Feature ideas: 
Constraints: 
Goal: 

Convert ideas to testable hypotheses:

1. Start with the OUTCOME you want
   NOT: "Build dark mode"
   YES: "Increase daily active users by 15% and reduce churn by 2%"
   
   Why the distinction?
   Because feature (dark mode) isn't the goal.
   Outcome (more active users) is the goal.
   
   Features are just guesses about how to achieve outcomes.

2. Convert each feature idea to hypothesis
   Feature: "API webhooks for integrations"
   
   Hypothesis: "If we build webhooks, customers will
   - Integrate us into their workflows more deeply
   - Have lower churn (more embedded)
   - Increase usage by 20%
   - Upgrade to paid at higher rate
   
   We'll know this is true if: 60% of webhook-using customers remain active after 6 months"

3. The test before building
   Question: "Do customers actually want this?"
   
   Test options (from cheapest to most expensive):
   1. Ask 10 customers: "Would webhooks solve a problem?"
   2. Build fake landing page: "Sign up for webhooks beta"
   3. Manual test: Do webhooks manually for a customer
   4. Build MVP: Basic webhook implementation
   5. Build full feature: Production-ready webhooks
   
   What's the cheapest test that proves/disproves hypothesis?

4. Success criteria (be specific)
   NOT: "See if people like it"
   YES: "60% of beta customers integrate within 2 weeks"
   
   Why specific?
   - Objective (not subjective)
   - Measurable (you'll know if you hit it)
   - Realistic (you could achieve it)
   - Relevant (it predicts success)

5. Hypothesis confidence
   Before testing, rate your confidence (1-10):
   - "I'm 90% confident dark mode will increase usage"
   - "I'm 40% confident webhooks will reduce churn"
   
   High confidence (8-10): Still build MVP first
   Medium confidence (5-7): Do manual test first
   Low confidence (1-4): Do customer interviews first
   
   Match test rigor to confidence level.

6. Build-first vs test-first decision
   BUILD FIRST if:
   - Easy to build (days not weeks)
   - Low cost of failure (doesn't impact business)
   - You'll learn by building (experimentation)
   - Speed matters (market timing)
   
   TEST FIRST if:
   - Hard to build (weeks/months)
   - High cost of failure (wrong direction wastes runway)
   - You can learn without building
   - Time matters less than accuracy

7. The roadmap as hypothesis queue
   Instead of:
   "Q1: Dark mode, Webhooks, Advanced analytics, Mobile app"
   
   Think:
   "Q1 Hypothesis Testing:
   - Dark mode → Test with 100 users, measure engagement
   - Webhooks → Test with 5 beta customers manually
   - Analytics → Talk to 10 customers first
   - Mobile → Validate demand before building
   
   Go deep on the hypothesis that passes."

8. Learning velocity
   The goal: Learn what customers want BEFORE spending engineering effort.
   
   Track:
   - Features built vs hypotheses tested (ratio)
   - Hypothesis accuracy (% that predict success)
   - Speed to hypothesis test (days)
   - Feature impact (% of features that move needle)
   
   As you improve ratio, you move faster and waste less time.

Examples

Input

{
  "featureIdeas": "Dark mode, Mobile app, Advanced analytics, API webhooks, Slack integration",
  "constraints": "2 engineers, 8 weeks to next funding round",
  "whatYouWantToAchieve": "Increase retention from 40% to 50% (D30), grow enterprise segment"
}

Output

Complete hypothesis framework with outcome-focused goals, feature-to-hypothesis conversion, test options, success criteria, confidence ratings, build vs test decision matrix, hypothesis queue roadmap, and learning velocity tracking.

Sources

https://www.ycombinator.com/library/6R-how-to-find-product-market-fit

Related Prompts

Discover more advanced prompting techniques

Customer Development Interview Playbook (Learn From Users)

Best customer insights come from unstructured conversations. 58% of pivots based on talking to customers vs just looking at data.

Hypothesis Testing Framework (Lean Startup for Ideas)

Turn vague ideas into testable hypotheses with clear success metrics. Prevents expensive pivots. 44% faster learning with documented hypotheses.

Customer Avatar Deep Research (Who Actually Buys)

Go beyond demographics to understand psychology, values, and decision-making. 52% better targeting when psychographics are explicit. Most companies target…