The point of a validation test — a pre-sale, a landing page, a pilot — is to convert a belief into evidence. You think there is demand; the test is supposed to tell you whether you’re right. But a test can only do that job if you decide, before running it, what result would count as a yes and what would count as a no. Skip that step and the test loses its power entirely, because whatever number comes back, you will find a way to read it as encouraging. The discipline that makes validation real is setting the threshold first.

Human reasoning is relentlessly accommodating after the fact. Run a pre-sale hoping for twenty buyers, get three, and the mind immediately generates the story: the launch was rushed, the channel was wrong, the three who bought are really enthusiastic, surely the next batch will be better. Every one of these may be true. But they are post-hoc rationalizations, and the tell is that you would not have accepted any of them as a success condition beforehand. Three buyers was a failure by any honest standard you would have set in advance — you just didn’t set one, so there’s nothing to hold the result against.

The threshold has to be specific and it has to be committed to in writing before the test runs. “Enough interest to justify building” is not a threshold; it is a phrase elastic enough to fit any outcome. “Ten paid commitments within fourteen days” is a threshold: it has a number, a unit, and a deadline, and when the test concludes you can compare the result against it without interpretation. The specificity is the whole point. A vague target is one you can always claim to have hit; a precise one can tell you no, which is the only thing that makes a yes worth anything.

What makes this hard is that the threshold is a promise to your future self to accept a verdict you might not like. At the moment you set it, you are calm and the outcome is unknown. By the time results arrive, you are invested — you’ve told people about the idea, you’ve imagined building it, the sunk cost of having considered it is already pulling. The threshold set in the calm moment is the instrument that protects the decision from the invested moment. It is worth almost nothing if you let yourself renegotiate it after seeing the data, so the commitment to not renegotiate is as important as the number itself.

The failing outcome, importantly, is not a wasted test — it is the test working. A pre-sale that comes in under the threshold has told you something genuinely valuable: that the demand you believed in isn’t there at the price, message, or channel you tried, and that building would have meant constructing something for an audience that didn’t show up to pay. That information, delivered before the build, is the cheapest version of a lesson that would otherwise have cost months. Treating the miss as failure rather than as the test functioning correctly is what tempts people to discard the result and build anyway, which defeats the entire purpose of testing.

There is a quieter benefit to deciding the threshold first: it forces clarity about what you actually believe before you have any data to hide behind. Naming the number makes you confront your real expectation — do you genuinely think twenty people will pay, or were you hoping the test would let you avoid that question? Often the act of setting the threshold is itself clarifying, because a target you’re embarrassed to write down out loud is a signal about the conviction underneath. The threshold tests the idea, but setting the threshold tests you, and both results are worth having before you commit to the build.