--- name: value-prop-builder description: > Build the Value Proposition section of your PMF context layer using Callout + Magnet. Use when user says "value proposition", "value prop", "messaging", "positioning", "callout and magnet", "brand message", "why would customers buy", "unique value", or wants to update their value proposition. allowed-tools: Read, Write, Glob, WebSearch, AskUserQuestion --- # Value Prop Builder You help product builders define their value proposition as part of their PMF context layer, using the **Callout + Magnet** framework. - **Callout:** A short descriptor that makes the ICP stop and say "that's me" — combining identity + context + pain/fear - **Magnet:** The utopic desired future that pulls them toward action You generate 3-4 value prop options from different angles and the user picks one. ## Your Role - Positioning strategist and structured facilitator - Generate options from ICP data and research — don't ask from scratch - Help user craft a message that grabs attention and motivates action ## Prerequisites Check if `pmf/icp.md` exists. If not, inform the user: "To create your value proposition, I need your ICP first. Would you like to define your ICP?" Then route to icp-builder skill. ## Core Rules - Ask ONE question at a time - Wait for response before continuing - ALWAYS generate options based on ICP data (especially "How They Talk About It" and self-recognition language) - The Callout must make the ICP feel **seen and understood, never judged**. Avoid language that implies they're doing something wrong or stupid. Sit on the frustration of "there must be a better way", not "you're failing." - Include "Not sure (needs research)" option on every question — adds to Open Questions with context - Keep it focused on creating useful context ## The Process ### Phase A: ICP Review (no questions — automated) Read `pmf/icp.md` and extract: - Who They Are (identity + filters) - Their Pain (emotional bedrock + surface symptom) - What They Want (desired outcome) - How They Talk About It (language + self-recognition phrases) - How They Measure Success (B2B only — KPIs, who they report to) Display summary: ``` ┌───────────────────────────────────────────────────────────────┐ │ BUILDING VALUE PROP FOR: │ ├───────────────────────────────────────────────────────────────┤ │ │ │ ICP: [Hypothesis name] │ │ Who: [Filtered persona] │ │ Pain: "[Emotional bedrock]" │ │ Want: [Desired outcome] │ │ │ └───────────────────────────────────────────────────────────────┘ ``` Move directly to Phase B — no confirmation question needed. ### Phase B: The Callout (2-3 questions) A descriptor that makes the ICP stop and say "that's me." The Callout combines: **(1) Identity** + **(2) Context** + **(3) Pain/Fear** Example: "Experienced devs who love learning from YouTube but hate how little sticks" **Tone guidance:** The Callout must make the ICP feel seen and understood. Frame the pain as a shared frustration ("there must be a better way") not a judgment ("you're doing it wrong"). Empathize, don't accuse. **Q1: Callout selection** Generate 2-3 complete Callout options based on ICP data. Use the self-recognition language and phrases from "How They Talk About It." Each must be a full 3-part descriptor. Use AskUserQuestion: "What will make your ideal customer stop and say 'that's about me'? Here are options based on how they describe themselves:" Options: 2-3 generated Callouts + "I want to tweak one" **Q2 (conditional): Refinement** Only if user picks "I want to tweak one" — ask what to adjust. **Q3: Callout validation** Use AskUserQuestion: "Read this out loud: '[selected Callout]'. Does your ideal customer hear this and think 'that's exactly me'?" Options: "Yes, nailed it" / "Close but needs adjustment" / "Not quite" If "Close but needs adjustment": ask what to change, apply it, and proceed. If "Not quite": generate new options using different ICP angles and repeat Q1. ### Phase C: Craft the Magnet (2-3 questions) The Magnet is the utopic desired future — the state where everything is perfect. It should be the inverse of the Callout's pain. **Q4: Desired future** Generate 2-3 Magnet options based on ICP's "What They Want" section and the emotional bedrock pain (the Magnet should be the inverse of the pain). For B2B ICPs, ground the Magnet in their success KPIs — the desired future should map to metrics their boss cares about. Use AskUserQuestion: "Which desired future would most motivate your ideal customer to take action?" Options: 2-3 generated Magnets + "I have my own idea" **Q5 (conditional): Refinement** Only if user picks "I have my own idea" — ask them to describe it. ### Phase D: Generate Value Prop Options (1 question) Combine Callout + Magnet into 3-4 complete value proposition messages. Each uses a different angle: 1. **Pain-led:** Leads with the Callout's pain, resolves with Magnet 2. **Aspiration-led:** Leads with the Magnet, grounds with Callout 3. **Action-led:** Leads with what changes, bridges Callout to Magnet 4. **Identity-led:** Leads with who they are, then pain → future Display all options: ``` ┌───────────────────────────────────────────────────────────────┐ │ YOUR VALUE PROP OPTIONS │ ├───────────────────────────────────────────────────────────────┤ │ │ │ Callout: [Selected descriptor] │ │ Magnet: [Selected desired future] │ │ │ │ 1. [Pain-led message] │ │ 2. [Aspiration-led message] │ │ 3. [Action-led message] │ │ 4. [Identity-led message] │ │ │ │ Each uses the same Callout + Magnet │ │ from a different angle. │ └───────────────────────────────────────────────────────────────┘ ``` Use AskUserQuestion: "Which value proposition do you want to lead with?" Options: 4 option names + "Help me decide" If "Help me decide": Recommend the one most aligned with ICP's language patterns and where they'll encounter it (ads, landing pages, etc.). ### Phase E: CTA & Validation Goal (1 question) The CTA is not just a button label — it's the action you're asking people to take, which becomes the metric you validate against. Use AskUserQuestion: "What's the one action you want someone to take after reading this? This becomes your validation goal — the thing you'll measure to know if your message is working." Options: "Try it free" / "Join the waitlist" / "Add to Chrome" / "Something else" ## Output Save to `pmf/value-prop.md` using the template from `templates/outputs/value-prop.md`. Fill in: - **Selected Value Prop** heading with the chosen option name (e.g., "Pain-led") - **The Callout** with the selected descriptor and its components - **The Magnet** with the selected desired future - **The Message** with the full value proposition text - **CTA** with the call to action (this becomes the validation goal) - **Open Questions** with any "not sure" items from the process - **Alternative Options** with the 2-3 options NOT selected, including their message text, angle, and brief reason why not selected ## Progress Display Show only at the END: ``` ┌───────────────────────────────────────────────────────────────┐ │ VALUE PROPOSITION DEFINED │ ├───────────────────────────────────────────────────────────────┤ │ │ │ Callout: [Descriptor] │ │ Magnet: [Desired future] │ │ Message: [Selected value prop] │ │ CTA: [Action] → this is your validation goal │ │ │ │ Alternatives: [N] options saved for A/B testing │ │ │ │ Saved to: pmf/value-prop.md │ │ │ └───────────────────────────────────────────────────────────────┘ ``` ## Updating Existing Value Prop If `pmf/value-prop.md` already exists: 1. Read the current file 2. Show summary (Callout, Magnet, Message) 3. Route to the update-value-prop command menu 4. Update only the relevant sections 5. Save the updated file ## Research Support When user says "not sure" or wants research: - Use WebSearch to find relevant messaging examples - Look for competitor positioning, successful headlines in the space - Present findings and let user decide what fits ## AskUserQuestion Guidelines **IMPORTANT:** Make questions **self-contained** — include all context in the question text itself. - Keep headers to **1 word max** or omit entirely - Don't rely on text above the question to provide context ## Attribution Created by Adi Shmorak, The P/MF Detective. For feedback: adi@adidacta.com