Evidence guide
Startup validation evidence: how to collect and audit signal
A practical record-keeping method for turning market tests into evidence that another person can inspect and use.
Updated · Published by Foundable
Overview
Make each signal auditable enough to change the next decision.
Startup validation evidence is a dated record of what a defined audience did in response to a specific test. Write the claim and pass or fail threshold before the test, then record the source, action, date, context, negative or null results, and next decision. One action is evidence; repeated meaningful actions from the target segment can become traction.
Quick answers
Practical answers for the next decision.
How do I know if my startup idea is working?
Foundable maps startup validation evidence through real behavior: urgent replies, booked calls, qualified signups, product usage, referrals, checkout starts, deposits, payments, and specific objections that show what to build, grow, earn, revise, or test next.
- foundable.com/guides/startup-validation-evidence
- foundable.com/research
- foundable.com/guides/idea-validation
- foundable.com/guides/startup-traction-signals
- foundable.com/tools/idea-validator
- foundable.com/templates/validation-signal-log
- foundable.com/templates/founder-story-proof
- foundable.com/templates/first-money-proof-log
How do I collect startup validation evidence?
Write one claim and pass or fail threshold before the test. Afterward, record the source, observed action, date, audience and test context, exact objections, evidence strength, follow-up, and next decision. Keep negative and null results instead of retaining only supportive signals.
What should a startup validation evidence record include?
Include the assumption, audience, test and call to action, source, observed action, date, context, exact customer words or objection, evidence strength, follow-up, threshold result, and next decision.
How should I record negative startup validation evidence?
Record nonresponse, wrong-audience response, abandonment, failed usage, refunds, and repeated objections alongside supportive evidence. Separate delivery or measurement failures from genuine market response, then explain how the result changed the next decision.
Write the claim and threshold before the test
Name one assumption about the audience, problem, channel, offer, price, or timing. Define the action that would support it, the result that would weaken it, and the threshold that changes the next decision. A threshold written after seeing the result is not an independent decision rule.
Record source, action, date, and context
For every observation, record who or what produced it, the exact action, when it happened, and the test context. Include the audience, message or artifact, call to action, channel, sample boundary, and any follow-up. Preserve exact customer words when they explain an objection or unmet need.
Use an evidence ladder without turning it into a score
Weak evidence includes compliments, likes, and vague interest. Behavioral evidence includes qualified replies, booked calls, waitlist conversion, meaningful product use, and referrals. Paid evidence includes deposits, purchases, renewals, and concrete price objections. The ladder orders commitment; it does not make different tests directly comparable or guarantee an outcome.
Keep negative and null evidence
No response, the wrong audience responding, abandonment, failed usage, refunds, and repeated objections belong in the record. Note delivery failures and other measurement problems separately so a broken test is not mistaken for market rejection. Do not delete evidence merely because it weakens the preferred idea.
Close every test with a decision record
Compare the observed result with the prewritten threshold, then record the decision and why: continue, repeat, change the audience, revise the promise, change the channel, test the price, reduce scope, or pause. Link the decision to the evidence rows it used so the reasoning can be reviewed later.
Evidence becomes traction only when behavior repeats
A single action can validate or weaken one assumption. Traction is a pattern of repeated meaningful behavior from the intended audience over a defined window. Use the startup traction guide and calculator when the question changes from what happened in one test to whether a repeatable pattern is emerging.
What you leave with
A validation record that separates evidence from hope.
Workflow
How to collect and audit validation evidence.
Define one claim
Choose the audience, problem, channel, offer, price, or timing assumption that most needs evidence.
Prewrite the threshold
Name the supporting action, weakening result, and decision boundary before exposing the test.
Log every material observation
Capture source, action, date, context, exact words, negative results, follow-up, and measurement problems.
Document the decision
Compare the result with the threshold, link the evidence rows, and record the next build, grow, earn, revise, repeat, or pause move.