to coach a rep on taking-control behaviors but needs evidence from the deal record
- You want to verify whether a rep is leading the customer's buying process or following it

---

## Step 1 — Load Deal Artifacts

Read all available deal documents in the working directory.

**Required:**
- `deal-brief.md` — account overview, current deal stage, which stakeholders have been engaged, whether an RFP is present, deal size, budget confirmation status

**Optional but valuable:**
- `call-transcript.md` or `discovery-notes.md` — actual rep behavior evidence: language used, whether next steps were defined, whether pushback was maintained or dropped
- `stakeholder-map.md` — who has been granted access, what level (junior/senior), whether decision-makers have been met

If `deal-brief.md` is absent, ask the user to provide a description of the deal covering:
- Deal stage and age
- How the opportunity originated (inbound RFP, rep-generated, referral)
- Which contacts have been engaged and at what level (junior procurement, senior executive)
- Whether access to senior decision-makers has been requested and what the response was
- Key behaviors from the last 1–2 customer interactions

Why this matters: the diagnosis requires both deal context (foil detection) and behavioral evidence (spectrum classification). The more specific the evidence, the more actionable the output.

---

## Step 2 — Run the Foil RFP Detection

Roughly 20% of all sales opportunities are foil RFPs: situations where the customer has already selected a vendor but runs a formal RFP process as a verification exercise. These deals have confirmed budget, a ready contact, and real meetings — which is exactly why most reps love them. But the customer has no real intention of changing vendors.

**Apply the four foil signals:**

| Signal | Foil indicator present? |
|--------|------------------------|
| Opportunity originated as an inbound RFP or unsolicited request | Possible foil — confirm with signals below |
| Rep has only met with junior procurement or administrative contact | Strong foil signal |
| Requests for access to senior decision-makers have been vague, delayed, or denied | Strong foil signal |
| Customer sets unusually tight RFP timelines that compress meaningful evaluation | Supporting signal |

**Apply the access test:**

If the rep has not yet explicitly tested for access, flag this as a required immediate action. The access test works as follows:

At the close of the first substantive interaction, the rep should directly ask:
> "For this type of decision, our experience is that the right senior leaders need to be involved. Is that the case in your organization? When would I be able to meet with them?"

**Interpret the response:**

- Customer agrees and provides a specific timeline → not a strong foil signal; continue and revisit
- Customer hedges, says it's too early, or defers indefinitely → strong foil signal
- Customer explicitly denies access to senior leaders → foil confirmed; recommend disengage

**Foil verdict:**

Assign one of three verdicts. **In every diagnosis, state the base rate explicitly: roughly 20% of all inbound B2B sales opportunities are foil RFPs** — verification exercises where the customer has already selected a vendor but runs a formal process to validate price or satisfy procurement. This 20% base rate anchors the rep's prior before they read signals.

- **Foil — disengage:** Access denied or strongly hedged; multiple foil signals present. The rep's time is better spent on real opportunities. Recommend cutting the sales effort and redirecting resources.
- **Foil risk — access test required:** RFP is inbound and only junior contact engaged; access has not been explicitly tested. Rep must run the access test at the next interaction before investing further.
- **Not a foil:** Senior stakeholder access has been granted or a credible path exists; deal shows genuine evaluation intent.

**Required in every output:** the final `taking-control-diagnosis.md` MUST cite the ~20% foil base rate in the Foil Verdict section so the rep can calibrate against it.

---

## Step 3 — Classify Rep Behavior on the Spectrum

Using the call transcript, discovery notes, or rep's own description of recent interactions, classify observable behavior against the passive/assertive/aggressive spectrum.

See `references/control-rubric.md` for the full diagnostic checklist. Key signals:

**Passive signals:**
- Rep proactively offered a discount, extended timeline, or expanded scope without being asked
- Rep did not push for a defined next step after a positive meeting ("let's be in touch")
- Rep accepted the customer's stated next-step timeline without proposing an alternative
- Rep immediately backed down when the customer pushed back on the reframe or value case
- Rep asked "Who needs to be involved?" and waited for the customer to answer, rather than prescribing who should be

**Assertive signals (target zone):**
- Rep maintained the value position across multiple pushback rounds, citing specific evidence
- Rep defined the next step at the close of the meeting and secured agreement
- Rep pushed back on a customer assumption with a specific counter-argument, then invited the customer to explore further
- Rep proactively coached the customer on who should be involved and what steps to take next
- Rep demanded senior stakeholder access as a condition of continued investment of time and resources

**Aggressive signals (off-target):**
- Rep used language the customer described as threatening or antagonistic
- Rep continued pushing a position without reading or responding to the customer's signals
- Rep escalated in a way that damaged a key relationship without a clear value case

**Power perception check:**
A key reason reps go passive is a false belief that the customer holds all the power. Industry data consistently shows both sides perceive the other as more powerful. If the rep believes they have no leverage, they are most likely wrong — flag this in the diagnosis.

**Spectrum verdict:** Passive / Borderline passive / Assertive / Borderline aggressive / Aggressive

---

## Step 4 — Misconception Scan

Check whether any of the three named control misconceptions are shaping the rep's or team's behavior. These are not abstract — each produces specific, diagnosable behavior patterns.

**Misconception 1 — Taking control is only about negotiation (end-of-sale)**

*Diagnostic:* Rep applies pressure or holds firm only during formal pricing discussions. Earlier stages of the deal are handled passively. Rep has not run the foil access test. Rep has not maintained the reframe under pushback during discovery or pitch stages.

*Correction:* Taking control spans the entire sale. The foil access test at the first meeting is the clearest early-stage example. Controlling the buying process (Lead and Simplify) happens mid-sale. Negotiation is a subset, not the whole.

**Misconception 2 — Taking control is only about money**

*Diagnostic:* Rep holds firm on price but does not push for next steps, does not maintain the commercial teaching reframe under customer pushback, and defers on the question of who to involve and at what pace.

*Correction:* Control also means controlling how the customer thinks (maintaining the reframe), controlling deal momentum (next steps at every meeting), and controlling the buying process (prescribing steps rather than following the customer).

**Misconception 3 — Assertive equals aggressive**

*Diagnostic:* Rep (or manager) avoids any form of direct challenge or push because they fear it will damage the relationship. Any pushback is treated as "too aggressive." Alternatively, the diagnosis here can be the opposite — rep is actually aggressive (see Step 3) and using "assertive" as cover.

*Correction:* Assertive uses strong, purposeful language while sensing and responding to the customer. Aggressive pursues the rep's agenda without reading the customer's reaction. The more common failure is excessive passivity — reps who were told to take control almost never jump to aggressive; they stay passive.

Mark each misconception as: **Active / Not present / Unclear (more evidence needed)**

---

## Step 5 — Lead and Simplify Assessment

Evaluate whether the rep is helping the customer navigate their own buying process, or deferring to the customer's lead.

**Learn and React (passive — red flags):**
- Rep asked the customer "Who else needs to be involved?" and accepted whatever answer was given
- Rep followed the customer's stated timeline without proposing a more efficient path
- Rep mapped stakeholders by name and role but did not investigate what each stakeholder cares about or why
- Rep deferred to the customer's stated purchasing process even when it was clearly suboptimal

**Lead and Simplify (assertive — target behaviors):**
- Rep drew on past successful implementations to prescribe which stakeholders should be involved and why
- Rep proposed a specific next step with a date and named participants, then secured agreement
- Rep investigated stakeholder goals, biases, and personal objectives — not just job titles
- Rep simplified the customer's complex buying process by framing it in terms the customer had not articulated themselves

**Assessment verdict:** Learning and Reacting / Mixed / Leading and Simplifying

---

## Step 6 — Produce taking-control-diagnosis.md

Write `taking-control-diagnosis.md` to the working directory with the following structure:

```
# Taking-Control Diagnosis — [Deal Name / Rep Name]
Generated: [date]

## Foil RFP Verdict
[Verdict + key evidence + recommended action]

## Behavioral Spectrum Position
[Verdict + 3–5 specific behavioral observations from the deal record]

## Active Misconceptions
[List each misconception with status: Active / Not present]
[For each Active misconception: one concrete corrective action]

## Lead and Simplify Assessment
[Verdict + 2–3 specific examples from the deal record]

## Priority Next Actions
[Ordered list of 3–5 specific actions the rep should take in the next interaction]

## If Foil Confirmed — Disengage Protocol
[If foil verdict = Disengage: specific script for exiting gracefully and redirecting time]
```

**Important:** If the foil verdict is "disengage," make this the first and most prominent recommendation. Continuing to invest in a foil opportunity is the most expensive mistake in this framework — every hour spent on a foil is an hour not spent on a real opportunity.

---

## Reference Materials

- `references/control-rubric.md` — Full passive/assertive/aggressive diagnostic checklist, misconception detail, Lead and Simplify signal table, power perception gap data
- Negotiation planning (DuPont roadmap, SSN template) → `plan-negotiation-with-constructive-tension`

---

## Self-Check Before Delivering Diagnosis

Before writing the final `taking-control-diagnosis.md`, verify:

- [ ] Foil verdict is one of three defined options with evidence cited
- [ ] Spectrum verdict is based on at least 3 specific behavioral observations, not general impressions
- [ ] Each active misconception has a concrete corrective action, not just a label
- [ ] Lead and Simplify assessment is based on observable behavior from the deal record
- [ ] If foil = disengage, the first recommendation is to exit — not to try harder
- [ ] Priority next actions are specific (who does what in which interaction) — not generic advice
- [ ] No verbatim book passages appear in the output artifact

---

## Example Output Prompts (Skill-Tester Scenarios)

**Scenario A — Inbound RFP, junior contact only:**
> "I have a deal: enterprise logistics company sent us an RFP 3 weeks ago. Only contact is a procurement coordinator. We've had two calls. Deal size is $400K. They've told us 'the right people are involved' but haven't offered a meeting with anyone above director level. Decision is supposedly in 6 weeks."

Expected outputs: Foil risk verdict, access test language, misconception 1 flag, disengage-if-denied recommendation.

**Scenario B — Rep transcript with passive behavior pattern:**
> "Here is a call transcript. The rep gave a 20% discount offer on the first call without being asked. When the customer said 'our CFO isn't involved at this stage,' the rep said 'no problem, we can work with you.' Meeting ended with 'I'll be in touch next week.'"

Expected outputs: Passive spectrum verdict, misconceptions 2 and 3 flagged, Lead and Simplify deficit identified, three corrective next-step behaviors.

**Scenario C — Rep accused of being "too aggressive":**
> "My manager says I'm being too pushy with this prospect. I pushed back twice on their request for a 30% price cut, asked to meet the CFO directly, and told them if they can't give us access to their operations team we can't do a proper scoping. Now the customer says I'm hard to work with."

Expected outputs: Assertive-not-aggressive verdict, misconception 3 analysis (is the rep actually assertive or crossing into aggressive?), behavioral evidence review, guidance on sensing-and-responding to customer reaction signals.

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill is standalone (no dependencies). After running this diagnosis, the natural next step is to invoke `plan-negotiation-with-constructive-tension`, which consumes the `taking-control-diagnosis.md` artifact this skill produces.

FILE:references/control-rubric.md
# Taking-Control Rubric

Reference for: **diagnose-taking-control-gaps**  
Source: The Challenger Sale, Ch. 7 — Taking Control of the Sale (Dixon & Adamson)

---

## 1. Three Misconceptions About Taking Control

These are the three named misconceptions from the book, with their corrections.

### Misconception 1 — Taking control is synonymous with negotiation

**The misconception:** Taking control only happens at the end of the sale, at the negotiating table. It's a deal-closing skill.

**Why it's wrong:** Challengers take control across the entirety of the sales process — including at the very beginning. The most concrete early-stage example is the foil RFP access test (see Section 2 below). Negotiation is a subset of taking control, not the definition of it.

**Diagnostic signal:** Rep only pushes back or exerts pressure during formal pricing negotiations. Earlier stages of the deal are handled passively — the rep accepts whatever access and pace the customer sets.

---

### Misconception 2 — Taking control only applies to money

**The misconception:** Control behavior is about resisting discounts and protecting pricing. It's a financial posture.

**Why it's wrong:** Challengers also take control of:
- How the customer thinks (the reframe in Commercial Teaching — pushing back on customer assumptions with constructive tension)
- Deal momentum (pushing for concrete next steps at the close of every meeting, even when the meeting ended positively)
- The buying process itself (Lead and Simplify — prescribing who to involve and how to proceed, rather than asking the customer to lead)

**Diagnostic signal:** Rep holds firm on price but does not push for next steps after positive meetings, does not push back on customer assumptions, and defers to the customer on who should be involved in the decision.

---

### Misconception 3 — Assertive means aggressive

**The misconception:** Telling reps to "take control" or "be more assertive" will turn them into antagonistic, relationship-damaging aggressors.

**Why it's wrong:** Assertive and aggressive are distinct positions on a behavioral spectrum. The real risk runs in the opposite direction: most reps, when told to be more assertive, will still gravitate toward passive. The concern about reps becoming too aggressive almost never materializes.

**Diagnostic signal:** Rep (or manager) interprets any form of pushback or pressure as "too aggressive." The rep acquiesces at the first sign of customer resistance. Alternatively, rep uses genuinely antagonistic language and does not sense/respond to customer reaction — that is aggressive, not assertive.

---

## 2. Passive / Assertive / Aggressive Spectrum

Source: Figure 7.1, CEB Sales Leadership Council, 2011.

```
PASSIVE ←————————————————→ ASSERTIVE ←——————————————→ AGGRESSIVE

Gives in to demands          Holds value position         Attacks with
Uses accommodating language  Uses strong but              antagonistic language
Allows personal boundaries   constructive language        Blindly pursues
to be breached               Senses and responds          own agenda
Proactively discounts        to customer reaction         Off-putting and
without being asked          Moves purposefully           offensive
Primary goal: please         Maintains constructive
the customer                 tension respectfully
                             Primary goal: sell a deal,
                             not just have a good meeting
```

### Rep Behavior Diagnostic Checklist

| Observable behavior | Spectrum position |
|---------------------|------------------|
| Proactively discounts before asked | Passive |
| Avoids pressing for next steps after a positive meeting | Passive |
| Immediately caves when customer pushes back on the reframe | Passive |
| Asks "Who needs to be involved?" and waits for customer to lead | Passive |
| Accepts limited stakeholder access without challenging it | Passive |
| Holds pricing position across multiple pushback rounds while citing value | Assertive |
| Pushes for concrete next steps at close of every meeting | Assertive |
| Maintains reframe under pushback: "You're right this is different, AND..." | Assertive |
| Proactively coaches customer on who should be involved in the decision | Assertive |
| Demands senior stakeholder access as a condition of continued engagement | Assertive |
| Uses threatening or belittling language with customer contacts | Aggressive |
| Continues own agenda without reading customer's signals | Aggressive |
| Escalates in a way that burns the relationship without a value case | Aggressive |

---

## 3. Power Perception Gap

BayGroup International survey finding:
- 75% of sales reps believe procurement has more power in the negotiation
- 75% of procurement officers believe the seller has more power

Implication: most reps are operating from a fundamentally false assumption about the power balance. When reps concede because they believe they have no leverage, the concession is not driven by reality — it is driven by misperception. Challengers understand this intuitively and do not give in prematurely.

---

## 4. Lead and Simplify — Signals

Challengers "lead and simplify" rather than "learn and react."

| Learn and React (passive) | Lead and Simplify (assertive) |
|--------------------------|------------------------------|
| "Who needs to be involved in this decision?" | Prescribes who should be involved based on past successful implementations |
| "What are your next steps on your end?" | "Here is what I recommend as our next step: X by [date]." |
| Defers to customer's pace and process | Teaches the customer how to buy a complex solution |
| Maps stakeholders by name and title | Maps stakeholder goals, biases, and personal objectives |
| Accepts the customer's stated buying process | Simplifies the process by drawing on supplier's implementation knowledge |

---

## 5. Foil RFP — See Skill Body

Full foil RFP detection protocol and access test language are documented in the SKILL.md body (Steps 2–3). This reference provides the rubric; the skill body provides the decision logic and output format.

---
name: diagnose-pitch-for-commercial-teaching-fit
description: Audit an existing sales pitch, deck, or call transcript against the Commercial Teaching rubric. Use this skill when you want to review your deck, diagnose why your pitch isn't working, check whether your pitch leads with solution instead of leading to it, run a commercial teaching check, get a pitch diagnostic, run a sales deck review, figure out why your pitch isn't differentiating, or check whether your pitch opens with your solution before establishing a customer problem. Detects lead-with-vs-lead-to errors, missing Reframes, Rational Drowning misfocus, teaching-into-the-desert traps, buzzword pollution, and sequence violations. Produces a scored per-step rubric with highlighted problem passages and rewrite recommendations.
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/diagnose-pitch-for-commercial-teaching-fit
metadata: {"openclaw":{"emoji":"🔍","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors: ["Matthew Dixon", "Brent Adamson"]
    chapters: [4, 5]
tags: [sales, b2b-sales, commercial-teaching, pitch-audit, sales-enablement, challenger-sale]
depends-on: []
execution:
  tier: 1
  mode: full
  inputs:
    - type: document
      description: "An existing pitch artifact: markdown export of a slide deck, a call transcript, a sales narrative document, or a written pitch script."
  tools-required: [Read, Write, Grep]
  tools-optional: []
  mcps-required: []
  environment: "Works offline from a provided pitch artifact. No live system access needed. Output: pitch-diagnosis.md with scored rubric, flagged passages, and rewrite recommendations."
discovery:
  goal: "Audit an existing pitch against the Commercial Teaching rubric to surface structural errors, anti-patterns, and missing elements — with specific rewrite guidance."
  tasks:
    - "Map each pitch section or slide to one of the six choreography steps"
    - "Score each step as Pass, Partial, or Fail against its specific criteria"
    - "Detect lead-with-vs-lead-to error: does the pitch open with the seller's solution before establishing customer problem?"
    - "Detect teaching-into-the-desert: does the insight anchor back to the seller's unique capabilities?"
    - "Detect missing or weak Reframe: does the pitch produce curiosity and surprise, or agreement and confirmation?"
    - "Check Rational Drowning focus: is the ROI on the customer's problem or on purchasing the solution?"
    - "Scan for buzzword pollution: leading, unique, innovative, solution, best, top, largest"
    - "Produce pitch-diagnosis.md with scored rubric, specific flagged passages, and rewrite recommendations"
  audience: "sales reps, sales enablement teams, sales operations, revenue leaders, marketing teams building sales collateral"
  when_to_use: "When a pitch exists and you want to improve it before a key meeting, after a lost deal, or as part of a systematic sales content audit"
  environment: "Pitch artifact as a document or text. Works from a deck export, transcript, or narrative script."
  not_for:
    - "Authoring a new pitch from scratch — use author-commercial-teaching-pitch instead"
    - "Evaluating pricing or negotiation tactics — those require different skills"
  quality: placeholder
---

# Diagnose Pitch for Commercial Teaching Fit

## When to Use

You have an existing sales pitch — a deck, a script, a call transcript, or a narrative document — and you want to know whether it is built to teach or built to sell. This skill applies when:

- You lost a deal and suspect the pitch led with your solution before establishing why the customer should care
- Your pitch feels like everyone else's, and you cannot explain why you are not differentiating
- Marketing delivered a deck and you are not sure it follows the right sequence
- You want to run a systematic audit of your sales collateral before a critical meeting
- Someone reviewed your pitch and said "you're just talking about yourself" — and you want to know exactly where

**The diagnostic framework:** Commercial Teaching pitches follow a six-step sequence — Warmer → Reframe → Rational Drowning → Emotional Impact → New Way → Your Solution. The seller's product does not enter the conversation until Step 6. Most pitches violate this by leading with their solution instead of leading to it. This audit locates the violations precisely.

**This skill does not author a new pitch.** It diagnoses an existing one. If you need to build a pitch from scratch, use `author-commercial-teaching-pitch`.

Before starting, confirm you have:
- A pitch artifact (deck export, transcript, or written narrative) to audit
- Enough content to identify sections or slides individually

---

## Context and Input Gathering

### Required Input

- **Pitch artifact:** A document that represents your actual pitch. Acceptable formats: markdown export of a slide deck, call transcript, written sales narrative, or capability brochure. If the pitch only exists as a live presentation, capture a written outline with slide titles and key talking points per slide.

### Observable Context

Before scoring, scan the artifact for these quick signals:

- **First 20% of content:** Does it mention your company name, product name, or offering? If yes, lead-with error is likely.
- **Presence of "About Us" or capability section early:** Almost always a lead-with indicator.
- **Buzzword count:** Search for: leading, unique, innovative, solution, best, top, largest, innovator. More than 3 occurrences in a short pitch is a pollution signal.
- **ROI or data section:** Is the ROI calculated on the customer's problem, or on buying your product?
- **Storytelling section:** Is there a moment where you describe how other companies experienced this problem? If not, Step 4 is likely missing.

### Default Assumptions

- If the pitch is a slide deck without speaker notes → audit the slide titles and visible content only; note that full scoring requires speaker notes
- If the pitch is a call transcript → treat each topic shift as a potential step boundary
- If sections are unlabeled → assign provisional step labels based on content, then flag uncertainty in the output
- If the pitch is a single paragraph → still apply the sequence test; the error pattern will be obvious

### Sufficiency Check

You have enough to proceed when:
1. You have a pitch artifact with enough content to identify what happens in the opening, middle, and close
2. You can tell what the seller is selling (at least at a category level)
3. You can tell who the customer is (at least at a role or industry level)

If the pitch is fewer than 5 sentences, ask the user to expand it or provide the full deck before continuing.

---

## Execution

### Step 1 — Load and Structure the Pitch Artifact

Read the pitch artifact in full. As you read, do not score yet — just parse.

Create a working map of the artifact's sections. For a slide deck: list each slide title. For a transcript: identify each topic shift. For a narrative: identify paragraph clusters by theme.

For each section, write a one-line summary of what that section is claiming or doing.

**Why this matters:** You cannot score choreography violations without first knowing what sequence the pitch actually follows. Most reps are not aware their pitch is out of sequence — they need the map made explicit.

---

### Step 2 — Map Sections to Choreography Steps

For each section in your working map, assign it to one of the six choreography steps, or mark it as Unclassified if it does not fit any step.

The six steps and their content signatures:

| Step | What the Content Does |
|------|-----------------------|
| 1 — Warmer | Discusses customer challenges, benchmarks, peer anecdotes. No product. Ends with a question. |
| 2 — Reframe | Introduces unexpected angle that connects challenge to a hidden problem or overlooked opportunity. Delivers surprise. |
| 3 — Rational Drowning | Quantifies the cost or size of the problem using data, charts, ROI models. |
| 4 — Emotional Impact | Tells a story about how similar companies experienced this problem. Makes it personal and recognizable. |
| 5 — New Way | Defines the capability or behavior change needed — without naming your company as the answer. |
| 6 — Your Solution | Introduces your specific offering as uniquely able to deliver what Step 5 described. |

Mark any content that is about your company (history, team, awards, logo wall, locations) as Unclassified — it does not belong in the sequence.

**Why this matters:** The mapping exposes sequence violations. A pitch may technically contain a Reframe — but if it appears after Your Solution, the choreography is broken.

---

### Step 3 — Score Each Step Against Its Criteria

For each step that appears in the pitch, assign a score: **Pass**, **Partial**, or **Fail**. For steps that are absent, mark **Missing**.

Use the detailed per-step criteria from `references/six-step-rubric.md`. The summary criteria are:

**Step 1 — Warmer**
- Pass: hypothesis-based, uses benchmarks or peer anecdotes, ends with reaction question
- Fail: opens with "what's keeping you up at night?" or company overview instead
- Fail: no customer context at all — pitch jumps to data or product

**Step 2 — Reframe**
- Pass: presents an insight the customer has not already considered; produces surprise
- Fail: customer's likely response is "I totally agree" rather than "I never thought of it that way"
- Fail: missing entirely — pitch goes from challenges directly to data or product
- Partial: mild reframe that challenges at the margins without genuine surprise

**Step 3 — Rational Drowning**
- Pass: ROI is on the customer's problem (cost of the reframe issue), not on buying your product
- Fail: ROI calculator shows "you'll save X% with our platform" — product ROI, not problem ROI
- Fail: missing or too thin — data does not disturb, only informs

**Step 4 — Emotional Impact**
- Pass: tells a story about similar companies that triggers recognition; customer sees themselves
- Fail: missing — pitch goes from data to product with no narrative bridge
- Fail: story is about your customer success cases, not about the customer's own behavior pattern

**Step 5 — New Way**
- Pass: describes solution capabilities without naming your company; customer agrees to the concept
- Fail: describes generic best practices available from any vendor (teaching-into-the-desert risk)
- Fail: accidentally names your product (premature reveal)
- Fail: missing — pitch jumps from emotional story directly to product

**Step 6 — Your Solution**
- Pass: connects your specific capabilities back to the New Way requirements; differentiation is earned
- Fail: product presentation leads the pitch (lead-with error — your solution appears in Step 1 position)
- Partial: solution is presented but connection to Step 5 is loose

---

### Step 4 — Detect Lead-With vs. Lead-To Error

Run a focused check on the opening of the pitch (first 20% of content).

**Check A — Solution-first opening:**
Does the pitch open with any of the following?
- The seller's product or platform name
- A product feature or capability statement
- A company overview or "About Us" section
- A value proposition or tagline referencing what the seller does

If yes, mark: **Lead-With Error Detected**. The pitch opens with the seller's story before establishing why the customer should care.

**Check B — First-slide diagnosis:**
For deck audits — what is on slides 1–3? If the answer is any combination of: company values, team credentials, solution capabilities, partner logos, or customer logo walls — those slides are all lead-with signals. They have no place in the opening of a Commercial Teaching pitch.

**Check C — Buzzword pollution scan:**
Search the full artifact for these terms (from Ch9 analysis of top overused pitch language): leading, unique, innovative, solution, best, top, largest, innovator, world-class, cutting-edge, state-of-the-art.

Count occurrences. More than 3 in a short pitch is a pollution signal. More than 6 means the pitch is indistinguishable from competitor materials.

Flag each occurrence with location and surrounding sentence.

**Why this matters:** Buzzwords do not differentiate — they signal sameness. Every competitor also claims to be "the leading solution." These words invite a price discussion rather than a conceptual sale. They also indicate that the pitch is leading with product identity rather than customer insight.

---

### Step 5 — Detect Teaching Into the Desert

This check addresses a subtler failure: the pitch has a real Reframe (Step 2), but the insight does not route back to capabilities that only this seller can deliver.

**The test:** For each insight or Reframe identified in Step 2, ask:

1. Does the New Way (Step 5) describe capabilities that are genuinely differentiating for this seller?
2. Can any of the seller's three largest competitors equally satisfy the Step 5 capability requirements?
3. If the customer takes this insight to bid, does the seller win — or does any qualified competitor win?

If competitors can equally satisfy the New Way requirements, mark: **Teaching-Into-the-Desert Risk Detected**.

**Manifestations to look for:**
- Step 5 describes generic technology adoption, process improvement, or organizational change that any consultant could recommend
- The insight is industry-available (customer has heard it before, just not from this rep)
- Step 6 (Your Solution) does not map cleanly to Step 5 (New Way) — there is a gap between what was described as needed and what the seller actually offers

**Why this matters:** Free consulting that benefits your competitors is the worst possible outcome of a teaching pitch. The four rules of Commercial Teaching require that every insight lead to a capability where you outperform the competition — otherwise you are educating the market on behalf of your rivals.

---

### Step 6 — Detect Missing or Weak Reframe

A separate targeted check on Step 2.

**Presence check:** Is there any moment in the pitch that introduces an unexpected angle — something the customer has not already considered? If no — the Reframe is missing. This is the most consequential single failure in the choreography.

**Quality check for present Reframes:**
- Would a customer's first reaction be "I never thought of it that way" — or "yes, exactly, that's what we're all dealing with"?
- Does the Reframe connect challenges from Step 1 to a larger problem the customer did not know they had?
- Is it delivered as a headline (creating curiosity) or as a full explanation (killing the tension)?
- Does it create mild dissonance — or does it feel like validation?

**Common Reframe failure modes:**
- Reframe describes a well-known industry problem (everyone has already heard it) → teaching at the margins
- Reframe is so gently framed it reads as agreement rather than challenge
- Reframe makes a claim the customer immediately agrees with → they already have a solution in mind → commoditization risk

Mark as: **Reframe Missing**, **Reframe Weak**, or **Reframe Effective**.

---

### Step 7 — Write the Pitch Diagnosis Report

Write the output to `pitch-diagnosis.md` in the same directory as the input artifact, or in the working directory if no artifact path is specified.

The report contains four sections:

**Section 1 — Sequence Map**

A table showing how the pitch sections map to choreography steps, with any sequence violations called out.

```
| Pitch Section | Assigned Step | Sequence Position | Notes |
|---|---|---|---|
| Slide 1: About Our Company | Unclassified | 1st | Lead-with signal — company content in opening |
| Slide 2: Our Leading Platform | Step 6 | 2nd | Solution appears at position 2 of 8 — severe lead-with |
| Slide 3: Industry Challenges | Step 1 (Warmer) | 3rd | Correct step, wrong position |
...
```

**Section 2 — Per-Step Scorecard**

```
Step 1 — Warmer:           [ Pass / Partial / Fail / Missing ]
Step 2 — Reframe:          [ Pass / Partial / Fail / Missing ]
Step 3 — Rational Drowning: [ Pass / Partial / Fail / Missing ]
Step 4 — Emotional Impact:  [ Pass / Partial / Fail / Missing ]
Step 5 — New Way:           [ Pass / Partial / Fail / Missing ]
Step 6 — Your Solution:     [ Pass / Partial / Fail / Missing ]

Lead-with error:            [ Detected / Not detected ]
Buzzword pollution:         [ N occurrences / None ]
Teaching-into-the-desert:   [ Risk detected / Safe ]
Reframe quality:            [ Missing / Weak / Effective ]
```

**Section 3 — Flagged Passages**

For each issue detected, quote the specific passage and label the problem:

```
[LEAD-WITH] Slide 1, sentence 2:
"Our platform is the leading solution for enterprise supply chain optimization."
Problem: Solution claim in opening position. Customer has not yet been given a reason to care.
Rewrite direction: Replace with a hypothesis about supply chain challenges at companies like theirs.

[BUZZWORD] Slide 3:
"Our innovative, world-class team delivers unique value."
Problem: 4 buzzwords in one sentence. These terms appear in every competitor's deck.
Rewrite direction: Replace with a specific, measurable capability or a customer outcome.

[TEACHING-INTO-THE-DESERT] Slide 6 (New Way):
"Companies need to adopt AI-powered analytics to reduce decision latency."
Problem: Any vendor with an analytics product can satisfy this requirement.
Rewrite direction: Specify a unique capability only this seller delivers — or revise Step 2 to teach an insight that naturally leads to a more differentiated capability.
```

**Section 4 — Priority Rewrite Recommendations**

Rank the top 3 changes the rep or enablement team should make, ordered by impact:

1. **Highest impact first** — typically: remove solution-first content from opening positions
2. **Second:** install or strengthen the Reframe
3. **Third:** correct the Rational Drowning focus from product ROI to problem ROI

For each recommendation, provide one concrete rewrite direction (not a full rewrite — that is `author-commercial-teaching-pitch`).

---

## Output

The primary output is `pitch-diagnosis.md` containing all four report sections.

Secondary outputs (inline in the conversation):
- A one-paragraph summary of the pitch's overall Commercial Teaching fitness
- The three highest-priority issues in plain language

The diagnosis report is not a rewrite. It is a precise diagnostic that the rep or enablement team uses to understand exactly what to fix — and why each fix matters to the customer experience.

---

## Examples

### Example 1 — Lead-With Opening

**Pitch opening:** "At Acme Corp, we are the leading provider of enterprise workflow automation. Our award-winning platform serves over 500 enterprise customers across 40 countries."

**Diagnosis:**
- Step 1 (Warmer): Fail. This is a company overview, not a hypothesis about customer challenges.
- Lead-with error: Detected. The pitch opens with the seller's identity and solution before the customer has any reason to care.
- Buzzword pollution: "leading" (1 occurrence in 2 sentences).
- Rewrite direction: Open with: "We've worked with companies similar to yours and consistently see three challenges in workflow adoption — [specific hypothesis]. Is that what you're experiencing, or would you add something?"

### Example 2 — Teaching Into the Desert

**Reframe:** "Most companies don't realize that 40% of approval bottlenecks happen outside their workflow tool — in email threads and Slack conversations. That's where decisions actually get made."

**Step 5 (New Way):** "You need a system that captures decisions wherever they happen and integrates them back into the workflow record."

**Diagnosis:**
- Reframe quality: Effective. "I never thought of it that way" is the likely reaction.
- Teaching-into-the-desert risk: Detected. The New Way describes a capability available from at least four major workflow vendors with inbox integrations. The insight does not route to a unique capability.
- Rewrite direction: Either specify what this seller does uniquely in the capture-and-integrate space, or revise the Reframe to lead to an insight where the seller's differentiation is genuinely unmatched.

### Example 3 — Missing Reframe

**Pitch structure:** Slide 1: About Us → Slide 2: Industry Challenges → Slide 3: Market Data → Slide 4: Our Solution → Slide 5: ROI Calculator → Slide 6: Next Steps

**Diagnosis:**
- Sequence map: No Reframe (Step 2) present. No New Way (Step 5) present. Pitch is: opener → warmer → Rational Drowning → solution. Three steps are missing.
- Reframe: Missing. The pitch presents challenges and data, but never introduces an unexpected angle.
- Emotional Impact: Missing. No storytelling bridge between the data and the solution.
- Impact: Without the Reframe, the customer has no reason to think this pitch differs from every competitor pitch they've seen this quarter.
- Rewrite direction: After Slide 2 (challenges), insert a Reframe slide that presents one unexpected angle the customer has not already considered. This is the single highest-impact addition.

---

## Reference Files

- `references/six-step-rubric.md` — Detailed per-step scoring criteria with pass/partial/fail indicators, failure mode examples, and the full anti-pattern quick-reference table

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill is standalone (no dependencies). It diagnoses an existing pitch. To author a new pitch from scratch, use `build-commercial-insight` followed by `author-commercial-teaching-pitch`.

FILE:references/six-step-rubric.md
# Six-Step Commercial Teaching Rubric

**Reference for:** diagnose-pitch-for-commercial-teaching-fit  
**Source:** Teaching for Differentiation chapters, The Challenger Sale (Dixon & Adamson)

Use this rubric to score each step of an existing pitch against its correct criteria. Each step receives one of three scores: **Pass**, **Partial**, or **Fail**.

---

## Step 1 — Warmer

**Function:** Establish credibility before anything else. Prove you understand the customer's world without asking them to teach you about it.

**Correct execution:**
- Opens with a hypothesis-based assessment of the customer's challenges ("We've worked with similar companies and see these three challenges most often…")
- Uses benchmarking data, peer anecdotes, or industry patterns — not generic pleasantries
- Closes with a reaction question: "Is that what you're seeing, or would you add something?"
- Signals: "I'm not here to waste your time asking questions you'd expect me to already know the answer to"

**Pass signals:**
- Customer acknowledges challenges as real
- Customer adds nuance without correcting the whole premise
- Customer is willing to keep talking

**Fail signals:**
- Opens with "What's keeping you up at night?" or open-ended needs discovery questions
- Opens with company overview, product intro, or agenda without any customer context
- No hypothesis — rep is mining for information rather than demonstrating expertise

**Partial:** Has some customer-context setup but relies on generic statements or clichéd industry pain without specificity or data.

---

## Step 2 — Reframe

**Function:** Introduce the core insight — the unexpected angle that connects the customer's acknowledged challenges to a bigger hidden problem or overlooked opportunity. This is the pivot of the entire conversation.

**Correct execution:**
- Presents an insight the customer has not already considered (not validation)
- Delivered as a headline, not a full argument (that comes in Step 3)
- Goal reaction: "Huh, I never thought of it that way before"
- Must catch the customer off guard — creates curiosity and mild dissonance

**Pass signals:**
- Customer pauses, shows surprise, asks a follow-up question
- Customer does NOT say "I totally agree" or "exactly what we've been working on"
- Tension increases — customer leans in

**Fail signals:**
- Customer enthusiastically agrees immediately — the "I totally agree!" response
- Reframe confirms a problem the customer has already articulated (teaching at the margins)
- Reframe is missing entirely — pitch moves from challenges directly to data or product
- Reframe is so hedged it creates no tension

**Partial:** Reframe exists but is mild, expected, or only marginally challenges assumptions.

**Note on missing Reframe:** If Step 2 is absent, the entire choreography fails — without the insight pivot, Steps 3–5 have no foundation and the pitch reverts to a conventional needs-based sell.

---

## Step 3 — Rational Drowning

**Function:** Quantify why the Reframe matters — the financial, operational, or strategic cost of the problem just revealed. Make the customer feel the weight of inaction.

**Correct execution:**
- ROI is calculated on the customer's problem (the reframe), not on buying your solution
- Uses data, benchmarks, models calibrated to the customer's context
- Emotional target: "Wow, I had no idea we were wasting that kind of money"
- Creates a sense of urgency — the problem is worse than they thought

**Pass signals:**
- Customer expresses surprise at the scale of the problem
- Customer begins to mentally accept the reframe as their own problem
- Numbers are credible and traceable

**Fail signals:**
- ROI calculator shows return on purchasing the seller's product/platform
- Quantification is generic ("companies like yours save 20-30%") with no customer calibration
- Data dump without emotional framing — rational but not disturbing
- Step 3 is skipped — pitch moves from Reframe directly to solution

**Partial:** Quantification is present but framed as product ROI, or data is too generic to create real urgency.

---

## Step 4 — Emotional Impact

**Function:** Ensure the customer sees themselves in the story — not just that the problem is real for someone, but that it is real for them specifically. Defeat the "we're different" objection before it is voiced.

**Correct execution:**
- Tells a story about how similar companies went down the painful path being described
- Story must feel immediately recognizable — customer should replay their own experience
- Goal: rueful head shake, wry smile, or "it's like you work here"
- Story is not about the seller's solution; it is about the customer's behavior pattern

**Pass signals:**
- Customer confirms "we do that all the time" or equivalent acknowledgment
- Customer volunteers a specific internal example
- The "we're different" objection does not appear

**Fail signals:**
- Customer says "I see what you're saying, but we're different"
- Step 4 is skipped — pitch goes from data directly to solution
- Storytelling is generic ("many companies") without specificity that triggers recognition
- Story is about the seller's product success rather than the customer's situation

**Partial:** Some narrative is present but too abstract to create emotional recognition.

---

## Step 5 — New Way

**Function:** Define the solution capabilities the customer needs — without yet naming your company as the provider. The customer must agree to the "what" before you introduce the "who."

**Correct execution:**
- Describes the behavioral change or capability set needed to capture the opportunity or solve the problem
- Framed as: "Here is how a company would need to operate differently…"
- Does NOT yet name or imply the seller's product as the answer
- Goal reaction: "You're right — that's exactly what we need to do" or "that's the kind of company I want us to be"

**Pass signals:**
- Customer agrees conceptually with the solution approach
- Customer does not yet have a specific vendor in mind
- The capability description maps clearly to Step 6 (your unique strengths)

**Fail signals:**
- Step 5 describes generic best practices any vendor can deliver (teaching-into-the-desert risk)
- Step 5 is skipped — pitch jumps from emotional story directly to product pitch
- Capabilities described in Step 5 are already commoditized in the market
- Seller accidentally names their own product in Step 5 (premature reveal)

**Partial:** New Way is described but does not clearly connect to unique seller capabilities, or is so generic it creates no differentiation.

**Teaching-into-the-desert check:** Ask — can any of your three largest competitors equally satisfy these Step 5 capability requirements? If yes, the insight does not lead to unique strengths.

---

## Step 6 — Your Solution

**Function:** Show that your specific offering is uniquely positioned to deliver the capabilities described in Step 5. This is the first moment the seller explicitly enters the conversation.

**Correct execution:**
- Connects each unique seller capability to the New Way requirements from Step 5
- Demonstrates differentiation based on capabilities the customer now values (because they followed Steps 1–5)
- Solution presentation feels like a natural conclusion, not a product pitch
- Competitive positioning is earned through the preceding teaching, not asserted upfront

**Pass signals:**
- Customer says "so you can actually help us do this?"
- No request to "just send a quote"
- Competitor is not obviously equally capable on the same criteria

**Fail signals:**
- Step 6 contains claims that contradict the uniqueness implied in Step 5
- Solution is presented without connecting to the New Way capabilities
- Step 6 is where most of the pitch lives — solution dominates the conversation rather than concluding it
- Pitch uses buzzwords ("leading solution," "unique platform," "innovative approach") that signal sameness

**Partial:** Solution is presented but connection to Steps 4–5 is loose, or competitive differentiation is asserted rather than demonstrated.

---

## Scoring Summary Template

```
Step 1 — Warmer:          [ Pass / Partial / Fail ]
Step 2 — Reframe:         [ Pass / Partial / Fail ]
Step 3 — Rational Drowning: [ Pass / Partial / Fail ]
Step 4 — Emotional Impact: [ Pass / Partial / Fail ]
Step 5 — New Way:         [ Pass / Partial / Fail ]
Step 6 — Your Solution:   [ Pass / Partial / Fail ]

Lead-with detection:      [ Clean / Triggered ]
Buzzword pollution:       [ Clean / N occurrences found ]
Teaching-into-the-desert: [ Safe / Risk detected ]
Sequence integrity:       [ Intact / Disorder detected at Step __ ]
```

**Scoring interpretation:**
- 6 Pass + all clean = pitch is well-structured for Commercial Teaching
- Any Fail in Step 2 (Reframe) = foundational gap; Steps 3–5 have no anchor
- Lead-with triggered = most urgent rewrite priority
- Teaching-into-the-desert detected = insight exists but benefits competitor as much as you

---

## Anti-Pattern Quick Reference

| Anti-Pattern | What It Looks Like | Where It Shows |
|---|---|---|
| Lead-with | First slides cover company, product, or capabilities | Steps 1–2 violated |
| Missing Reframe | No insight pivot; pitch goes Warmer → data → product | Step 2 absent |
| Rational Drowning on the wrong problem | ROI calculator shows return on buying your product | Step 3 misfocused |
| Teaching into the desert | Great insight, but Step 5 New Way is generic | Steps 5–6 mismatch |
| Buzzword pollution | "Leading," "unique," "innovative," "solution" in pitch text | Most common in Steps 1, 6 |
| Solution-first opening | Pitch opens with product name, offering, or differentiators | Violates entire sequence |
| Enthusiastic agreement Reframe | Customer says "I totally agree!" in Step 2 | Step 2 failed to reframe |

---
name: diagnose-manager-effectiveness
description: "Diagnose a frontline sales manager's effectiveness against the CEB four-driver model. Use when someone asks: 'sales manager effectiveness', 'am I coaching the right reps', 'sales manager diagnostic', 'manager coaching ROI', 'where should I spend coaching time', 'sales manager assessment', 'Challenger sales manager', 'manager drivers', 'front-line manager performance', 'am I a good sales manager', 'sales manager self-assessment', 'coaching time allocation', 'am I spending coaching time correctly', 'democratic coaching', 'why is my team not improving'. Produces a manager-effectiveness-diagnosis.md with weighted driver scores, active anti-patterns, and a concrete time-reallocation plan based on team composition."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/diagnose-manager-effectiveness
metadata: {"openclaw":{"emoji":"🧭","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors: ["Matthew Dixon", "Brent Adamson"]
    chapters: [8]
tags: [sales, sales-management, challenger-sale, manager-diagnostic, coaching-roi]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "coaching-log.md or manager self-description plus team composition (star/core/laggard breakdown)"
  tools-required: [Read, Write, AskUserQuestion]
  tools-optional: [Grep]
  mcps-required: []
  environment: "Any working directory. No codebase required."
discovery:
  goal: "Diagnose a sales manager against the CEB four-driver model and identify where coaching time is being wasted"
  tasks:
    - "Gather manager time allocation across four activity types"
    - "Apply Management Fundamentals gate — stop if not cleared"
    - "Score manager against four driver weights (Selling 27%, Coaching 28%, Innovating 29%, Resource Allocation 16%)"
    - "Detect democratic-coaching anti-pattern in coaching time distribution"
    - "Distinguish coaching activity (known behavior development) from innovating activity (deal-level obstacle solving)"
    - "Produce manager-effectiveness-diagnosis.md with reallocation plan"
  audience:
    roles: ["Frontline sales managers", "sales VPs", "sales enablement professionals", "sales operations"]
    experience: intermediate
  when_to_use:
    triggers:
      - "Manager wants to know whether they are spending their time on the highest-impact activities"
      - "Sales VP wants to diagnose which managers need development and in which driver"
      - "Enablement team wants to baseline manager cohort before a Challenger training initiative"
      - "Manager is not seeing rep performance improvement despite active coaching efforts"
    prerequisites: []
    not_for:
      - "Running a coaching session with a rep (use coach-rep-with-pause-framework instead)"
      - "Planning a Challenger rollout across a sales organization (use plan-challenger-model-rollout)"
      - "Assessing individual rep selling profiles (use classify-rep-profile)"
  environment:
    codebase_required: false
    codebase_helpful: false
    works_offline: true
  quality:
    scores:
      with_skill: null
      baseline: null
      delta: null
    tested_at: null
    eval_count: null
    assertion_count: 6
    iterations_needed: null
---

# Diagnose Manager Effectiveness

Diagnoses a frontline sales manager against the CEB five-factor effectiveness model (derived from 2,500+ managers, 12,000+ reps). Identifies whether the manager has cleared the Management Fundamentals gate and where their time allocation is misaligned with the four driver weights that predict manager excellence: Sales Innovation (29%), Coaching (28%), Selling (27%), Resource Allocation (16%). Produces a structured diagnosis artifact with anti-pattern flags and a reallocation plan.

The core counterintuitive finding from the CEB data: most managers think sales leadership is about resource allocation and pipeline management. The data says those are the least important activities. The single biggest driver is sales innovation — a skill most sales leaders have never systematically thought about.

Driver weights and the star/core/laggard coaching ROI table are in `references/manager-drivers.md`.

## When to Use

- A manager asks "am I spending my time correctly?" or "why isn't my team improving despite my coaching?"
- A sales VP wants to rank managers by development priority and identify which driver to invest in
- An enablement team wants to baseline the management cohort before a Challenger training program
- A manager feels like they're working hard but the numbers aren't moving

**Not for:** executing a coaching session with a rep (that's coach-rep-with-pause-framework), planning an organization-wide Challenger rollout, or assessing individual rep selling profiles.

## Step 1 — Gather Manager Context

Before scoring, collect the following. If the manager has provided a document (coaching-log.md or self-description), read it first. Then ask for anything missing.

**Time allocation** — Ask the manager to estimate their average week:
- What percentage of time goes to selling or deal support (covering territories, supporting complex deals, modeling behaviors for the team)?
- What percentage goes to coaching reps (1:1s, ride-alongs, call debriefs focused on rep skill development)?
- What percentage goes to deal-level innovation (working with reps to unstick specific stuck deals, finding creative paths through customer obstacles)?
- What percentage goes to resource allocation (pipeline reviews, CRM, territory management, process compliance, activity tracking)?

If percentages don't sum to roughly 100%, ask for clarification. If the manager cannot distinguish coaching from deal innovation, note that — it is a diagnostic signal in itself.

**Team composition** — Ask:
- How many reps on the team?
- Roughly what fraction are stars (consistently above quota), core (at or near quota), or laggards (chronically below, not on a recovery track)?
- Which tier gets the most coaching time currently?

**Management fundamentals check** — Ask the manager to reflect on five binary questions:
1. Do your reps consider you reliable — that you follow through on what you commit to?
2. Do your reps trust that you act with integrity in performance conversations?
3. Do you listen before directing in conversations?
4. Do you recognize rep contributions visibly and specifically?
5. Do you actively build team cohesion and protect psychological safety?

These are pass/fail. A "no" to any one of them triggers the gate in Step 2.

## Step 2 — Management Fundamentals Gate

Management fundamentals are a prerequisite, not a driver to develop alongside the others. They are binary traits — either present or not — and they cannot be coached into existence. CEB found ~4% of managers fail on at least one.

**If the manager fails on any fundamental:**
- Stop the downstream analysis. Flag this clearly in the output.
- The recommendation is not "work on reliability" — it is organizational: this manager should be moved to a non-management role. Investing in their Selling, Coaching, or Innovating skills while management fundamentals are broken produces no return.
- Output a gate-failure diagnosis and end there.

**If the manager passes all five:**
- Proceed to Step 3.

**Why this matters:** Management fundamentals provide the foundation every other driver builds on. A technically skilled manager with an integrity problem creates a fundamentally unsafe environment where coaching cannot land, innovation cannot happen, and rep retention collapses.

## Step 3 — Score Time Allocation Against Driver Weights

The four sales-side drivers and their empirically derived weights:

| Driver | Weight | What It Covers |
|---|---|---|
| Sales Innovation (Innovating) | 29% | Unsticking stuck deals, collaborative deal-level problem solving |
| Coaching | 28% | Developing rep skills toward known behaviors, behavior-focused 1:1s |
| Selling | 27% | Modeling Challenger behaviors, supporting complex deals, covering vacancies |
| Resource Allocation | 16% | Pipeline reviews, CRM compliance, territory management, activity tracking |

For each driver, compare the manager's reported time to the weight. Calculate the gap:

```
Gap = Reported time % − Driver weight %
```

Positive gap = over-indexed. Negative gap = under-indexed.

**Flag as high-priority when:**
- Innovation gap is more negative than −10 points (most common under-investment)
- Coaching gap is more negative than −10 points
- Resource allocation gap is more positive than +15 points (most common over-investment)

**Common mismatch pattern:** Managers over-invest in resource allocation (pipeline reviews, CRM) because it feels like management and is easy to measure. The data says this is the least impactful driver. The most impactful driver — innovation — is rarely tracked, rarely recognized, and rarely developed.

## Step 4 — Detect Democratic Coaching Anti-Pattern

The democratic coaching anti-pattern is one of the most expensive mistakes in sales management: spreading coaching time equally across all reps regardless of performance tier.

**Why it produces near-zero return:**
- Coaching **laggards** (low performers) has near-zero impact. You cannot coach away a poor job fit. These reps need a different role, not better coaching.
- Coaching **stars** (high performers) produces marginal gains. Like a professional golfer shaving one stroke off their average — small incremental improvements on an already high baseline.
- Coaching **core** (median) performers is where all the upside lives. The CEB data shows a significant coaching quality improvement can boost core performer output by up to **19%**. Even moving from bottom-third to top-third coaching quality produces a **6-8% performance gain** for core reps.

**Detection questions:**
- Does the manager spend roughly equal time coaching each rep regardless of tier?
- Does the manager spend disproportionate time on their two most challenging reps (laggards)?
- Does the manager treat star 1:1s as coaching sessions rather than innovation conversations?

**Flag the anti-pattern when:**
- Coaching time is distributed roughly equally across all tiers
- More than 40% of coaching time goes to laggards
- Stars receive structured behavior-development coaching rather than deal-level collaboration

**Correct allocation:** Shift the majority of scheduled coaching time to core/median performers. Reserve time with stars for deal innovation conversations (collaborative problem solving on stuck deals). Move chronic laggards through performance management rather than coaching.

## Step 5 — Distinguish Coaching From Innovating

Many managers confuse these or default entirely to one mode. This is a second diagnostic check independent of time allocation.

**Ask the manager to describe a recent interaction where they helped a rep:**
- If the manager directed the rep toward a specific behavior they knew was missing → that is **coaching**
- If the manager and rep worked through an unknown together to find a path forward on a stuck deal → that is **innovating**

**Red flags indicating confusion:**
- Manager describes all rep interactions as "coaching" including deal problem-solving
- Manager cannot recall any instance of genuine deal-level innovation (suggests innovating is absent)
- Manager describes innovation as "coming up with new value propositions" (that is not what sales innovation means — it is creatively connecting existing capabilities to specific customer obstacles)

**Why the distinction matters:** Coaching builds repeatable rep skills over time. Innovating closes specific stuck deals now. A manager who only coaches leaves revenue on the table. A manager who only innovates never improves rep capability. Both are required — and both count toward the 28%/29% of manager excellence.

## Step 6 — Write manager-effectiveness-diagnosis.md

Write the diagnosis artifact to `manager-effectiveness-diagnosis.md` in the current directory. Structure:

```markdown
# Manager Effectiveness Diagnosis
Generated: [date]

## Management Fundamentals — [PASS / GATE FAILURE]
[If gate failure: stop here with recommendation to redeploy]

## Time Allocation vs. Driver Weights
| Driver | Reported % | Target Weight | Gap | Status |
|---|---|---|---|---|
| Sales Innovation | X% | 29% | ±N | Over/Under/Aligned |
| Coaching | X% | 28% | ±N | Over/Under/Aligned |
| Selling | X% | 27% | ±N | Over/Under/Aligned |
| Resource Allocation | X% | 16% | ±N | Over/Under/Aligned |

## Active Anti-Patterns
[List detected anti-patterns with evidence. If none, say so.]

### Democratic Coaching — [DETECTED / NOT DETECTED]
[Evidence: how coaching time is distributed across tiers]
[Estimated coaching time wasted on low-ROI tiers]

### Coaching/Innovation Confusion — [DETECTED / NOT DETECTED]
[Evidence from manager's description of their interactions]

## Team Composition Analysis
[Star/core/laggard breakdown]
[Where coaching ROI is currently being captured vs. where it should be]

## Time-Reallocation Recommendations
1. [Specific reallocation: e.g., "Shift 15% of time from pipeline review to structured core-rep coaching 1:1s"]
2. [Specific reallocation: e.g., "Replace star coaching sessions with deal-innovation conversations on stuck deals"]
3. [Specific reallocation: e.g., "Reduce resource allocation from 40% to 16% target — delegate CRM reviews"]

## Priority Development Focus
[Single most impactful driver to invest in, with rationale]
[If innovating is the gap: emphasize this is the most overlooked skill in sales management]
```

Keep the artifact factual and specific. Avoid vague recommendations like "coach more." Every recommendation should specify which rep tier, which activity type, and what to reduce to make room.

## Self-Check Before Delivery

Before writing the final artifact, verify:

- Management fundamentals gate was applied — if any fundamental fails, the downstream analysis stops
- All four driver weights are present in the table (27/28/29/16)
- The democratic coaching check is explicit (not just implied)
- Coaching and innovating are assessed separately — not combined
- Recommendations are specific to this manager's team composition and actual time allocation, not generic
- Star/core/laggard ROI is used to justify coaching reallocation recommendations

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill is standalone (no dependencies). To execute a coaching session against the diagnosed gaps, invoke `coach-rep-with-pause-framework`. To plan a team-wide methodology rollout informed by manager diagnoses, invoke `plan-challenger-model-rollout`.

FILE:references/manager-drivers.md
# Reference: Manager Effectiveness Driver Definitions

Source: The Challenger Sale, Chapter 8 — The Manager and the Challenger Selling Model
CEB Sales Leadership Diagnostic: 65+ companies, 12,000+ reps, 2,500+ managers, 64 attributes

---

## The Five-Factor Model

CEB's factor analysis reduced 64 behavioral attributes to 5 statistically significant categories. Regression against actual manager performance (rep-rated + company-rated) produced the driver weights below.

---

## Management Fundamentals (Prerequisite Gate)

**What it covers:** Integrity, reliability, listening skills, recognition, team-building

**Critical property:** Binary, not developmental. These are not skill gaps to close — they are hire/no-hire traits. A manager either has them or doesn't.

**Threshold:** Roughly 4% of managers in the CEB sample fail on at least one management fundamental.

**Action when failed:** Redeploy. Do not invest coaching or development resources in a manager who has not cleared this gate.

**Why it matters:** You cannot build sales-specific excellence on a broken management foundation. Management fundamentals form the essential base; everything else is built on top.

---

## Sales-Side Driver Weights

These four drivers account for the sales-specific portion of manager effectiveness (the remaining ~75% after management fundamentals). Weights are recalibrated to 100% of the sales side.

### Selling — ~27%

**Definition:** Manager's ability to model and execute Challenger selling behaviors when needed.

**When it activates:**
- Covering vacant sales territories
- Providing critical support on the largest, most complex deals
- Modeling specific Challenger behaviors for the team

**Key attributes (from CEB factor analysis):**
- Offers the customer unique perspectives
- Tailors the offer to the specific needs of each customer
- Is comfortable discussing money and investment
- Can navigate complex negotiations

**Important limitation:** Selling well does not make someone a great manager. The driver says great managers tend to be skilled sellers — not that selling skill predicts management success.

### Coaching — 28%

**Definition:** Working side-by-side with reps to diagnose and correct specific rep behaviors known to hinder high performance. The manager already knows what "good" looks like and imparts that understanding to the rep.

**Key coaching behaviors (from CEB data):**
- Guiding reps to tailor their pitch effectively
- Showing reps how and when to assert control in a sale
- Helping reps navigate complex negotiations and stakeholder dynamics

**Prerequisites for effective coaching:**
- A hypothesis of what good looks like before entering the conversation
- A coaching environment separate from performance management (psychological safety)
- A skill the rep actually needs to develop (coaching cannot fix a bad fit)

**Coaching vs. managing:** Many managers believe they coach but actually tell, direct, and manage. True coaching is hypothesis-based, rep-centered, and behavior-specific.

### Sales Innovation (Innovating) — 29%

**Definition:** Creatively connecting existing organizational capabilities to specific customer obstacles that are blocking a deal from closing. Neither the manager nor the rep knows the path forward — they discover it together.

**When it applies:**
- A deal is stuck despite correct Challenger behaviors from the rep
- A customer is choosing "no decision" over a good option
- The path to consensus is blocked by an unexpected obstacle

**What makes it different from coaching:**

| | Coaching | Sales Innovation |
|--|--|--|
| Manager's role | Impart known answer | Collaborate to discover answer |
| Manager's knowledge | Knows what "good" looks like | Doesn't know the answer either |
| Situation | Behavior improvement | Deal-level obstacle |
| Method | Directive, hypothesis-based | Collaborative, exploratory |

**Why it's the single biggest driver:** Unlike coaching (which has received enormous attention), sales innovation is rarely discussed as a systematic management skill. It is the "missing link" — even a team of skilled Challenger reps will leave deals stuck without an innovative manager.

**Commander's Intent analogy:** Like military leaders who issue intent rather than step-by-step orders, innovative managers help reps adapt deal strategy to "the reality on the ground."

### Resource Allocation — 16%

**Definition:** Territory management, process compliance, pipeline reviews, activity efficiency, and resource distribution across reps and accounts.

**Why it scores lowest:** Most sales leaders intuitively believe resource allocation is the core of sales leadership. The data says it's the least impactful driver of manager excellence. It is necessary but not differentiating.

**Practical implication:** Managers who spend the majority of their time on pipeline reviews, CRM hygiene, and activity tracking are optimizing the least impactful driver while underinvesting in innovation and coaching.

---

## Star / Core / Laggard Coaching ROI Table

| Rep Tier | Coaching Impact | Why |
|---|---|---|
| **Laggards** (bottom performers) | Near-zero | Cannot coach away a poor job fit. The rep needs a different role, not better coaching. |
| **Stars** (top performers) | Marginal | Already self-directed. Like a professional golfer shaving one stroke — incremental at best. |
| **Core** (median performers) | High — up to 19% output boost | Most responsive to coaching quality. The entire upside of coaching investment lives here. |

**The distribution effect:** When coaching quality improves, the performance curve tips rather than shifts. The core (middle) moves substantially; the tails (stars and laggards) barely move.

**Quantified impact on core performers:**
- Significant coaching quality improvement: up to **19% performance boost**
- Moving from bottom-third to top-third coaching: **6-8% performance gain**
- Real-world case (insurance company): **10% improvement** for coached reps vs. uncoached

---

## Democratic Coaching Anti-Pattern

**Pattern:** Spreading coaching time equally across all reps regardless of tier.

**Why it fails:** Concentrating coaching time on laggards or stars produces near-zero return. The entire ROI of coaching investment lives in core performers — but democratic coaching systematically underweights them.

**Detection signals:**
- Manager spends equal time across all reps in 1:1s
- Manager focuses on their two "problem reps" or their top performer
- Coaching conversations happen reactively rather than on a scheduled, structured basis

**Correct allocation:** Shift the majority of coaching time to core/median performers. Reserve star time for innovation conversations (deal-level problem solving, not behavior coaching). Redeploy or manage out chronic laggards through performance management rather than coaching.

---
name: coach-rep-with-pause-framework
description: "Plan a structured coaching session for a sales rep using the PAUSE framework (Prepare → Affirm → Understand → Specify → Embed). Use when a manager needs to: run a 'sales coaching session', 'coach a sales rep', use 'PAUSE coaching', ask 'pre-call coaching questions', run a 'post-call debrief', do 'challenger rep coaching', follow a 'sales manager coaching guide', 'coach the rep not the deal', answer 'how to coach my rep', or build a '1:1 coaching agenda'. Reads rep-profile-assessment.md from classify-rep-profile to identify the rep's weakest Challenger pillar, selects the Appendix A question set for that pillar, and produces a coaching-session-plan.md with pre-call questions, observation focus, post-call debrief, and an embed homework assignment. Also includes a decision gate: if the issue is deal-specific (stalled deal, unknown obstacle) rather than a rep-behavior gap, routes to the sales innovation mode (Investigate → Create → Share + optional SCAMMPERR) instead of PAUSE coaching."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/coach-rep-with-pause-framework
metadata: {"openclaw":{"emoji":"🎓","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors: ["Matthew Dixon", "Brent Adamson"]
    chapters: [8, "Appendix A"]
tags: [sales, sales-management, challenger-sale, sales-coaching, manager-coaching, rep-development]
depends-on: [classify-rep-profile]
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "rep-profile-assessment.md (from classify-rep-profile) + deal-brief.md for the live deal the coaching session will anchor to + optional call-transcript.md from the most recent interaction with that customer"
  tools-required: [Read, Write, AskUserQuestion]
  tools-optional: [Grep]
  mcps-required: []
  environment: "Any working directory. No codebase required."
discovery:
  goal: "Plan a PAUSE coaching session tailored to the rep's weakest Challenger pillar, producing a ready-to-run coaching-session-plan.md"
  tasks:
    - "Read rep-profile-assessment.md to identify weakest Challenger pillar"
    - "Apply the coaching vs. innovating decision gate"
    - "If coaching: run PAUSE five-step protocol with pillar-specific Appendix A questions"
    - "If innovating: run Investigate → Create → Share on the stalled deal"
    - "Produce coaching-session-plan.md with pre-call, observation focus, post-call, and embed homework"
  audience:
    roles: ["sales managers", "front-line sales leaders", "sales enablement professionals"]
    experience: intermediate
  when_to_use:
    triggers:
      - "Manager has a rep diagnostic (rep-profile-assessment.md) and wants to run a coaching session"
      - "Manager wants to structure a 1:1 coaching conversation around a specific live deal"
      - "Manager wants pre-call and post-call coaching questions for a specific Challenger pillar"
      - "Manager wants a post-call debrief agenda for a call the rep just completed"
    prerequisites:
      - "rep profile assessment — run classify-rep-profile first to produce rep-profile-assessment.md"
    not_for:
      - "Manager self-diagnosis (use diagnose-manager-effectiveness)"
      - "Hiring or candidate screening"
      - "Rep profile classification (use classify-rep-profile)"
  environment:
    codebase_required: false
    codebase_helpful: false
    works_offline: true
  quality:
    scores:
      with_skill: null
      baseline: null
      delta: null
    tested_at: null
    eval_count: null
    assertion_count: 7
    iterations_needed: null
---

# Coach Rep with PAUSE Framework

Plans a complete, pillar-targeted coaching session using CEB's PAUSE framework (Prepare → Affirm → Understand → Specify → Embed) from Chapter 8 of _The Challenger Sale_, anchored to the Appendix A coaching questions per Challenger pillar.

The key prerequisite: run `classify-rep-profile` first. Its output (`rep-profile-assessment.md`) tells you which of the three Challenger subscales — Teaches for Differentiation, Tailors for Resonance, Takes Control — is the rep's weakest. This skill selects the matching Appendix A question set and builds a coaching session around that single pillar.

One session, one pillar, one observable change. That is the design principle.

## When to Use

- You have a rep who has completed the Appendix B self-diagnostic and you want to run their first structured coaching session
- You observed a call or meeting and want to structure specific feedback using a documented coaching protocol
- You want pre-call questions to prime the rep before an important customer meeting, tied to a specific Challenger behavior
- You have a deal you want to use as the coaching anchor and need a structured post-call debrief agenda
- Your rep is executing the right behaviors sometimes but not consistently — you want to embed the change, not just name it

**Not for:** deals that are stalled because neither you nor the rep knows what is blocking progress (that is the innovating mode — see Step 3). Not for manager self-diagnosis (use `diagnose-manager-effectiveness`).

## Context and Input Gathering

Before generating a coaching plan, collect:

1. **rep-profile-assessment.md** — produced by `classify-rep-profile`. Contains three subscale scores and identifies the weakest pillar. Read it before starting.
2. **deal-brief.md** — the specific live deal the coaching session will anchor to. A coaching conversation is most effective when it is grounded in a real deal the rep is actively working.
3. **call-transcript.md** (optional) — the most recent call or meeting with the anchor deal customer. If available, read it to identify specific moments where the target behavior was missing or weak.

If no deal brief is available, ask: "Which customer deal do you want to use as the coaching anchor for this session? Describe the customer, deal stage, and what happened in the most recent interaction."

## Process

### Step 1 — Read rep-profile-assessment.md and identify the target pillar

Read `rep-profile-assessment.md`. Locate the Challenger subscale scores:

| Subscale | Questions | Score | Threshold |
|---|---|---|---|
| Teaches for Differentiation | Q2+Q3 | x/10 | Strong: 8+ / Foundation: 5-7 / New territory: ≤4 |
| Tailors for Resonance | Q5+Q6 | x/10 | Strong: 8+ / Foundation: 5-7 / New territory: ≤4 |
| Takes Control | Q8+Q9 | x/10 | Strong: 8+ / Foundation: 5-7 / New territory: ≤4 |

Identify the single lowest-scoring subscale. That is the target pillar for this coaching session.

WHY: Splitting coaching attention across multiple pillars dilutes the session. CEB's PAUSE framework is designed to land one specific, observable behavior change per session. Trying to address all three pillars at once produces generic feedback that does not change behavior.

If two subscales are tied at the same score, ask the manager: "Between these two areas, which one showed up most clearly as a gap in the rep's most recent customer interaction?" Use the answer to select one.

### Step 2 — Load the deal brief and read the call transcript if available

Read `deal-brief.md`. Note:
- Customer name and deal stage
- What happened in the most recent interaction
- What is blocking progress or causing concern

If `call-transcript.md` is available, read it. Look specifically for moments where the target Challenger behavior (from Step 1) was either absent, weak, or handled differently than the Appendix A standard suggests.

WHY: The PAUSE framework is Hypothesis-Based Coaching. The hypothesis is: "In this specific interaction, the rep was not executing [target behavior]. Here is the specific moment I observed it." Coaching without a specific observed moment becomes abstract and hard for the rep to act on.

### Step 3 — Coaching vs. innovating decision gate

Before running PAUSE, apply this decision gate:

**Is this a rep-behavior gap?**
- The manager can name specifically what "good" looks like for the target behavior
- The rep knows the right behavior in principle but is not executing it consistently
- The gap is about skill, habit, or confidence — not about the deal situation being unusual

→ If yes: **COACH** using PAUSE (Steps 4–8 below)

**Or is this a deal-specific obstacle?**
- The deal is stalled and neither the manager nor the rep knows what is blocking it
- The obstacle is unique to this customer's situation and no standard playbook applies
- The question is "how do we get this deal unstuck?" not "how does the rep sell better?"

→ If yes: **INNOVATE** using Investigate → Create → Share (Step 9 below)

WHY: Coaching and innovating are distinct activities that require different manager modes. Coaching assumes the manager knows the answer and imparts it. Innovating assumes neither party knows the answer and they must collaborate to discover it. Applying coaching when the situation requires innovation means giving prescriptive direction to a problem that has no known solution — which often accelerates a deal loss rather than preventing it.

### Step 4 — PAUSE: Prepare

Before the coaching session begins, the manager completes preparation:

**What to review:**
- The rep's subscale score and the specific threshold it falls into
- Prior coaching notes or action items from the last session (maintain continuity)
- The deal brief and any call transcript for the anchor deal
- The Appendix A pre-call questions for the target pillar (see `references/appendix-a-questions.md`)

**Form the coaching hypothesis:**
Write one sentence that captures what "good" would have looked like in the specific interaction that anchored this session. Example: "In the discovery call with [Customer], good Teach behavior would have been introducing a specific insight about [industry problem] before asking about their situation — instead, the rep led with product features."

WHY: The hypothesis gives the coaching session its direction. Without it, the session becomes a general discussion rather than a targeted development interaction. The hypothesis also prevents the manager from coaching based on outcome (the deal didn't advance) rather than behavior (what the rep did that contributed to that outcome).

**Choose one observation focus:**
Select the specific Challenger behavior to observe in the next live interaction. Keep it narrow: one behavior, not a checklist.

### Step 5 — PAUSE: Affirm

Open the coaching conversation by naming what the rep is doing well — before surfacing the gap.

**How to affirm:**
- Reference a specific recent moment, not a general compliment
- Connect it to a Challenger behavior the rep has shown, even partially
- Keep it brief — one to two sentences

Examples:
- "I noticed in that call that you stayed with the pricing conversation rather than pivoting away from it. That took confidence."
- "Your prep work on the customer's competitive position was thorough — you clearly understood their market pressure."

WHY: If a rep does not feel safe in the coaching session, feedback is wasted. The Affirm step is not a softening technique — it is a precondition for the rep to hear what follows. Separating coaching from performance management requires the rep to experience this interaction as development, not evaluation. Starting with a genuine observation of strength establishes that.

Do not skip Affirm even if the session is urgent or the manager feels the rep needs direct correction. A coaching session where the rep is defensive accomplishes nothing.

### Step 6 — PAUSE: Understand

Invite the rep to self-diagnose before offering the manager's view.

Use the Appendix A pre-call planning questions for the target pillar (from `references/appendix-a-questions.md`) as the structure for this conversation:

**If target pillar is Teach:**
Ask the rep the Teach pre-call questions — but framed as reflection on the past interaction rather than preparation for a future one. Example: "Before that call, what did you think the business problem you were going to surface was? How did you know it was genuinely critical to them?"

**If target pillar is Tailor:**
Ask the Tailor questions as post-interaction reflection. Example: "Going into that meeting, what did you know about the trends hitting their industry right now? What was unique about their competitive position?"

**If target pillar is Take Control:**
Ask the Take Control questions. Example: "What was your intended next step at the end of that call? What do you know about how their organization actually makes this kind of purchase decision?"

WHY: Reps who self-identify a gap own the change. Reps who are told they have a gap may agree verbally but resist behaviorally. The Understand step is not a quiz — it is an opportunity for the rep to arrive at the gap diagnosis through their own reflection, with the manager holding the framework that structures that reflection.

Listen carefully. If the rep's self-diagnosis is accurate, confirm it and move to Specify. If the rep's self-diagnosis misses the actual gap, ask one follow-up question to guide them closer before naming what you observed.

### Step 7 — PAUSE: Specify

Agree on one specific, observable behavior change for the rep's next customer interaction.

**Format of the specification:**
"In your next call with [Customer / account type], when [specific moment], you will [specific observable action] instead of [current pattern]."

Examples by pillar:

**Teach:** "In your next discovery call, you will open with the industry insight about [problem] before asking about their situation — even if it feels like leading with a conclusion. The question to ask yourself afterward: did I teach them something, or did I just ask them questions?"

**Tailor:** "Before your next stakeholder meeting, you will write down one thing that is unique about this company's market position and one place where they are most vulnerable. You will find a way to reference at least one of those observations during the conversation."

**Take Control:** "At the end of your next meeting with this customer, you will name a specific next step with an owner and a date — out loud, in the meeting — rather than leaving it for a follow-up email."

WHY: Generic feedback ("be more assertive," "teach better") cannot be observed, measured, or coached at the next session. A specific observable behavior can be tracked, which gives the Embed step something to anchor to. CEB's PAUSE design requires the Specify step to be concrete enough that both the manager and rep can answer "did they do it or not?" after the next interaction.

Limit to one change per session. A list of behaviors is not a Specify — it is a performance review.

### Step 8 — PAUSE: Embed

Plan how the behavior change will be reinforced beyond this session.

**Three components of the Embed plan:**

**1. Homework before the next call:**
The rep completes a brief preparation task using the Appendix A pre-call questions for the target pillar before their next customer interaction with the anchor deal. Manager reviews it before the call.

**2. Observation checkpoint:**
If the manager can observe the next interaction (live, recorded, or ride-along), schedule it now. If not, agree on when the rep will debrief immediately after the call.

**3. Post-call debrief agenda:**
Use the Appendix A post-call questions for the target pillar (from `references/appendix-a-questions.md`) as the debrief structure. Schedule the debrief for within 24 hours of the interaction.

WHY: A coaching session that ends without continuity into the next interaction creates awareness but not change. The Embed step is what separates a coaching conversation from a one-time event. It ensures that the next coaching session begins with evidence of whether the specific behavior occurred — which gives coaching its cumulative effect over time.

Also: update the coaching log after the session. Note the target pillar, the specific observable change agreed on, and the scheduled debrief date. This creates the continuity that the next session's Prepare step will use.

### Step 9 — Innovating mode (when the gate routes here instead of PAUSE)

When the situation is a deal-specific obstacle — neither the manager nor the rep knows what is blocking progress — use the three-part sales innovation process:

**Investigate:** Work with the rep to map the customer's decision-making process in as much detail as possible. Ask:
- Who is involved in this decision beyond the contacts you have already spoken with?
- What decision criteria is this customer applying — and how do you know?
- What financial, organizational, or political concerns might be blocking progress that the customer has not stated directly?

The goal is to understand what is actually happening, not assume the obstacle is what it appears to be.

**Create:** Co-create options with the rep without immediately evaluating them. This is thought partnership, not deal inspection. Consider:
- Can we reposition our capabilities to better connect to a challenge this customer actually has?
- Can we shift risk from their side to ours in exchange for a longer commitment or expanded scope?
- What has worked in a similar stalled deal elsewhere that we could adapt here?

Use the SCAMMPERR framework (from `references/scammperr-framework.md`) to expand thinking before defaulting to a price concession or standard escalation. The tool generates options — it does not require answering every question.

**Share:** After resolving the deal (or during the next team meeting), document what worked and why. The innovation that unsticks one deal often applies to another.

WHY: Stalled deals fail two ways: the manager gives the rep a prescriptive answer that does not fit the specific situation, or the manager and rep iterate through the same options without generating new ones. Innovating mode is explicitly collaborative because the manager does not know the answer — and pretending otherwise produces bad advice delivered confidently.

### Step 10 — Write coaching-session-plan.md

Produce a written coaching plan the manager can carry into the session.

```markdown
# Coaching Session Plan

**Rep:** [Name]
**Date:** [YYYY-MM-DD]
**Deal anchor:** [Customer name + deal stage]
**Target pillar:** [Teach / Tailor / Take Control]
**Subscale score:** [x/10] — [Strong / Foundation / New territory]

---

## Prepare (manager completes before session)

**Coaching hypothesis:**
[One sentence: what "good" would have looked like in the observed interaction]

**Observation focus for next call:**
[One specific behavior to watch for]

**Prior coaching continuity:**
[What was agreed last session? Did the rep do it?]

---

## Pre-call Questions (Understand step — ask rep to self-diagnose)

[Insert the Appendix A pre-call questions for the target pillar]

---

## Affirm (open the session with)

[One specific genuine observation of a recent strength]

---

## Specify (agreed behavior change)

"In your next call with [Customer], when [moment], you will [action] instead of [current pattern]."

---

## Embed Plan

**Homework before next call:**
[Pre-call planning task using Appendix A questions]

**Observation checkpoint:**
[Date/format: live, recorded, or immediate post-call debrief]

**Post-call debrief agenda:**
[Appendix A post-call questions for the target pillar]

**Next coaching session date:**
[YYYY-MM-DD]

---

## Coaching Log Update

[Note target pillar, specific observable change, debrief date — for next session's Prepare step]
```

## Inputs

| Input | Required | Description |
|---|---|---|
| rep-profile-assessment.md | Yes | From classify-rep-profile — contains subscale scores |
| deal-brief.md | Yes | The live deal used as the coaching anchor |
| call-transcript.md | Optional | Most recent customer interaction for this deal |
| Prior coaching notes | Optional | Continuity from the last PAUSE session |

## Outputs

**Primary artifact:** `coaching-session-plan.md`

Includes: coaching hypothesis, Affirm opener, Appendix A pre-call questions (pillar-specific), Specify statement (one observable change), Embed plan (homework + observation checkpoint + post-call debrief agenda + next session date).

## Key Principles

**1. One pillar per session.** The PAUSE framework produces one observable behavior change. Attempting to address multiple pillars in a single session produces generic coaching and no lasting change. If the rep has two weak pillars, schedule two sessions.

WHY: CEB's research found that coaching effectiveness depends on specificity. The more targeted the feedback, the more likely the rep can act on it in their next interaction. Broad coaching produces awareness without behavior change.

**2. Coaching assumes the manager knows the answer.** If the manager does not know what "good" looks like for the target behavior in the specific deal context, run the innovating mode instead. Coaching from uncertainty produces advice that sounds authoritative but isn't.

WHY: The fundamental distinction between coaching and innovating is epistemic. Coaching is teaching. Innovating is discovering. Using a teaching approach when you are actually discovering together creates false confidence in the wrong direction.

**3. The Specify step must be observable.** "Be more assertive" is not a Specify. "In the meeting, when the customer pushes back on price, stay with the conversation for at least two exchanges before offering an alternative" is a Specify.

WHY: The Embed step cannot work without a concrete behavior to verify. If neither the manager nor the rep can answer "did it happen?" after the next interaction, the coaching session did not produce a Specify.

**4. Affirm is not optional.** The psychological safety created by the Affirm step is a precondition for the rep to hear the Understand and Specify steps without becoming defensive. Skipping it to "save time" destroys the conditions that make the rest of the session work.

WHY: CEB's PAUSE framework separates performance management from coaching development. Performance management creates defensiveness (because there are consequences). Coaching creates openness (because it is safe). Affirm is the mechanism that establishes which mode the rep experiences the session as.

**5. Embed creates the cumulative effect.** A single PAUSE session that ends without a scheduled post-call debrief and the next session date is just a conversation. The behavior change happens across sessions, not within them. Embed is where the ROI of coaching accrues.

WHY: CEB found that the biggest gap in sales manager coaching is not the quality of individual sessions — it is continuity between them. Managers who coach inconsistently produce reps who improve inconsistently. The Embed step institutionalizes continuity.

## Examples

### Example 1 — Teach pillar coaching session (complex SaaS deal)

**Setup:** Rep scored Teach=6, Tailor=8, Take Control=7. Weakest: Teach. Anchor deal: mid-market VP Sales at a SaaS company, discovery call coming up next week for a $180k annual contract.

**Coaching hypothesis (Prepare):** "In the last call, the rep led with product features instead of surfacing a business problem the customer hasn't seen yet. Good Teach behavior would have been opening with the insight that companies at this customer's growth stage typically underinvest in pipeline visibility until it causes a missed quarter — before asking about their current setup."

**Affirm:** "I noticed you came into that last call with detailed knowledge of their competitor's recent pricing move. That preparation shows you were thinking about their world, not just your product."

**Understand (pre-call questions for Teach):**
- What business problem are you planning to surface in the discovery call next week?
- How do you know this is genuinely high-priority for a VP Sales at a company this size?
- Why hasn't this customer solved this problem on their own already?

**Specify:** "In your discovery call next week, you will open with the pipeline visibility problem — before asking a single question about their current setup. If you catch yourself asking questions first, pause and introduce the insight before continuing."

**Embed:** Homework — use the Teach pre-call questions from Appendix A to write a one-page call prep before the discovery call. Debrief within 24 hours using the Teach post-call question: "How provoked or intrigued was the customer when you surfaced the insight? What told you that?"

---

### Example 2 — Innovating mode (stalled deal, unknown obstacle)

**Setup:** Rep's deal with a large enterprise customer has been in legal review for six weeks. No standard objection — the deal just isn't moving. Manager does not know what is blocking it.

**Decision gate:** This is not a rep-behavior gap. Neither the manager nor the rep knows what is stalling the deal. Route to innovating mode.

**Investigate:** Work with the rep to map the customer's internal process. Who is in legal? Who is sponsoring this deal internally at the customer? What does the customer's standard vendor approval process look like — is six weeks normal or abnormal? What did the rep hear most recently from any contact inside the account?

**Create:** Generate options without evaluating them immediately. Could we modify contract terms to reduce the customer's perceived risk? Could we offer a 30-day pilot to get internal momentum before full legal review? Is there an executive sponsor at the customer who could accelerate the legal process — and has the rep engaged them?

**SCAMMPERR prompt:** Under "Adapt" — what have we done to move past a legal stall in a similar deal? Under "Put to other uses" — is there a part of our offering we could scope down to get a faster-path signature while the main contract moves through review?

**Share:** After resolution, document what worked in the internal coaching log for the team.

## References

- [Appendix A Coaching Questions by Challenger Pillar](references/appendix-a-questions.md) — Full pre-call and post-call question sets for Teach, Tailor, and Take Control
- [SCAMMPERR Framework](references/scammperr-framework.md) — Structured deal repositioning tool for the innovating mode

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

- **classify-rep-profile** ← Run this first. Produces rep-profile-assessment.md with the subscale scores this skill reads.
- **diagnose-manager-effectiveness** — For managers who want to assess their own coaching/innovating time allocation before building a coaching practice.
- **diagnose-taking-control-gaps** — Deep-dives into Take Control skill gaps when that subscale is the coaching target.

FILE:references/appendix-a-questions.md
# Appendix A Coaching Questions by Challenger Pillar

Paraphrased from the Challenger Coaching Guide excerpt (Appendix A, _The Challenger Sale_). Use these question sets in the Understand and Embed steps of the PAUSE framework to anchor coaching conversations to specific observable behaviors per pillar.

---

## Teach for Differentiation

**Before the call — preparation questions for the rep:**

1. What business problem are you planning to surface with this customer — and why this one, at this point in the relationship?
2. How do you know this problem is genuinely high-priority for them? What evidence do you have?
3. How have other companies in this customer's situation approached the same problem? What worked and what didn't?
4. How novel will your perspective be to this customer? Is it likely to teach them something they don't already know, or confirm what they already believe?
5. Why hasn't the customer solved this problem on their own already? What has stood in the way?

**After the call — debrief questions:**

- How engaged or provoked did the customer appear when you introduced your perspective? What signals told you that — their words, questions, body language, next-steps energy?
- Was the customer's reaction to your insight what you expected? If not, what surprised you?

**Coaching exercise — building a stronger teaching conversation:**

Pick one account together. Answer these:
- What are the customer's top strategic priorities over the next one to three years?
- Where are they outperforming their competitors? Where are they falling behind?
- How does your contact's specific role connect to those priorities and gaps?
- Where does your company's capability create a genuine advantage that connects to the customer's situation?

Use the answers to help the rep build a teaching conversation that links the customer's competitive position to your solution — without leading with the product.

---

## Tailor for Resonance

**Before the call — preparation questions for the rep:**

1. What industry trends are likely top-of-mind for this customer right now, and how would those trends hit their specific company?
2. What is distinctive about this company's market position — what do they do better or differently than their peers?
3. Where are they most exposed or vulnerable in their competitive environment?
4. Which stakeholders will be in this conversation, and what does success look like from each of their perspectives?

**After the call — debrief questions:**

- What did you learn about this customer's business drivers that you didn't know going in?
- Did you encounter any goals, concerns, or motivations you hadn't anticipated? How did you adapt in the moment?
- Did your message land differently with different stakeholders in the room? Why?

---

## Take Control

**Before the call — preparation questions for the rep:**

1. What specific next step does a successful outcome of this call produce — not a vague "move forward" but a concrete decision, date, or action?
2. What do you know about how this customer's organization actually makes purchase decisions? Who approves? Who blocks? What does their internal process look like?
3. Where in the buying process are you, and what does the customer need to do in their organization to get to the next stage?

**After the call — debrief questions:**

- Did this conversation produce a concrete next step that moves the purchase process forward? What is it, and who owns it?
- When tension arose — around price, timing, consensus, or commitment — what was your instinct: defuse the tension or hold your position? What did you actually do?
- What are your next steps to ensure the deal does not stall between now and the next interaction?

---

## Which Questions to Use

Select the question set that matches the rep's weakest Challenger pillar as identified in `rep-profile-assessment.md`:

| Weakest subscale | Use |
|---|---|
| Teaches for Differentiation (Q2+Q3 < 7) | Teach question set |
| Tailors for Resonance (Q5+Q6 < 7) | Tailor question set |
| Takes Control (Q8+Q9 < 7) | Take Control question set |

If two subscales are equally weak, focus on one per coaching session — do not blend question sets. The Specify step requires the manager to land on ONE observable change, and mixing pillars makes that impossible.

FILE:references/scammperr-framework.md
# SCAMMPERR Framework

A structured brainstorming tool for repositioning a stalled deal without defaulting immediately to a price concession or standard escalation. Adapted from Michal Michalko's creative thinking methodology; applied to deal-level sales innovation in _The Challenger Sale_ (Chapter 8).

Use this in the **Create** step of the innovating mode — after Investigate has surfaced what is actually blocking the deal.

---

## How to Use It

Work through the prompts with the rep. You do not need to answer every row. The framework is a forcing mechanism to expand the universe of options before converging on a solution. Each row opens a different angle on the deal.

| Letter | Lens | Prompting Question |
|--------|------|-------------------|
| **S** | Substitute | What element of our offer or process could we replace with something different? What if we substituted a different stakeholder, a different product scope, or a different commercial structure? |
| **C** | Combine | What could we add to this deal or combine with it to make it more compelling? Could we pair this with a service, a pilot, or a complementary offering? |
| **A** | Adapt | What approach has worked in a similar deal elsewhere that we could adapt to this customer's specific situation? What has another team, region, or rep done to move a similar obstacle? |
| **M** | Modify / Magnify | What could we change about how we are presenting this deal? What if we made one element larger, more prominent, or differently framed? |
| **P** | Put to other uses | Is there a different use case, department, or entry point at this customer that could get a faster decision while the main deal moves through the process? |
| **E** | Eliminate | What if we removed a component that the customer finds risky or complex? Could a smaller scope get an initial signature, with expansion planned separately? |
| **R** | Reverse | What if we changed the sequence? Instead of asking for a decision before showing value, could we show value first? Could we flip who is taking the risk in the transaction? |
| **R** | Rearrange | What if we restructured the entire commercial arrangement — payment terms, duration, stakeholder sequence, or pricing model — to fit how this customer actually makes decisions? |

---

## Applied Example

**Situation:** Enterprise deal stalled because purchasing is resisting a price increase for services the rep has been delivering for two years.

Working through SCAMMPERR:

- **Substitute:** What if we substituted the standard service package with a performance-based contract — customer pays for outcomes, not inputs?
- **Combine:** Could we combine the renewal with a small pilot of our new capability, giving purchasing a reason to approve now in exchange for preferred pricing on the pilot?
- **Adapt:** We've moved past similar pricing objections with [other company] by showing the productivity ROI at the site level. Can we adapt that same business case for this customer's plant manager?
- **Modify:** What if we modified the frequency rather than the price — same annual cost, delivered in smaller monthly increments to match their budget cycles?
- **Eliminate:** What custom packaging are we providing that they do not actively value? Removing it might offset the increase without reducing perceived value.

The rep and manager do not need to use all eight lenses. Two or three strong options discovered through this process are enough to generate a creative path forward — and prevent the default of offering an immediate discount.

---

## When to Use vs. PAUSE

| Situation | Use |
|---|---|
| Rep has a known behavior gap (not executing Teach, Tailor, or Take Control consistently) | PAUSE coaching |
| Deal is stalled and neither manager nor rep knows why or what will move it | Investigate → Create → Share + SCAMMPERR |
| Rep is new to a pillar and needs skill building | PAUSE coaching |
| The obstacle is deal-specific (legal, budget cycle, consensus, procurement) | Investigate → Create → Share + SCAMMPERR |

Key distinction: SCAMMPERR is for deal obstacles, not rep skill gaps. Do not use it to diagnose a rep's Challenger behaviors — that is what the Appendix A pillar questions are for.

---
name: classify-rep-profile
description: "Classify a sales rep against the five CEB selling profiles (Challenger, Hard Worker, Relationship Builder, Lone Wolf, Reactive Problem Solver) and score their Teach/Tailor/Take-Control subscales using the Appendix B self-diagnostic. Use when a rep wants to know 'am I a Challenger', 'which sales profile am I', 'sales style assessment', 'classify sales rep', 'Challenger profile quiz', 'CEB sales profile', 'Teach Tailor Take Control self-diagnostic', 'what kind of rep am I', 'Relationship Builder Hard Worker Lone Wolf Reactive Problem Solver assessment', 'B2B sales rep diagnostic', 'rep profile classification', 'sales enablement assessment'. Applies CEB's empirical study of 6,000+ reps and 44 behavioral attributes. Produces a rep-profile-assessment.md artifact with profile classification, three subscale scores, threshold interpretation, dominant profile flags, and prioritized development recommendations."
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/classify-rep-profile
metadata: {"openclaw":{"emoji":"📊","homepage":"https://github.com/bookforge-ai/bookforge-skills"}}
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors: ["Matthew Dixon", "Brent Adamson"]
    chapters: [2, 3, "Appendix B"]
tags: [sales, b2b-sales, sales-methodology, rep-diagnostic, sales-enablement, challenger-sale]
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - type: document
      description: "Rep's self-description, recent deal notes, or behavioral observations (optional — agent will gather via questions if not provided)"
  tools-required: [Read, Write, AskUserQuestion]
  tools-optional: [Grep]
  mcps-required: []
  environment: "Any working directory. No codebase required."
discovery:
  goal: "Classify a sales rep against the five CEB profiles and score their Teach/Tailor/Take-Control subscales"
  tasks:
    - "Guide rep through the 10-statement Appendix B self-diagnostic"
    - "Compute Teach / Tailor / Take-Control subscale scores"
    - "Classify dominant selling profile and secondary profile flags"
    - "Interpret scores against empirical thresholds"
    - "Produce development recommendations with prioritized subscale focus"
  audience:
    roles: ["B2B sales reps", "sales managers", "sales enablement professionals"]
    experience: intermediate
  when_to_use:
    triggers:
      - "Rep wants to self-assess their selling style"
      - "Manager wants to diagnose a rep's behavioral profile before coaching"
      - "Enablement team wants to baseline the team before a Challenger training initiative"
    prerequisites: []
    not_for:
      - "Hiring assessment or candidate screening (use a separate hiring guide)"
      - "Personality-type profiling (this measures observable behaviors, not traits)"
  environment:
    codebase_required: false
    codebase_helpful: false
    works_offline: true
  quality:
    scores:
      with_skill: null
      baseline: null
      delta: null
    tested_at: null
    eval_count: null
    assertion_count: 12
    iterations_needed: null
---

# Classify Rep Profile

Guides a sales rep (or their manager) through the 10-statement Appendix B self-diagnostic from CEB's research. Computes three Challenger subscale scores — Teach for Differentiation, Tailor for Resonance, Take Control — against empirical thresholds derived from a study of 6,000+ reps across 90 companies. Produces a structured assessment artifact with profile classification, subscale scores, profile flag analysis, and prioritized development recommendations.

The empirical basis: CEB's factor analysis of 44 behavioral attributes collapsed to five selling profiles. Only one profile dominates star performance (~40% of all high performers; >50% in complex solution sales): the Challenger.

## When to Use

- A rep asks "am I a Challenger?", "what's my selling style?", or "which sales profile fits me?"
- A manager wants a structured baseline before starting Challenger coaching with a rep
- A sales enablement team wants to assess where reps cluster before a training rollout
- A rep received feedback that they're "too accommodating" or "not pushing enough" and wants to understand why
- A team lead wants to check whether their perceived high performers are actually Challengers before scaling their behaviors

**Not for:** candidate screening, personality assessment, or competency-based performance reviews.

## Context & Input Gathering

Before running the diagnostic, collect two pieces of context:

**1. Deal complexity context** — Ask: "Are you primarily selling complex, bundled solutions with long sales cycles and multiple stakeholders, or simpler transactional products with short cycles?"

WHY: The Challenger profile dominates complex sales. In transactional/simple sales, the Hard Worker profile performs comparably. Context determines which profile classification carries the most urgency.

**2. Existing self-description (optional)** — Ask: "Is there a deal log, coaching note, or brief description of how you typically sell that I can read before we run the diagnostic?"

WHY: Pre-reading lets the agent validate self-reported scores against observed behaviors and flag potential score inflation.

## Process

### Step 1 — Establish context (2 questions)

Ask the user:
1. "Are your deals primarily complex solution sales (multi-stakeholder, long cycle, bundled offers) or simpler transactional sales?"
2. "Are you completing this for yourself (self-assessment), or are you assessing someone else (coaching/enablement)?"

WHY: The two answers calibrate how to interpret the results — which profile to prioritize developing toward, and whether to phrase questions in first or third person.

If existing deal notes or a coaching log are available, read them before proceeding.

### Step 2 — Administer the diagnostic

Present all 10 statements from `references/self-diagnostic-questionnaire.md`, one at a time (or in a single grouped prompt). For each statement, collect a score from 1 to 5.

WHY: The self-diagnostic was designed by CEB as the validated scoring instrument for the five-profile taxonomy. Each statement maps to a specific behavioral dimension — three map to the Challenger subscales (Q2, Q3, Q5, Q6, Q8, Q9); four flag other profile tendencies (Q1, Q4, Q7, Q10).

Guidance for the agent while presenting:
- Paraphrase each statement clearly in the current conversation context
- Do not hint at which answer is "more Challenger" — that biases results
- If a rep seems unsure, offer a concrete example relevant to their deal context

### Step 3 — Compute subscale scores

Calculate the three Challenger subscale scores:

| Subscale | Statements | Max |
|---|---|---|
| Teaches for Differentiation | Q2 + Q3 | 10 |
| Tailors for Resonance | Q5 + Q6 | 10 |
| Takes Control | Q8 + Q9 | 10 |

Apply empirical thresholds per subscale:
- **8+** → Strong Challenger behavior in this dimension
- **5–7** → Foundation established; targeted development needed
- **4 or below** → New territory; start with the most approachable behavior in this subscale

WHY: The thresholds come directly from CEB's scoring guide (Appendix B). They are not normative relative rankings — they are calibrated against the 6,000-rep study to indicate readiness to employ Challenger behaviors in each dimension.

Also note: profile flags for Q1 (Relationship Builder), Q4 (Lone Wolf), Q7 (Reactive Problem Solver), Q10 (Hard Worker). A score of 4 or 5 on any flag statement signals natural selling tendency outside the Challenger model.

### Step 4 — Classify dominant profile

Use this logic to determine primary profile:

1. If any Challenger subscale scores 8+ → the rep has Challenger capability in that subscale; note which ones
2. If all three subscales score 5–7 or higher → classify as Challenger in development
3. If one or more subscales score 4 or below AND a profile flag scored 4+ → the flag profile is the rep's dominant current profile
4. If multiple flags scored 4–5 → list all as secondary tendencies; the rep is a profile blend

WHY: CEB's model holds that every rep has some presence across all five profiles — the classification identifies the dominant behavioral "major," not a binary label. A rep can be a Challenger with Hard Worker tendencies.

**Complexity overlay:**
- Complex sales: Flag Relationship Builder classification as high-risk (nearly zero star performance in complex deals)
- Transactional sales: Hard Worker classification is viable — don't push Challenger development where deal complexity doesn't warrant it

### Step 5 — Write the assessment artifact

Create `rep-profile-assessment.md` in the working directory (or a specified output path).

WHY: A written artifact makes the assessment portable — it can be shared with a manager, filed in a coaching system, or revisited at the next development checkpoint.

## Inputs

| Input | Required | Description |
|---|---|---|
| Self-assessment responses (Q1–Q10) | Yes | Collected interactively during the diagnostic |
| Deal complexity context | Yes | Complex solution vs. transactional (gathered in Step 1) |
| Self/manager mode | Yes | Determines first/third person framing |
| Deal notes or coaching log | Optional | Used to validate or contextualize self-reported scores |

## Outputs

**Primary artifact:** `rep-profile-assessment.md`

```markdown
# Sales Rep Profile Assessment
**Rep:** [Name or "Self-Assessment"]
**Date:** [YYYY-MM-DD]
**Assessed by:** [Self / Manager: Name]
**Deal complexity context:** [Complex solution / Transactional]

---

## Diagnostic Scores

| Statement | Score (1–5) |
|---|---|
| Q1 — Lasting customer relationships | [score] |
| Q2 — Delivers unique customer perspectives | [score] |
| Q3 — Deep product/service expertise | [score] |
| Q4 — Willing to risk disapproval | [score] |
| Q5 — Adapts message to customer value drivers | [score] |
| Q6 — Identifies customer business drivers | [score] |
| Q7 — Personally resolves customer requests | [score] |
| Q8 — Guides customers to decisions in difficult situations | [score] |
| Q9 — Comfortable discussing pricing/money | [score] |
| Q10 — Extensive pre-call preparation | [score] |

---

## Challenger Subscale Scores

| Subscale | Score | Threshold |
|---|---|---|
| Teaches for Differentiation (Q2+Q3) | [x/10] | [Strong / Foundation / New territory] |
| Tailors for Resonance (Q5+Q6) | [x/10] | [Strong / Foundation / New territory] |
| Takes Control (Q8+Q9) | [x/10] | [Strong / Foundation / New territory] |

---

## Profile Classification

**Dominant profile:** [Challenger / Hard Worker / Relationship Builder / Lone Wolf / Reactive Problem Solver]

**Secondary tendencies:**
- [Profile if flag scored 4+, else "None detected"]

**Performance context:** [Challenger wins in complex sales (>50% of stars). Relationship Builder underperforms in complex sales (nearly zero star rate). Hard Worker is viable in transactional/simple deals.]

---

## Profile Flag Summary

| Flag | Statement | Score | Interpretation |
|---|---|---|---|
| Relationship Builder | Q1 | [score] | [High ≥4: tendency detected / Low: not dominant] |
| Lone Wolf | Q4 | [score] | [High ≥4: tendency detected / Low: not dominant] |
| Reactive Problem Solver | Q7 | [score] | [High ≥4: tendency detected / Low: not dominant] |
| Hard Worker | Q10 | [score] | [High ≥4: tendency detected / Low: not dominant] |

---

## Development Recommendations

**Priority 1 — [Lowest-scoring Challenger subscale]:**
[Specific development focus for this subscale — what to practice in the next 30 days]

**Priority 2 — [Second-lowest subscale]:**
[Specific development focus]

**Priority 3 — [If a profile flag scored high and it creates risk in current deal context]:**
[Specific behavioral shift to address the flag]

**Note on profile classification:** ~40% of star performers are Challengers; only ~7% of stars are Relationship Builders overall, dropping to nearly zero in complex solution sales. This diagnostic measures current behavior, not fixed capacity — all three Challenger subscales are learnable with deliberate practice.

---

## Next Steps

- [ ] Share with manager for validation against observed behaviors
- [ ] Identify one Challenger subscale to develop over the next 30 days
- [ ] Re-take diagnostic in 90 days to measure progress
```

## Key Principles

**1. Behavior, not personality** — The five profiles describe observable selling behaviors, not personality types or fixed traits. Every rep has a behavioral "major" today that can shift with deliberate practice. Never treat the classification as a permanent label.

WHY: CEB specifically designed the study around demonstrated behaviors because skills and behaviors are actionable immediately, whereas personality traits cannot be coached in any practical timeframe.

**2. Challenger advantage is context-dependent** — The Challenger profile dominates complex solution sales. In transactional, high-volume, simple sales, the Hard Worker profile is equally viable. Avoid pushing Challenger development in contexts where deal complexity doesn't warrant it.

WHY: CEB's data shows all five profiles perform roughly equally in simple/transactional deals. The Challenger advantage emerges specifically as deal complexity increases.

**3. High performer ≠ Challenger** — Only ~40% of star performers are Challengers. When managers identify "their Challengers" without using a diagnostic, they typically just select their high performers regardless of selling style. This risks scaling Lone Wolf or Relationship Builder behaviors instead.

WHY: The Appendix B diagnostic exists precisely to solve this problem — it separates profile from performance so organizations scale the right behaviors.

**4. The Relationship Builder risk is real in complex sales** — Relationship Builders represent only 7% of star performers overall, and nearly zero in complex solution sales. A service-oriented, tension-avoidant approach cannot drive the behavioral change complex buying decisions require.

WHY: Complex sales ask customers to change how they operate. A rep who defuses tension and accommodates every demand cannot push the customer to think differently — which is the core commercial mechanism of complex buying.

**5. Self-report needs behavioral validation** — Self-assessment scores are a starting point, not a verdict. High performers sometimes understate their Challenger behaviors (false modesty); Relationship Builders sometimes overstate them (not recognizing the gap). Supplement with manager observations or call-review evidence when the stakes are high.

WHY: CEB's original methodology used manager observations of rep behaviors, not self-reports. Appendix B is a practical approximation — calibrate it against behavioral evidence wherever possible.

## Examples

### Example 1 — Sales rep self-assessment (complex B2B SaaS)

**Scenario:** A mid-market account executive at a SaaS company asked "I always get good reviews from customers but I'm not hitting quota. What profile am I?"

**Trigger:** "What sales profile am I? I think I might be a Relationship Builder."

**Process:** Agent established deal complexity (complex, multi-stakeholder SaaS deals, 60-90 day cycles). Administered the diagnostic — rep scored: Q1=5, Q2=3, Q3=4, Q4=2, Q5=3, Q6=3, Q7=4, Q8=2, Q9=2, Q10=3.

**Output excerpt from rep-profile-assessment.md:**

```
Challenger Subscale Scores:
- Teaches for Differentiation (Q2+Q3): 7 — Foundation
- Tailors for Resonance (Q5+Q6): 6 — Foundation
- Takes Control (Q8+Q9): 4 — New territory

Dominant profile: Relationship Builder (Q1=5 flag; Takes Control at floor)

Priority 1 — Takes Control (score: 4): Practice holding the silence after a proposal.
In the next 3 deals, script one moment where you guide the customer to a decision rather
than waiting. Specifically: become comfortable discussing pricing on the customer's terms.

Note: In complex SaaS deals (your context), Relationship Builder profile correlates with
nearly zero star performance. Your teaching and tailoring foundations are buildable —
but Take Control is the most urgent gap between your current approach and quota attainment.
```

---

### Example 2 — Manager diagnosing a rep before coaching

**Scenario:** A regional sales manager wants to understand her top performer's profile before scaling his approach to the team.

**Trigger:** "My best rep is crushing quota. I want to understand his selling style before I ask him to mentor others."

**Process:** Agent flagged the anti-pattern — high performer ≠ Challenger. Manager provided deal notes from the rep's last three wins. Agent administered diagnostic in third-person ("How would [rep] rate himself on each statement?"). Rep scored: Q1=3, Q2=5, Q3=5, Q4=4, Q8=4, Q9=5 (Lone Wolf flag on Q4; strong Challenger subscales).

**Output excerpt:**

```
Dominant profile: Challenger (all three subscales 8+)
Secondary tendency: Lone Wolf (Q4=4 — willing to risk disapproval)

Safe to scale: Yes — this rep's performance is driven by Challenger behaviors
(Teach/Tailor/Take Control), not Lone Wolf rule-breaking. The Lone Wolf flag
is a secondary tendency, not the primary mechanism of their success.

Recommendation: Observe and codify this rep's commercial teaching conversations
specifically — how they frame unique perspectives and hold tension before closing.
Those are the replicable behaviors.
```

---

### Example 3 — Team enablement baseline

**Scenario:** A sales enablement leader wants to understand where her 12-person team clusters before rolling out a Challenger training program.

**Trigger:** "Can you help me assess what profile mix I have on my team so I know where to focus training?"

**Process:** Agent recommended running the diagnostic for each rep individually (or self-administered in a group session) and aggregating subscale scores. Agent produced a team profile summary template.

**Output excerpt:**

```
Team Profile Summary (12 reps):
- 3 reps: Challenger (all subscales 5+, no dominant profile flags)
- 4 reps: Relationship Builder dominant (Q1=5, Takes Control at 4 or below)
- 2 reps: Hard Worker dominant (Q10=5, transactional deal focus)
- 2 reps: Challenger in development (subscales 5-7, no strong flags)
- 1 rep: Lone Wolf flag (Q4=5, inconsistent subscale scores)

Training priority: Takes Control subscale (team average: 5.2 — lowest).
8 of 12 reps score below 7 on Takes Control. Recommend leading training with
money conversation practice and guided commitment techniques before moving
to teaching content — the team can teach but cannot close the tension.
```

## References

- [Five Sales Rep Profiles](references/five-profiles.md) — Full profile definitions, behavioral signatures, and CEB performance data
- [Self-Diagnostic Questionnaire](references/self-diagnostic-questionnaire.md) — 10-statement instrument, scoring guide, threshold interpretations, and sample score interpretation

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/).

The skill was generated by the [BookForge](https://github.com/bookforge-ai/bookforge) pipeline from _The Challenger Sale_ by Matthew Dixon and Brent Adamson (Portfolio/Penguin, 2011). Content has been paraphrased and structured as an executable skill — it does not reproduce verbatim passages from the copyrighted work. Attribution required on redistribution.

## Related BookForge Skills

This skill has no dependencies. It is the foundational classification layer for the Challenger Sale skill set — other skills in this set assume the rep has already completed a profile assessment.

FILE:references/five-profiles.md
# Five Sales Rep Profiles — Reference

Source: The Challenger Sale (Dixon & Adamson), Chapter 2
Empirical basis: CEB factor analysis of 44 behavioral attributes across 6,000+ sales reps in 90 companies. Managers assessed reps on performance against quota; star performers = top 20%.

---

## Overview

Factor analysis of 44 rep attributes produced five statistically distinct behavioral clusters. These are not personality types — they describe observable selling behaviors. Each profile represents a "major" in how reps engage with customers. All reps show some behavior across all profiles, but one profile typically dominates their approach.

Core performers distribute roughly evenly across all five profiles. Star performers do not — the Challenger profile dominates.

---

## Profile 1: Hard Worker

**Core behavioral signature:** High call volume, self-motivated, process-adherent, persistent.

Key behaviors:
- Shows up early, stays late, puts in extra effort
- Maintains high activity — more calls per hour, more visits per week than peers
- Follows sales process with discipline
- Frequently seeks feedback to improve
- Believes consistent right actions produce results over time

**Performance context:** Competes strongly in transactional, high-volume, low-complexity deals. In simple sales where call quality matters less than call quantity, Hard Workers win the day. In complex solution sales, their activity-volume approach does not translate to star performance.

**Profile flag statement:** "I am likely to spend more time in preparation in advance of any sales calls or meetings than everybody else." (Diagnostic Q10)

---

## Profile 2: Relationship Builder

**Core behavioral signature:** Service-oriented, tension-avoidant, accessibility-first, accommodating.

Key behaviors:
- Builds and nurtures strong personal and professional relationships across the customer organization
- Generous with time; posture of "whatever you need"
- Works hard to ensure customer needs are met in the moment
- Avoids difficult conversations to preserve harmony
- Acquiesces to customer demands; won't push on next steps
- Focuses on customer convenience rather than customer value
- Relies on finding business through existing relationships rather than creating new opportunities

**Performance data:**
- 7% of all star performers (lowest of any profile)
- In complex solution sales: likelihood of star status falls to nearly zero

**Why the profile fails in complex sales:** Complex sales require customers to change behavior and think differently about their business. A service-oriented relationship without commercial push cannot drive that change. Familiarity with a customer is not enough when the buyer needs insight to justify a disruptive purchase. "A service-oriented quarterly check-in call is a great way to find business, but not a very good way to make business."

**Profile flag statement:** "I often form enduring and useful relationships with customers." (Diagnostic Q1)

---

## Profile 3: Lone Wolf

**Core behavioral signature:** Instinct-driven, self-confident, process-averse, high individual performance.

Key behaviors:
- Deep self-confidence; follows own instincts over established process
- "Prima donna" of the sales force — does things their way or not at all
- No CRM entries, no trip reports, no process compliance
- Achieves strong individual results despite — or because of — rule-breaking
- Non-replicable approach by design

**Performance data:**
- ~25% of high performers overall (high individual probability of star status)
- Competes in complex sales, but is "hard to find and even harder to control"

**Why not scalable:** Lone Wolf behavior succeeds precisely because it breaks rules. You cannot teach systematic rule-breaking. Lone Wolves self-select out of coaching and development. Managers cannot observe and codify their methods.

**Anti-pattern warning:** Because Lone Wolves score disproportionately high among star performers, managers often mistake them for Challengers. The diagnostic separates these: Lone Wolves score high on risk-taking (Q4) but not necessarily on the three Challenger subscales.

**Profile flag statement:** "I often risk disapproval in order to express beliefs about what is right for the customer." (Diagnostic Q4)

---

## Profile 4: Reactive Problem Solver

**Core behavioral signature:** Detail-oriented, post-sale focused, reliable, reactive to existing customer needs.

Key behaviors:
- Highly reliable; ensures all post-sale promises are fulfilled
- Detail-oriented; thorough follow-up on implementation and execution issues
- Responds to existing customer problems immediately — even at the expense of generating new business
- Often described as "a customer service rep in sales rep clothing"
- Arrives with plans to generate new sales; leaves having solved existing customer problems

**Performance context:** Reliability and follow-through matter in all sales contexts, but the Problem Solver's reactive posture prevents proactive pipeline development. In complex sales requiring prospecting and insight delivery, this profile under-indexes.

**Profile flag statement:** "When it comes to fulfilling customer requests, I usually resolve everything myself." (Diagnostic Q7)

---

## Profile 5: Challenger

**Core behavioral signature:** Insight-driven, constructively assertive, customer-value focused, comfortable with tension.

Key behaviors:
- Deep understanding of customer business; uses it to push customer thinking
- Teaches customers something new and valuable about how to compete
- Shares views even when different or controversial
- Assertive on both customer thinking and commercial terms (pricing, next steps)
- Creates constructive tension rather than defusing it
- Focuses on customer value, not customer convenience
- Challenges own managers and internal stakeholders, not just customers

**Six statistically significant defining attributes (from the 44-attribute study):**
1. Offers unique perspectives to customers
2. Has strong two-way communication skills
3. Knows the individual customer's value drivers
4. Can identify economic drivers of the customer's business
5. Comfortable discussing money (pricing, reimbursement)
6. Can pressure the customer appropriately

**Three capability clusters:**
- **Teach for Differentiation** — Attributes 1 + 2: delivers unique insight and engages in two-way dialogue
- **Tailor for Resonance** — Attributes 3 + 4: customizes message to individual value and economic drivers
- **Take Control** — Attributes 5 + 6: comfortable with money conversations and willing to push for decisions

**Performance data:**
- ~40% of all star performers overall
- >50% of star performers in complex solution sales
- The only profile that wins consistently as deal complexity increases

**Contrast with Relationship Builder:**

| Dimension | Challenger | Relationship Builder |
|---|---|---|
| Primary focus | Customer value | Customer convenience |
| Tension posture | Creates constructive tension | Defuses/avoids tension |
| Customer comfort zone | Pushes customer out of it | Seeks acceptance into it |
| Business generation | Makes business (proactive) | Finds business (reactive) |

---

## Complexity Decision Rule

| Sales Environment | Winning Profile |
|---|---|
| Complex, solution, multi-stakeholder, long cycle | Challenger |
| Simple, transactional, high-volume, short cycle | Hard Worker |
| Any complexity | Avoid over-indexing on Relationship Builders |

---

## Anti-Pattern: High Performers Are Not All Challengers

Only ~40% of high performers are Challengers. When managers are asked to "identify their Challengers," they typically nominate their high performers regardless of selling style. This risks observing and scaling the behaviors of high-performing Lone Wolves or Relationship Builders instead.

**Correct sequence:** Run the diagnostic first. Confirm selling style. Then observe and codify Challenger behaviors.

FILE:references/self-diagnostic-questionnaire.md
# Selling Style Self-Diagnostic — Reference

Source: The Challenger Sale (Dixon & Adamson), Appendix B
Purpose: A 10-statement self-assessment that scores a rep's Challenger capability across three subscales (Teach / Tailor / Take Control) and flags tendencies toward the other four selling profiles.

---

## Instructions

For each statement below, rate how accurately it describes how you sell to your customers, using this scale:

| Score | Meaning |
|---|---|
| 1 | Strongly disagree |
| 2 | Disagree |
| 3 | Neutral |
| 4 | Agree |
| 5 | Strongly agree |

---

## Statements

**Statement 1** (Profile flag: Relationship Builder)
Rate your agreement with the idea that you regularly form lasting, productive relationships with your customers — relationships that persist beyond individual transactions.

**Statement 2** (Challenger subscale: Teaches for Differentiation)
Rate your agreement with the idea that you are able to deliver genuinely new perspectives to customers — insights that shift how they think about their business and naturally connect to what you offer.

**Statement 3** (Challenger subscale: Teaches for Differentiation)
Rate your agreement with the idea that your product and industry knowledge goes well beyond what even an experienced buyer would possess, making you a credible expert in the sales conversation.

**Statement 4** (Profile flag: Lone Wolf)
Rate your agreement with the idea that you are willing to express your honest assessment of what is right for a customer even when doing so risks their disapproval.

**Statement 5** (Challenger subscale: Tailors for Resonance)
Rate your agreement with the idea that during negotiations you have a clear sense of what each customer values most, and you adapt your message and offer accordingly rather than delivering a standard pitch.

**Statement 6** (Challenger subscale: Tailors for Resonance)
Rate your agreement with the idea that you can identify the underlying business drivers — financial, operational, strategic — of your customers' organizations and use that understanding to shape your approach.

**Statement 7** (Profile flag: Reactive Problem Solver)
Rate your agreement with the idea that when a customer makes a request, you are the one who personally ensures it gets resolved, end-to-end, rather than delegating to others.

**Statement 8** (Challenger subscale: Takes Control)
Rate your agreement with the idea that in challenging or stalled sales situations, you feel at ease guiding the customer toward a commitment and decision rather than waiting for them to move on their own.

**Statement 9** (Challenger subscale: Takes Control)
Rate your agreement with the idea that you can engage comfortably and confidently in conversations about price, budget, and financial concerns with your customers, meeting them on their own terms.

**Statement 10** (Profile flag: Hard Worker)
Rate your agreement with the idea that you invest more time than your peers in preparation before every sales call or customer meeting.

---

## Scoring Guide

### Challenger Subscale Scores

Calculate three subscale totals. Each subscale ranges from 2 (minimum) to 10 (maximum).

| Subscale | Statements to Add |
|---|---|
| **Teaches for Differentiation** | Statement 2 + Statement 3 |
| **Tailors for Resonance** | Statement 5 + Statement 6 |
| **Takes Control** | Statement 8 + Statement 9 |

### Threshold Interpretation (per subscale)

| Score | Interpretation |
|---|---|
| **8 or above** | Strong Challenger behavior in this area. Continue pushing for ways to challenge customers' thinking in this dimension. |
| **5 to 7** | Solid foundation to build from. Identify a specific development focus within this subscale and begin pushing harder. |
| **4 or below** | This dimension represents new territory. Identify the behavior within it that feels most approachable and start development there. |

---

## Profile Flag Interpretation

A high score (4 or 5) on Statements 1, 4, 7, or 10 signals a natural tendency toward one of the other four selling profiles. These are not disqualifiers — they identify areas where a rep has strong instincts outside the Challenger model.

| Statement | High score signals |
|---|---|
| Statement 1 | Relationship Builder tendencies |
| Statement 4 | Lone Wolf tendencies |
| Statement 7 | Reactive Problem Solver tendencies |
| Statement 10 | Hard Worker tendencies |

**Note:** A rep can show Challenger subscale strength AND a profile flag simultaneously. The goal is to develop all three Challenger subscales to 8+ while being aware of which non-Challenger behaviors are most natural.

---

## Sample Score Interpretation

**Example Rep A:**
- Teaches for Differentiation: Q2=4, Q3=5 → Score **9** (Strong)
- Tailors for Resonance: Q5=3, Q6=4 → Score **7** (Foundation — develop further)
- Takes Control: Q8=2, Q9=3 → Score **5** (Foundation — develop further)
- Flag: Q1=5 (Relationship Builder tendency — may be defusing tension instead of holding it)

**Interpretation:** Rep A has strong teaching ability but is likely accommodating customers rather than pressing for decisions. Development priority: Take Control behaviors — specifically, comfort with money conversations and willingness to guide customers to commitments.

---

## Usage Notes

- This is a self-report instrument. Self-assessments should be supplemented with manager observation or call review to validate.
- Do not use this diagnostic to label reps as "not a Challenger" permanently — it identifies current behavior patterns, not fixed capacity.
- When using for team profiling, aggregate subscale scores across the team to identify organization-level development priorities.
- High-performing reps are not automatically Challengers. Run the diagnostic before observing behaviors to codify — only ~40% of star performers are Challengers.

---
name: build-commercial-insight
description: |
  Reverse-engineer a Commercial Insight from seller strengths and validate it against the four-criteria test.

  Trigger this skill when you need to:
  - Build a commercial insight or reframe for a teaching pitch
  - Figure out "what insight should I teach" in a B2B sales conversation
  - Find a unique angle for a commercial teaching pitch
  - Reverse engineer a pitch from your strengths (the "Deb Oler question")
  - Identify which insight to lead with in an insight-led selling approach
  - Differentiate your pitch from competitors who sell the same category
  - Determine whether a proposed reframe qualifies as commercial teaching
  - Start a sales narrative from your solution's unique differentiator rather than customer pain
  - Create an insight that "leads to" your strengths rather than "leading with" them

  NOT for: building the full 6-step pitch choreography (use author-commercial-teaching-pitch), tailoring to specific stakeholders (use tailor-pitch-by-stakeholder), or scoring pitch boldness (use SAFE-BOLD framework).
version: 1.0.0
homepage: https://github.com/bookforge-ai/bookforge-skills/tree/main/books/the-challenger-sale/skills/build-commercial-insight
metadata:
  openclaw:
    emoji: "💡"
    homepage: "https://github.com/bookforge-ai/bookforge-skills"
status: draft
source-books:
  - id: the-challenger-sale
    title: "The Challenger Sale"
    authors:
      - Matthew Dixon
      - Brent Adamson
    chapters:
      - 4
      - 5
tags:
  - sales
  - b2b-sales
  - commercial-teaching
  - sales-messaging
  - differentiation
  - challenger-sale
depends-on: []
execution:
  tier: 1
  mode: hybrid
  inputs:
    - solution-capabilities.md — what your solution does, what competitors cannot match
    - target-customer-segment — industry, role, size, shared behaviors of the segment
  tools-required:
    - Read
    - Write
    - AskUserQuestion
  works-offline: true
discovery:
  goal: "Reverse-engineer a Commercial Insight from the seller's unique strengths, express it as a validated Reframe, and output a commercial-insight.md artifact ready to use as the foundation for a Commercial Teaching pitch."
  triggers:
    - "What insight should I teach in my next sales call?"
    - "I need a reframe for my pitch — where do I start?"
    - "How do I build an insight-led pitch from my product's strengths?"
    - "I want to stop leading with features — how do I lead to them instead?"
    - "Deb Oler question: why should customers buy from us?"
    - "Is my teaching insight actually differentiating?"
    - "What's costing my customers money that only we can fix?"
  audience:
    - Sales reps and account executives building teaching-led pitches
    - Sales enablement teams building repeatable insight narratives
    - Product marketing managers creating differentiated messaging
    - Sales managers coaching reps on insight-led conversation design
  prerequisites: none
  not-for:
    - Building the full 6-step teaching pitch — use author-commercial-teaching-pitch
    - Tailoring the pitch to specific stakeholder roles — use tailor-pitch-by-stakeholder
    - Scoring pitch boldness against the SAFE-BOLD continuum
---

# build-commercial-insight

Reverse-engineer a Commercial Insight from the seller's unique strengths and validate it against the four-criteria test. Outputs a `commercial-insight.md` artifact ready to anchor a Commercial Teaching pitch.

---

## When to Use

Use this skill when you need to identify *what to teach* before building a teaching pitch. This is the upstream step: you cannot build a strong teaching conversation without a validated insight that:

1. Routes naturally to what you do better than competitors
2. Genuinely reframes how the customer thinks (vs. confirming what they already know)
3. Creates urgency to act
4. Works across a segment — not just one deal

If you already have a validated insight and want to structure the 6-step delivery, move to `author-commercial-teaching-pitch`.

---

## Context and Input Gathering

Before starting, collect or ask for:

1. **Solution capabilities:** What does your solution do that competitors cannot or do not do at the same level? Be specific — avoid "innovative," "customer-focused," and "solutions-oriented." Those are not differentiators.
2. **Target customer segment:** What type of customer (role, industry, company size, common behavior pattern) is this insight for?
3. **What you already know about this segment's worldview:** What do they believe today about the problem area you operate in? What do they under-value or under-appreciate?

If the user has not provided a `solution-capabilities.md` file, use `AskUserQuestion` to collect the three inputs above before proceeding.

---

## Process

### Step 1 — Inventory Unique Strengths

List the specific capabilities your solution provides that competitors cannot match. For each capability, ask:

- Can a competitor offer this too, even partially?
- Would a customer lose this capability if they chose a competitor?

Keep only capabilities where the honest answer is "no, competitors cannot match this."

**Why:** The Deb Oler Question — "Why should our customers buy from us over anyone else?" — has a specific answer that most companies struggle to state clearly. Before you can identify what insight to teach, you must know where your solution actually outperforms. Without this, any insight you identify risks becoming free consulting for competitors.

Capture the result as a short list: **[Capability] → [Why competitors cannot match it]**.

### Step 2 — Focus on Under-Appreciated Capabilities

From your capability list, identify the ones customers *currently under-value* — not the ones they already know and appreciate.

Ask: "If a customer already values this capability, do they need to be taught anything about it?" If no, that capability is not the right anchor. The teaching opportunity lives specifically at the gap between the value you could deliver and the value customers currently perceive.

**Why:** Teaching is only necessary where perception lags reality. If customers already value a capability, they will buy on it without teaching. The Commercial Insight lives in the gap — the hidden value customers are missing because their worldview prevents them from seeing it.

### Step 3 — Apply the Reverse-Engineering Diagnostic

For each under-appreciated capability identified in Step 2, work backward using this diagnostic sequence:

```
Your Solution (unique capability)
    ↓
New Way (what the customer would need to do differently to capture that value)
    ↓
Emotional Impact (a peer company story where the customer sees themselves)
    ↓
Rational Drowning (quantified cost of NOT doing the New Way — in customer's own metrics)
    ↓
Reframe Candidate (the altered perspective that makes the cost suddenly visible)
```

The core diagnostic question at each step:
> "Why don't my customers value this capability already?"

That question uncovers the worldview or assumption blocking appreciation. The Reframe is the correction to that worldview.

**Why:** Building a teaching pitch forward (starting from the customer's pain) often produces generic insights that competitors can copy. Building backward from your unique capability ensures the insight naturally leads back to what only you can deliver. This is "leading to" your strengths rather than "leading with" them.

### Step 4 — Write the Reframe Candidate

Draft a one-to-two sentence Reframe statement that:
- Describes a problem or opportunity the customer has not fully seen
- Connects to a cost, risk, or missed revenue expressible in their own business metrics
- Does not mention your solution or company name

Target reaction from the customer: **"Huh, I never thought of it that way before."**

Disqualifying reaction: "I totally agree!" — enthusiastic agreement means the customer already believed this. It is not a Reframe; it is confirmation of existing beliefs.

**Why:** A Reframe that produces enthusiastic agreement is not a teaching insight — it is validation of the customer's current worldview. The commercial teaching model is built on the premise that the insight must genuinely shift perspective, not reinforce it.

### Step 5 — Validate Against the Four-Criteria Test

Score the Reframe Candidate against each rule. All four must pass before proceeding.

**Rule 1 — Leads to Unique Strengths**
Gate question: After teaching this insight, can you say: "Let me show you why we're better able to help you act on this than anyone else"?

- PASS: The insight naturally requires a capability only your solution provides.
- FAIL (Teaching into the Desert): The insight is compelling but does not require your unique capabilities. Customers can fix the problem with a generic solution or a competitor's product. If you teach this insight and the customer goes to bid, a competitor can win using your free consulting.

**Rule 2 — Challenges Assumptions**
Gate question: Does this insight tell customers something they did not already believe?

- PASS: The insight contradicts or significantly extends the customer's current worldview.
- FAIL: The customer already has this belief on their radar. Enthusiastic agreement is a failure signal, not a success signal.

**Rule 3 — Catalyzes Action**
Gate question: Does this insight create sufficient urgency that the customer will feel compelled to act, not just agree?

- PASS: The insight quantifies a cost, risk, or missed revenue large enough to make inaction uncomfortable.
- FAIL: The customer finds the insight interesting but not urgent enough to change behavior.

**Rule 4 — Scales Across Customers**
Gate question: Is this insight applicable across a recognizable segment (by role, industry, or behavioral pattern) — or does it require full customization for each individual deal?

- PASS: The insight can be pre-built by marketing and delivered consistently across multiple customers with similar characteristics.
- FAIL: The insight requires bespoke research for each customer. One-off insights do not justify the construction investment and create a free consulting risk.

If any rule fails, return to Step 3 and adjust the Reframe candidate. Common fixes:
- Rule 1 failure → check if insight is too generic; narrow to a problem only your specific capability can resolve
- Rule 2 failure → the insight is already common knowledge; go deeper or find a less obvious angle
- Rule 3 failure → the Rational Drowning data layer is missing or weak; find the quantified cost
- Rule 4 failure → the insight is too deal-specific; abstract to the pattern that holds across the segment

### Step 6 — Write the Commercial Insight Statement

Once the Reframe passes all four rules, synthesize a Commercial Insight statement in this format:

> "Most [customer segment] are unknowingly [cost/risk/missed revenue description] because of [flawed assumption or behavior]. [Your solution type] can address this in a way [alternative approaches] cannot."

This becomes the anchor for the teaching pitch. The full Reframe, Rational Drowning data, and Emotional Impact narrative are built around it.

### Step 7 — Write the commercial-insight.md Artifact

Create `commercial-insight.md` with the following structure:

```markdown
# Commercial Insight: [Working Title]

## The Deb Oler Answer
[One paragraph: what your solution does that competitors cannot. The honest, specific answer to "why should customers buy from us over anyone else?"]

## Under-Appreciated Capability
[Which specific capability is currently under-valued, and why customers do not currently value it]

## The Blocking Worldview
[What assumption or belief prevents customers from appreciating this capability today]

## The Reframe
[The corrected perspective — 1-2 sentences. Target reaction: "I never thought of it that way."]

## Four-Criteria Validation
- Rule 1 (Leads to unique strength): [PASS/FAIL + evidence]
- Rule 2 (Challenges assumptions): [PASS/FAIL + evidence]
- Rule 3 (Catalyzes action): [PASS/FAIL + quantified cost/risk]
- Rule 4 (Scales across customers): [PASS/FAIL + segment definition]

## Commercial Insight Statement
[The synthesized 1-3 sentence insight statement ready to anchor the teaching pitch]

## Target Segment
[Role, industry, size, and behavioral characteristics of the customer segment this insight is built for]

## Next Step
Build the 6-step teaching pitch using: author-commercial-teaching-pitch
```

---

## Inputs

| Input | Required | Description |
|-------|----------|-------------|
| `solution-capabilities.md` | Yes | Your solution's specific capabilities and how they compare to competitors |
| Target customer segment | Yes | Industry, role, company size, common behaviors or challenges |
| Existing customer research or purchase data | Optional | Accelerates Step 3 by providing quantified cost/risk evidence |
| Draft pitch or messaging | Optional | Existing materials to audit against the four-criteria test |

---

## Outputs

| Output | Path | Description |
|--------|------|-------------|
| `commercial-insight.md` | `./commercial-insight.md` | The validated Reframe and insight statement ready for pitch construction |

---

## Key Principles

**Start from strength, not pain.** Discovery-led selling starts from the customer's stated pain and maps to your capabilities. Commercial Teaching reverses this: start from your unique capabilities and work backward to the insight that makes customers value them. This ensures the insight is differentiating, not generic.

**Teach what customers do not know, not what they want to hear.** The target reaction is thoughtful pause and curiosity — not enthusiastic agreement. If the customer already believed the insight, no teaching occurred.

**"Leading to" vs. "leading with."** Your solution should appear at the end of the teaching conversation as the natural answer to a problem the customer now believes is urgent — not at the beginning as the subject of the pitch. Your company is mentioned last, not first.

**Only 14% of claimed unique benefits are perceived as both unique and relevant by customers.** Most differentiation claims are not actually differentiating. The four-criteria test exists to verify — not assume — that an insight is genuinely differentiated.

**Teaching into the desert is worse than not teaching at all.** A compelling insight that does not lead to your unique capabilities creates urgency without a solution. The customer goes to bid. Competitors win using your free consulting investment.

---

## Examples

See `references/worked-examples.md` for two detailed worked examples:
- Industrial distributor reverse-engineering a "hidden MRO spend" insight from their catalog breadth advantage
- Enterprise software provider reverse-engineering a "fragmentation cost" insight from their integration advantage

Both examples trace the full reverse-engineering sequence from unique capability through validated Reframe, including four-criteria validation and emotional impact narrative construction.

---

## References

- `references/worked-examples.md` — Two full reverse-engineering walkthroughs

**Source:** *The Challenger Sale* by Matthew Dixon and Brent Adamson. Chapters 4 and 5 (Teaching for Differentiation, Parts 1 and 2).

---

## License

This skill is licensed under [CC-BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). You are free to use, adapt, and redistribute with attribution.

Source book content is copyrighted by the authors. This skill contains no verbatim passages — all content is paraphrased, structured, and extended for agent execution.

---

## Related BookForge Skills

| Skill | Relationship |
|-------|-------------|
| `author-commercial-teaching-pitch` | Use after this skill — builds the 6-step delivery choreography around the validated Commercial Insight |
| `tailor-pitch-by-stakeholder` | Use after `author-commercial-teaching-pitch` — tailors the pitch to individual stakeholder roles |
| `classify-rep-profile` | Use before this skill — confirms whether the rep is operating in Challenger mode |

FILE:references/worked-examples.md
# Worked Examples: Commercial Insight Construction

Two canonical examples from *The Challenger Sale* showing how sellers reverse-engineered a Commercial Insight from their unique strengths.

---

## Example 1: Industrial Distributor — Managing Hidden Spend

**Seller profile:** A broad-line industrial supplies distributor covering hundreds of thousands of products across maintenance, repair, and operations (MRO) categories. Competitors exist but none match the catalog breadth or cross-category planning capabilities.

**The positioning problem:** Customers treated the distributor as a transactional supplier — calling when they needed a specific part, then negotiating on price at contract renewal. The seller's strategic advisory value was invisible.

**Starting question (Deb Oler Question):** "Why should customers buy from us instead of a narrower competitor?"
Answer: Only this seller can manage the full MRO portfolio across all categories with integrated planning tools. No point-solution competitor can do this.

**What customers currently under-appreciate:** The total cost and hidden waste inside reactive, unplanned purchasing. Customers view individual unplanned purchases as minor nuisances — not as a systemic budget problem worth fixing.

**Why they under-appreciate it:** They categorize MRO as low-value facilities spend, not as strategic procurement. No one owns the problem cross-functionally.

**Reverse-engineered Reframe:** "The way most companies handle unplanned equipment purchases creates a cascading cost pattern that consumes significant budget they can't see on any single invoice."

**Validation check:**
- Rule 1 (leads to unique strength): Only the broad-portfolio seller with cross-category planning tools can solve this — a point-solution vendor cannot.
- Rule 2 (challenges assumptions): Customers think individual reactive buys are just "how things work." The reframe reveals a fixable systemic problem.
- Rule 3 (catalyzes action): Once customers see the aggregated cost (expressed in their own MRO spend data), the urgency to act is high.
- Rule 4 (scales): The unplanned purchasing pattern is role-predictable — facility managers at mid-to-large manufacturers share this behavior regardless of geography or industry vertical.

**Emotional impact narrative constructed:** A scenario tracing the chain of events when a critical piece of equipment fails unexpectedly — the multi-hour search across suppliers, workers pulled off productive tasks, emergency transport costs, and then over-buying out of fear of recurrence, leaving unused inventory tied up in storage. The narrative is built from aggregated real customer behavior, not hypothetical. Customers recognize themselves immediately.

**Output — the Commercial Insight statement:**
"Most companies are unknowingly burning a significant portion of their MRO budget on reactive buying patterns — and no current supplier structure they use can fix it. We can."

---

## Example 2: Enterprise Software Platform — The Fragmentation Cost

**Seller profile:** An integrated software platform serving a specific industry vertical (automotive dealerships). The seller's unique strength is cross-functional coverage — every major operational area (sales, service, finance, parts, marketing) on one system. Small competitors serve only one department.

**The positioning problem:** In a market downturn, dealerships were cutting software costs by selecting point solutions from smaller vendors at lower per-module prices. The seller's reps were countering with capability comparisons — and losing on price.

**Starting question (Deb Oler Question):** "Why should customers buy from us instead of assembling point solutions from cheaper competitors?"
Answer: Only the integrated platform eliminates the hidden cost of running disconnected systems — data re-entry, reconciliation errors, redundant licenses, training overhead across vendors.

**What customers currently under-appreciate:** The true operational cost of fragmented software. Customers see the per-module price of each point solution and compare it to the integrated platform's total price — the integrated system looks more expensive.

**Why they under-appreciate it:** The costs of fragmentation are distributed across departments and invisible on any single invoice. No one has aggregated them.

**Reverse-engineered Reframe:** "The money you're saving by buying cheaper, narrower software is actually being consumed — and then some — by the operational overhead those disconnected systems create."

**Validation check:**
- Rule 1 (leads to unique strength): Only the integrated platform eliminates the fragmentation cost — a point-solution vendor cannot solve this by definition.
- Rule 2 (challenges assumptions): Customers believe they are cutting costs. The reframe reveals they are creating costs.
- Rule 3 (catalyzes action): Quantifying the redundancy cost (up to 40% of total software spend attributable to overlap and re-work) creates urgency. The number is large enough to justify immediate review.
- Rule 4 (scales): The multi-vendor fragmentation pattern is common across similarly sized operators in the vertical. The insight can be pre-built for the segment and delivered consistently.

**Delivery format:** Free educational seminars structured as Commercial Teaching choreography. No mention of the seller for the first two-thirds of each session. The Reframe, Rational Drowning data, and Emotional Impact are about the customer's problem — not the seller's capabilities. The seller appears only at the solution stage, as the natural answer to a problem the customer now sees clearly.

**Output — the Commercial Insight statement:**
"The software savings you believe you're achieving by using multiple specialized vendors are likely being erased — and exceeded — by the hidden integration and redundancy costs those vendor relationships create. We can quantify this for your business specifically."

**Measured result:** During an industry downturn when the addressable market contracted sharply, the seller held revenue within a few percentage points of prior year while competitors lost significant share. Customers cited the educational seminars as the primary reason for renewed trust.

---

## Construction Pattern Across Both Examples

| Step | Industrial Distributor | Enterprise Software |
|------|----------------------|-------------------|
| Unique strength | Cross-category MRO planning breadth | Cross-functional integration |
| Under-appreciated benefit | Portfolio cost management | Single-system efficiency |
| Blocking worldview | "MRO is low-priority facilities spend" | "Cheaper per module = lower cost" |
| Reframe | Reactive buying is a hidden P&L leak | Fragmentation costs more than the savings |
| Rational Drowning data | Customer's own MRO spend data | Quantified redundancy cost |
| Emotional Impact hook | Equipment failure cascade narrative | Recognition of operational friction |
| Scales to segment | Facility managers at mid-to-large manufacturers | Operators in the target vertical of similar size |
| Delivery format | Sales call with 6-slide deck | Free educational seminar |

Both examples share the same underlying logic: start from what only you can do, find the worldview gap that prevents customers from valuing it, build a Reframe that crosses that gap, then validate against all four rules before building the pitch.