How to measure product-market fit
Product-market fit is measurable with two tests. The survey test, from Sean Ellis: ask active users how they would feel if they could no longer use the product, and treat 40 percent or more answering very disappointed as the threshold of fit. The structural test: plot cohort retention curves, and look for the flatten. A curve that levels at a stable plateau means a durable segment has adopted the product; a curve that declines toward zero means no acquisition program can outrun the leak.
The tests answer the same question from two directions, stated preference and revealed behavior, and serious operators run both, because either alone can mislead.
Running the Ellis survey properly
Survey active users, not signups: people who used the product at least twice in the last two weeks.
Ask the canonical question with three answers: very disappointed, somewhat disappointed, not disappointed.
Read the very-disappointed share against the 40 percent threshold.
Segment the very-disappointed group: who are they, what do they use, what brought them.
Act on the split. The worked example is Rahul Vohra’s account at Superhuman (First Round Review, 2019): double down on what the very-disappointed love, fix what holds the somewhat-disappointed back, ignore the rest, and re-survey. Vohra reported moving Superhuman’s score from 22 percent to 58 percent in under a year with that engine.
Reading retention curves
Measure at the product’s natural frequency: daily for messaging, weekly for project tools, quarterly for tax software; the wrong frequency makes a healthy product look sick. Then read the shape. Declining toward zero: no fit, and the correct investment is product work and customer research, not channels. Flattening: fit with the segment the plateau represents, and the plateau’s height is the foundation growth will compound on. Bending upward as cohorts age, through resurrection or expanding usage, is the signature of an unusually strong product.
What to do below the threshold
Not growth spend. The Startup Genome project’s failure research identified premature scaling, spending on growth before the system warranted it, as the most common pattern among failed startups in its dataset. Below threshold, the sequence is narrow the segment, deepen the value for it, re-test. Fit is found in a segment first and expanded from, almost never achieved on average.
Frequently asked questions
What is the 40 percent rule for product-market fit?
Sean Ellis’s benchmark: when 40 percent or more of surveyed active users say they would be very disappointed without the product, the product has crossed the threshold where growth investment tends to pay.
Is product-market fit permanent once reached?
No. It is a gate to verify, not a memory. Markets shift, competitors arrive, and curves that flattened can start declining; the test is worth repeating whenever the strategy leans on the answer.
Can revenue prove product-market fit?
Revenue can mask its absence: heavy acquisition spend produces revenue from a leaking product right up until it cannot. Retention curves and the survey measure fit directly; revenue measures it late.
Fit is verified with curves, not conviction. It is the first dimension a structured growth audit scores, because everything after it amplifies whatever the answer is.
Find out exactly where your growth stands.
The Growthmarkt audit measures your growth system across 8 dimensions and turns the result into a prioritized roadmap.