Product-market fit, before the metric exists

You ask twenty users how they'd feel if they could no longer use your product. Twelve say very disappointed. That's 60%, well past the famous 40% line. It doesn't tell you whether you have product-market fit. It tells you whether those twenty do.

The 40% threshold is useful. It's not a test you pass to declare product-market fit achieved — it is a forcing function that makes you pick the twenty users. The picking is the work.

How the Sean Ellis test works

The test is one question: how would you feel if you could no longer use the product? The answers are very disappointed, somewhat disappointed, and not disappointed, with a fourth for people who no longer use it. Sean Ellis, who ran early growth at Dropbox and Eventbrite, set the bar after benchmarking nearly a hundred startups. Companies with strong traction almost always cleared 40% very disappointed. Companies that struggled to grow almost always fell short.

The score only means something when the people answering have used the product for real. Superhuman asked people who had used the product at least twice in the past two weeks, and its first score was 22%. Rahul Vohra, its founder, found results turn directionally correct at around 40 responses. Twenty is enough to start. Forty is where it starts pointing somewhere.

Picking the twenty users

The Sean Ellis threshold tells you whether your selected cohort feels strongly. It doesn't tell you which cohort to select. It doesn't tell you which question to ask. It doesn't tell you whether the answers mean what you think they mean.

If you run the survey on your entire user base, get 25%, and conclude "not product-market fit," you have skipped the only interesting step. Product-market fit is a relationship between a specific kind of user and a specific moment of your product. The aggregate hides the relationship. You need the slice.

Product-market fit is local

You don't have product-market fit "in general." You have it with a specific kind of user, doing a specific job, at a specific moment in their workflow.

Two users in the same logical segment can have wildly different relationships to your product — the one who got onboarded by a founder over a video call and the one who self-served at two in the morning on a Tuesday. The first signed up for the relationship. The second signed up to try a thing. Their answers to "how would you feel" mean different things and should be counted separately.

This is why the small focused cohort approach works. You don't need a thousand users to love what you built. You need fifty users in the same job, doing the same thing, who would be miserable without it. Find those fifty before you ask whether you have product-market fit.

The signals before the metric

Before there's a survey response, there's a usage pattern. Before there's a usage pattern, there's a conversation. Before there's a conversation, there's someone describing what they were trying to do — and what got in their way.

Most early-stage product-market fit work is being in those conversations, with the right people, and noticing when the shape changes. A founder who talks to five users a week for three months has more information than a dashboard could ever provide — and the information arrives in time to act on.

The metric becomes useful later, when you need to compress what you have learned into a number you can show to a board or compare against a previous cohort. Before then, the metric is downstream of the work.

Tighten the cohort, not the product

When product-market fit feels weak, the founder instinct is to change the product. Sometimes that's right. More often, the product is right for somebody — just not the people you happened to listen to.

Tightening who you serve is usually faster than rebuilding what you serve. Sharper segment. Narrower job. More specific moment. If your survey returns 25% across all users but 58% among "early-stage SaaS founders in their first 18 months who self-serve onto a paid plan within seven days," the move is not to change the product. The move is to be honest about who the product is for.

Pivoting the product is dramatic and gets quoted in startup biographies. Tightening the cohort is unglamorous and almost always the move that works.

What the work looks like

The work is a small set of routines, run honestly, for longer than feels productive. Five user conversations a week, transcribed and re-read. A weekly check of cohort retention by signup-week. A short, current list of who you are building for, updated when a conversation forces it. The survey, eventually, on a cohort you can defend.

None of this looks like a playbook. It looks like a habit.

The moment you "achieve product-market fit" is the moment you notice that the work you've been doing keeps producing the same kind of conversation — people telling you they can't imagine using anything else, in language you didn't put in their mouths. The 40% rule shows up later, as a number that confirms what you already heard.