| name | startup-idea-validation |
| description | Use when validating a startup idea before building. Produces evidence-based GO/NO-GO decisions using a 9-dimension scorecard (problem, market, timing, moat, unit economics, founder-market fit, feasibility, GTM, risk), a validation ladder (interviews -> smoke test -> concierge/WoZ -> paid pilot), and riskiest-assumption-first experiments. |
Startup Idea Validation
Systematic validation for testing ideas before building: define hypotheses, collect evidence, score the opportunity, and make a decision you can defend.
Operating Principles (2026)
- Prefer decisions over inventories: each dimension ends with
GO / CONDITIONAL / PIVOT / NO-GO and a next action.
- Separate evidence quality from confidence: weak evidence cannot justify a high score.
- Pre-register thresholds and stop rules before running experiments (avoid moving goalposts).
- Validate willingness-to-pay and time-to-value early (price is part of the product).
- Calibrate thresholds to the target outcome (venture-scale vs cash-flow business) and business model (B2B SaaS, B2C, marketplace, services).
- Stay safe and ethical: no misrepresentation, respect ToS, and handle customer data with minimization and retention limits.
Intake Checklist (Ask First)
- One-sentence idea + target user + job-to-be-done
- Business model: B2B/B2C, SaaS/usage-based/marketplace/services, ACV/ARPU range
- Geography, constraints (regulated domain, procurement/security requirements, data access)
- Target outcome: venture-scale, profitable small business, or thesis-driven R&D
- Current evidence: interviews, pilots, pre-sales, traffic, competitor list, pricing assumptions
Choose the Right Output
Workflow
- Clarify the target outcome and business model; set default thresholds accordingly.
- Identify the RAT (the assumption that kills the business if wrong).
- Plan the validation ladder: interviews -> smoke test -> concierge/WoZ -> paid pilot.
- Run the cheapest falsifiable test first; pre-register PASS/FAIL thresholds and stop rules.
- Score all 9 dimensions using evidence; downgrade scores when evidence is weak.
- Produce a decision memo: verdict, why, what would change the decision, and the next smallest reversible step.
9-Dimension Scorecard
| Dimension | Weight | What it measures |
|---|
| Problem severity | 15% | Urgency, cost of inaction, current workarounds |
| Market size | 12% | Sufficient demand for the target outcome |
| Market timing | 10% | Clear “why now” and tailwinds |
| Competitive moat | 12% | Defensibility over time |
| Unit economics | 15% | Profit path (incl. payback and margins) |
| Founder-market fit | 8% | Access, expertise, and execution capability |
| Technical feasibility | 10% | Buildability, dependencies, constraints |
| GTM clarity | 10% | ICP, channels, motion, first customers |
| Risk profile | 8% | What can kill it and likelihood |
Verdict thresholds (default):
80–100: GO60–79: CONDITIONAL (validate RAT first)40–59: PIVOT<40: NO-GO
Deep scoring rubrics and calibration live in validation-methodology.md.
Evidence Rules
- Strong evidence is behavioral commitment with cost (time, money, switching, access); weak evidence is opinions and hypotheticals.
- Triangulate important claims across at least two sources (especially market sizing and competitor state).
- Keep an evidence trail: link + capture month; separate “fact” vs “assumption”.
Validation Ladder (Default)
| Step | Goal | Strong signal |
|---|
| Interviews | Validate the problem and context | Repeated pain with real workarounds and spend |
| Smoke test | Validate demand | Qualified conversion with price shown |
| Concierge/WoZ | Validate workflow value | Users complete the job and return |
| Paid pilot | Validate willingness-to-pay | Paid, renewed, or expanded |
AI / Automation Notes (2026)
If the idea depends on AI (agents, copilots, automation), validate these explicitly:
- Data rights and access: can you legally and reliably access required data?
- Reliability: define success metrics, failure modes, and human fallback; validate on real workflows.
- Cost-to-serve: model inference + retrieval + human-in-the-loop costs in
assets/financial-modeling-calculator.md.
See hypothesis-testing-guide.md for AI-specific experiment patterns.
Integration Points
Receives From
Feeds Into
Resources
| Resource | Purpose |
|---|
| validation-methodology.md | Scoring rubrics and calibration |
| hypothesis-testing-guide.md | Experiment design and RAT workflows |
| market-sizing-patterns.md | TAM/SAM/SOM methods and pitfalls |
| moat-assessment-framework.md | Defensibility analysis |
| customer-interview-guide.md | Interview methodology, scripts, and analysis |
| landing-page-validation.md | Smoke tests, conversion benchmarks, landing page tools |
| competitive-landscape-assessment.md | Competitive scan, gap analysis, market mapping |
| pivot-framework.md | Pivot triggers, types, decision framework, case studies (Slack, Instagram, Shopify) |
Templates
Data
