You've Researched Enough When Three People Who Live the Problem Agree It's Real

There's a sentence that ends more good decisions than any wrong answer ever has: "Let's table this and revisit it next month." Its cousin is worse – "let's form a committee to look into it." Both sound responsible. Both are usually avoidance wearing the costume of diligence, and if you listen for them, you can hear the exact moment a team stops diagnosing a problem and starts hiding from the decision it implies.

I say this as someone who tells everyone else to watch for the opposite failure – rushing to a conclusion before you understand the problem. Both are real. The hard part isn't knowing that you can over- or under-research; it's knowing where the line is, in the moment, when you genuinely can't tell whether you need more information or you're just afraid to commit. So here's the rule I actually use.

Why we keep going past enough

The honest reason teams over-research usually isn't diligence. It's that a hypothesis is a thing you can be wrong about, out loud, in front of people – and more research feels like insurance against that. As long as you're still gathering, you haven't committed, and if you haven't committed, you can't be wrong yet. That pull is strongest for exactly the people with the most to lose from a public miss: the founder, the senior leader, anyone whose name ends up attached to the call. But the insurance is fake. More research doesn't make you less wrong. It makes you wrong later, after you've spent more getting there. The only thing that actually lowers the cost of a wrong call is finding out sooner, cheaply, on purpose – which is the exact opposite of what stalling does. Once you see that, "we need more data" stops being a neutral request and starts being a question worth interrogating: more data toward what decision, and what would it have to say to change your mind?

You've researched enough when three or four people who live it agree

There's a heuristic from The Mom Test about customer interviews: when you've heard the same thing about four times in a row, you have enough – you don't need more interviews, you need to act on what you heard. The same logic works for diagnosis. Gather enough to form a worldview and a real hypothesis, then float it as an educated guess to the people closest to the problem. If three or four of them – the ones who actually live it, not the ones adjacent to it – say "yeah, that's a reasonable read," you have enough to move.

Notice what that rule is not. It's not "enough when you're confident," because confidence never arrives on its own schedule, and waiting for it is how a two-week question becomes a two-quarter study. It's not a target number of interviews, either. It's agreement among the people who feel the problem in their day, which is a very different signal from volume of research. You can run forty interviews and learn less than you'd learn from four hard conversations with people who live the thing, if the forty were with people one step removed.

The signal is agreement, not applause

The rule has a trap worth naming: the three or four have to be the people who live the problem, not the people who like the idea. A room full of stakeholders nodding at your slide is applause, and applause is easy to get and worth almost nothing – people are agreeable, they want to be supportive, and most of them don't feel the problem in their day. The signal you're actually after is narrower and harder to earn: does the person who wakes up inside this problem recognize your description of it as their own? When you say the hypothesis back to them and they interrupt to add the detail you missed, or correct the part you got slightly wrong, that's the signal – they're living it with you. When they say "yeah, that makes sense," politely, and move on, that's applause. The whole rule depends on whose agreement you're counting, so learn to tell the two apart before you count to four.

Then you stop researching and place the bet

The threshold isn't the finish line. It's the starting gun. Once three or four people who live it agree your hypothesis is real, you set the signals and key results that would tell you whether you're right – in advance – and you run a bounded experiment against them. If the results don't show up, the hypothesis was wrong, and you want to learn that sooner rather than later. I'd rather find out in three weeks than three quarters.

Deciding the signals before you start is the discipline that separates a real test from a rubber stamp. One operator who has run market entries across a hundred-plus countries frames it as agreeing, before the first dollar, on the result that earns more investment and the result that ends the test. Skip that step and every weak signal gets a sympathetic explanation and another month. And judge the early signal on the right thing: when you're testing a new direction, the first weeks are about resonance, not volume – are the right people leaning in, does the framing land, is a pattern forming – not a pipeline number that won't mean anything yet. The disciplined version of this is a short, bounded validation, measured in weeks and kept cheap, not an open-ended study that quietly becomes the work.

Both directions are a way to get it wrong

Moving too slowly is the failure I've described – the committee, the table-it, the search for a certainty that was never coming. But moving too fast is just as real, and I've done it. Earlier in my career I formed weaker hypotheses and tested them too soon, and once I ran a team through changes so quickly they told me they could never find their footing. We slowed the cadence way down until they could. Another team found that changing direction every two weeks meant they never got into a rhythm, so they committed to changing course only every couple of months, and that worked for them. Every team has its own tempo.

What doesn't change is where the accountability sits. A committee's real function is to dilute ownership until no one can be blamed for the call. The job asks the opposite of you: form the hypothesis, make the decision, put your name on it, and if it goes wrong, that's yours to fix. "Enough" was never a feeling you were waiting to have. It's a threshold you agreed to in advance, and a decision you were willing to own the moment you crossed it.