A threshold is a pre-commitment. You set it before you have results so that when results arrive, the decision is already made. Ten people paid: build. Nine: pass. The threshold isn’t the interesting part — the interesting part is that you set it in advance, when you’re not holding actual data that you’re tempted to reinterpret.

Without a threshold, results are infinitely interpretable. Eight people paid — is that close enough? Maybe the price was wrong. Maybe the timing was off. Maybe if you ran it another two weeks. Maybe you should survey the people who didn’t pay and find out what would have moved them. These questions are all reasonable, and every one of them is a path to deferring the decision. The threshold closes that path. You defined “enough” when you weren’t rationalizing, so “enough” means something now.

The number itself matters less than its specificity. A threshold of ten is not meaningfully different from a threshold of eleven. What matters is that you committed to a number before you started. The commitment is what the threshold does — it converts a judgment call into a binary outcome. Without the commitment, you’re just collecting data in a void, waiting until you feel ready to decide, which is not a process with a predictable ending.

This is why thresholds work better when you set them with a rationale. Not just “ten” but “ten, because if ten CRE professionals who have never met me are willing to pay twenty dollars a month for this, that’s a 1.3% conversion rate from a known-size audience, which validates that the pain exists at this price for this offer from this person.” When the rationale is explicit, you can evaluate whether results are actually meeting the threshold or skirting it. Eight paid but two of them were close friends who wanted to be supportive — that’s a different data point than ten strangers. The threshold is the floor; the rationale tells you what the floor is measuring.

The failure mode is soft thresholds: ranges, bands, phrases like “around ten” or “a handful.” These feel like flexibility but they’re just undecided thresholds wearing a number. When results arrive at seven, “around ten” becomes “well, seven is close.” Close to what? To a number that was already approximate. The threshold has to be a line, not a zone, or it’s not doing its job.

Pre-commitments feel constraining before you have data and freeing after. The constraint is the point. Decisions made in advance, without results in hand, are made by the version of you who set up the test without any attachment to how it ends. That version of you is more objective than the version holding the data. Trust the pre-committed version. Set the number now.