demand-check

Overview

Lightweight evaluation: only runs Demand Authenticity and the Elevator Pitch test. No full agent dispatch; runs directly in the main context for speed. This is a 5-minute first-pass filter.

Process

Step 1: Quick Product Understanding

• Local project: Read README + scan top-level directory structure. Identify the core value proposition.
• External app / idea: Use the user's description. Do a single WebSearch for the product if needed.
• No argument: Use current working directory.

Do not deep-dive into code. This is a surface-level scan.

Step 2: Detect Platform and Load Framework

Same platform detection logic as evaluate skill:

• Check for .xcodeproj / Package.swift → iOS
• Check for package.json → Web
• Otherwise ask or infer

Locate reference files by searching for **/product-lens/references/dimensions/01-demand-authenticity.md.

Read:

1. dimensions/01-demand-authenticity.md — extract:
   • Universal sub-questions
   • Platform-specific sub-questions (from ### iOS if iOS detected, otherwise ### Default)
   • Scoring anchors
2. modules/elevator-pitch.md — extract:
   • Default constraints
   • Platform-specific constraints (from ### iOS under ## Platform Constraints if iOS detected)
   • Judgment criteria

Step 3: Demand Dimension Assessment

Answer the dimension's sub-questions using available evidence:

• README claims and problem statement
• Quick code structure scan (is 80% of code serving core value?)
• User's own description of the problem being solved

Score 1-5 stars with one-sentence justification.

Step 4: Elevator Pitch Test

Attempt to write:

• Tagline (<=30 chars) — apply platform-specific constraints if iOS
• First description sentence

Judge: Clear / Vague / Cannot articulate

Step 5: Quick Alternative Scan

Run 2-3 WebSearches:

• "best [category] app [year]"
• "[problem description] alternative"
• "[category] indie app" (if relevant)

Note top 3-5 alternatives and their approach. This is not a full market scan — just enough to answer "is someone already doing this well?"

Step 6: Verdict

Output one of three verdicts:

Verdict	Criteria	Meaning
Pass	Demand >=3★ AND Pitch not "Cannot articulate" AND alternatives are beatable	Green light: proceed to full evaluation or building
Caution	Demand 3★ with Pitch "Cannot articulate", OR Demand 2★ with Pitch "Clear", OR dominant alternative exists but has clear gaps	Investigate further: run full /evaluate before committing
Fail	Demand <=2★ AND no differentiation angle, OR Pitch "Cannot articulate" AND problem already solved well by free tools	Reconsider: this idea needs significant rethinking

Validation Playbook (Caution verdict only)

When the verdict is Caution, append a ## Validation Playbook section to the output. Select 2-3 methods from the list below based on what is uncertain:

• If demand unclear: "AI-analyze 200+ competitor 1★-3★ App Store reviews to extract unmet needs. Success: >=3 recurring complaint themes that this product addresses."
• If differentiation unclear: "Post pain-point description (not product pitch) in target community (Reddit/Discord). Success: >=20 upvotes or >=5 'I have this problem' replies within 48 hours."
• If willingness-to-pay unclear: "Soft-launch in a small market (e.g., New Zealand) at target price for 2 weeks. Success: >=3 paid conversions without promotion."
• If dogfooding viable: "Use your own product daily for 2 weeks, logging every friction point. Success: friction log has <=5 entries AND you genuinely miss the product when you stop."
• If AI replacement risk unclear: "Ask ChatGPT, Claude, and Perplexity to recommend tools for this product's core job. Success: AI tools recommend the category but can't fully replace it."

Each selected method must include: method description, success criteria, failure criteria, and timeline (all <=2 weeks).

Step 7: Save Report

If the evaluation target is a local project:

1. Check if docs/08-product-evaluation/ directory exists. If not, create it.
2. Save the report to docs/08-product-evaluation/{YYYY-MM-DD}-demand-check.md
3. Check for previous demand-check reports in the same directory. If found, note score changes in the verdict section.

Output Format

# Demand Check: [Product Name]

## Elevator Pitch
> **Tagline:** [<=30 chars]
> **Description:** [first sentence]
> **Verdict:** [Clear / Vague / Cannot articulate] — [reasoning]

## Demand Authenticity (需求真伪) [★★★☆☆]

[Sub-question answers with evidence]

## Alternatives Found

| Alternative | Approach | Pricing | Threat Level |
|-------------|----------|---------|--------------|
| [name] | [how it solves the problem] | [price] | [High/Medium/Low] |

## Verdict: [Pass / Caution / Fail]

[One paragraph explaining the verdict and suggested next steps]

## Validation Playbook
(Only present when verdict is Caution)

Before committing to build, validate these uncertainties:
1. **[Uncertainty]:** [Method] — Success: [criteria] / Fail: [criteria] — Timeline: [<=2 weeks]
2. **[Uncertainty]:** [Method] — Success: [criteria] / Fail: [criteria] — Timeline: [<=2 weeks]