🇺🇸English

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

1. Clarify the target outcome and business model; set default thresholds accordingly.
2. Identify the RAT (the assumption that kills the business if wrong).
3. Plan the validation ladder: interviews -> smoke test -> concierge/WoZ -> paid pilot.
4. Run the cheapest falsifiable test first; pre-register PASS/FAIL thresholds and stop rules.
5. Score all 9 dimensions using evidence; downgrade scores when evidence is weak.
6. Produce a decision memo: verdict, why, what would change the decision, and the next smallest reversible step.

9-Dimension Scorecard

Verdict thresholds (default) :

• 80–100: GO
• 60–79: 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)

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

• startup-review-mining - Pain point evidence
• startup-trend-prediction - Market timing inputs
• startup-competitive-analysis - Competitor landscape

Feeds Into

• router-startup - Startup decision routing
• product-management - Validated requirements and roadmap inputs
• startup-business-models - Monetization and packaging decisions

Resources

Templates

Data

Weekly Installs

202

Repository

vasilyu1983/ai-…s-public

GitHub Stars

46

First Seen

Jan 23, 2026

Security Audits

Gen Agent Trust HubPassSocketPassSnykPass

Installed on

opencode168

gemini-cli164

codex163

github-copilot145

claude-code142

cursor141