Ask ten founders to define product-market fit and you'll get ten different answers, most of them some version of "you'll know it when you feel it." That's not a measurement framework — it's a mood, and moods are a bad basis for a fundraising narrative or a roadmap decision.
Why the usual proxies fall short
Revenue growth gets treated as proof of fit, but revenue can climb on the strength of aggressive sales motion or a single anchor client long before the product has earned durable demand. Net Promoter Score gets treated as proof of fit, but a good NPS often reflects goodwill toward the team more than the product's ability to retain and expand a customer base. Retention alone gets treated as proof of fit, but retention within a small, hand-held early cohort doesn't say much about what happens once support attention thins out across a larger, less patient customer base.
Each of these is a real signal. None of them is sufficient on its own, and used alone, each is prone to a specific kind of false positive.
The Sean Ellis test, and where it stops being enough
The most widely cited PMF proxy asks users how disappointed they'd be if the product disappeared, with 40%+ answering "very disappointed" treated as the fit threshold. It's a reasonable starting signal — but it measures attitude, not behavior, and it says nothing about whether the segment answering "very disappointed" is large enough, or reachable enough, to build a business on.
A product can clear the 40% threshold with a small, passionate niche that will never scale past a few hundred accounts. Treating that as fit and pouring go-to-market spend behind it produces a company-shaped hole where a product should be.
A framework built on three layers of evidence
Fit is better measured as convergence across three separate signals, not a single test.
Behavioral evidence comes from usage data: cohort retention curves segmented by acquisition channel and use case, not blended across the whole user base. A cohort that keeps returning to a core workflow without prompting is showing fit. A cohort that requires re-engagement campaigns to stay active is not, no matter what it says on a survey.
Attitudinal evidence comes from a properly designed survey — the disappointment test, but segmented by usage intensity and cross-referenced against what respondents say they currently use to solve the same problem. "Very disappointed" from someone using the product weekly to replace a paid alternative is a different finding than "very disappointed" from someone who logged in twice and hasn't been back.
Economic evidence comes from willingness to pay and willingness to expand — not stated intent, but actual behavior: upgrade rates, expansion revenue within existing accounts, and price sensitivity revealed through real objections in sales conversations rather than survey hypotheticals.
Fit shows up where these three layers agree: a segment that keeps coming back, says the product would be missed, and is willing to pay more for more of it. Any one layer alone is a hypothesis. Convergence across all three is closer to evidence.
What a false positive looks like
The most common false positive is a small segment that lights up on all three signals while the broader target market shows none of them. Founders often round this up to "we have fit" when the more accurate read is "we have fit with a segment too narrow to build the current plan around." The finding isn't wrong — the generalization from it is.
The second most common false positive is timing: a cohort acquired during a launch promotion or a founder's personal network shows strong numbers that don't hold once acquisition shifts to colder, less-invested channels. Segmenting cohorts by acquisition source, rather than looking at blended retention, is what catches this before it becomes an expensive lesson.
Bottom line
Product-market fit is a measurable finding, not a feeling — but only when it's measured as convergence across behavior, attitude, and economics within a properly segmented cohort, not read off a single survey question or a top-line growth number. Get the measurement right, and the answer to "do we have fit" stops being a gut call and starts being something you can defend to a board, a fundraising audience, or your own roadmap decisions.