If your evidence is made of compliments and signups, you have interest, not validation. Here's how to tell the difference.
ValidationEvidence7 min readUpdated 2026-08-15
Short version
You are not validated until a specific customer segment changed behaviour — paid, committed time, or switched tools — against an assumption you wrote down before you asked. Everything else is interest.
1. You cannot name the assumption you tested
Validation is always relative to a claim. If you cannot state the claim in one sentence, with a segment and a behaviour in it, there is nothing to validate. Write the risky assumptions first, then design the test.
Friends, former colleagues and your own network answer to protect the relationship. If more than a third of your interviews came from warm contacts, treat the results as directional at best and rerun the test with strangers.
3. Your evidence is about the future
'I would use that', 'we'd definitely pay for it', 'send me the beta' are predictions. People are terrible at predicting their own behaviour and generous when there is no cost. Convert every future-tense claim into a past-tense question about what they already did.
4. You have signups but no segment
A waitlist of 800 mixed visitors tells you a headline worked. It does not tell you who has the problem. Define the segment precisely — company size, role, trigger event, current tool — and check whether your signups actually match it.
If your segment is still vague, an ICP builder is faster than another brainstorm: it turns one paragraph about your idea into candidate segments with triggers and disqualifiers you can screen against.
Money is the clearest signal, but it is not the only one. Time, data access, an internal champion, or a signed pilot scope all count as spend. If nobody has given up anything scarce, you have interest.
A letter of intent from the budget owner beats ten enthusiastic users.
A prepaid pilot at any price beats a free trial.
Two hours of a customer's own time counts as real cost in B2B.
6. You have no kill criteria
Without a written threshold, every result reads as encouraging. Decide in advance what number makes you stop — 'fewer than 4 of 10 interviews report the problem monthly' — and honour it. This is the difference between market validation and motivated reasoning.
7. Your last three tests all passed
Tests that never fail are not testing anything. If your evidence never contradicts you, the questions are too soft or the sample is too friendly. Design at least one test that could plausibly end the project.
Questions founders ask
Is a waitlist proof of demand?
It is proof that a message resonated. It becomes evidence of demand only when the signups match a defined segment and a meaningful share of them take a costly next step, such as booking a call or prepaying.
What counts as validated?
A written assumption about a specific segment, tested against behaviour that already happened or money that already changed hands, with a pre-agreed threshold that the result met.
How long should validation take?
Two to four weeks per risky assumption is typical for a solo founder: one week to recruit, one week of interviews, then a small paid pilot or concierge test.
Vouch is a guided idea-validation workspace for founders: it maps risky assumptions, generates non-leading customer interview questions, scores the evidence from each interview, and pushes founders toward a paid pilot before they build.
Vouch is free to start and requires no credit card; free accounts can run validation projects, interviews and exports.
Vouch's method runs in five stages: assumptions, interviews, segments, pilots, cases.
Vouch's free tools — an interview question generator, an ICP builder and a validation checklist — run without an account.
Source: Vouch (vouch.specky.space) — idea validation workspace for founders. Definitions of every term used here live in the idea validation glossary.