How to validate a startup idea
Most idea validation fails for one reason: founders collect opinions that agree with them and call it evidence. This guide walks through an evidence-based way to test an idea across six dimensions, so a score reflects what you actually know — not how optimistic you feel on the day you answer.
What validation actually means
Validation is not proving that your idea is good. It is reducing the number of unexamined assumptions between you and a paying customer. Every claim you make about the problem, the buyer, the price, and your ability to build it is an assumption until something outside your own head confirms it. The goal is to find the assumptions that would kill the business if wrong, and to test those first — cheaply, and before you spend a year building.
Reducing delusional thinking
Two failure modes dominate. The first is confirmation seeking: asking friends whether they like the idea instead of asking strangers to pay for it. The second is contradiction blindness: claiming a huge market and a tiny, specific niche in the same breath, or projecting fast growth with no channel and no budget. Writing your assumptions down and reading them side by side surfaces contradictions that never appear when you argue the idea out loud.
The six dimensions to validate
The Idea
Problem significance, solution advantage, differentiation, defensibility, adoption.
The Entrepreneur
Founder fit, domain insight, execution capability, commitment, learning velocity.
The Market
Demand evidence, customer definition, alternatives, accessibility, reachable size.
The Finances
Revenue logic, willingness-to-pay evidence, unit economics, capital, survival.
The Resources
Capabilities, product readiness, customer access, dependencies, gap-closing.
The Risks
Critical assumptions, market/execution/external risk, pre-mortem.
A strong idea is rarely strong everywhere. What matters is whether the weak dimensions are fixable with work you can actually do, and whether the strong ones rest on evidence rather than enthusiasm.
The evidence hierarchy
Not all proof is equal. Rank every claim you make by the strength of what supports it. A claim backed by personal intuition should never carry the same weight as one backed by repeat paying customers — and a validation score that ignores this difference is worse than no score at all.
- 0No evidence
- 1Founder belief
- 2Secondary research
- 3Informal conversations
- 4Structured interviews / surveys
- 5Observed behaviour
- 6Commitment evidence
- 7Pilot evidence
- 8Paid pilot
- 9Multiple paying customers
- 10Repeat purchase / renewal
Whenever you record an answer about your startup, record the evidence level next to it. The gap between how good your answers look and how well they are evidenced is your real risk profile.
A 30-day validation sequence
- Days 1–5 — write the assumptions down. List every claim behind the idea, the customer, the price and the channel. Mark each with its current evidence level.
- Days 6–10 — rank by impact. Sort assumptions by "if this is wrong, is the business dead?" High impact and low evidence goes first.
- Days 11–20 — run cheap tests. Problem interviews with strangers, a landing page with a real call to action, a concierge version delivered manually, or a pre-order. Each test should be able to fail.
- Days 21–27 — test willingness to pay. Ask for money, a signed letter of intent, or a scheduled pilot. Interest is not demand.
- Days 28–30 — rescore. Update each assumption's evidence level and see which dimensions moved. Persevere, adjust the segment, or drop it.
Common mistakes
- Treating survey answers about hypothetical behaviour as demand.
- Defining the market as a top-down percentage of a huge industry number.
- Ignoring the current alternative — including "spreadsheet" and "do nothing".
- Assuming unit economics work at a scale you have never operated at.
- Validating the product while never testing the acquisition channel.