Product market fit: what it means and how to measure it

Sean Ellis, the growth marketer who coined the term "growth hacking", proposed a one-question survey:

How would you feel if you could no longer use this product?

The answers are "very disappointed", "somewhat disappointed", "not disappointed" and "I no longer use it". His rule of thumb, drawn from comparing results across many startups, is that products where 40% or more answer "very disappointed" tend to find it much easier to grow. Treat the 40% as a useful line, not a law.

How to run it well:

• Survey the right people. Ask users who have experienced the core of the product, used it recently and used it more than once. Surveying everyone who ever signed up drags the score down with people who never activated.
• Wait for enough answers. With a handful of responses, one or two people swing the result. Aim for a few dozen at minimum before reading much into it.
• Ask the follow-ups. What is the main benefit you get? What type of person would benefit most? How could we improve it? The "very disappointed" group's answers tell you who your product is really for and what to protect.

The score is most useful as a trend. Run it every quarter and segment it: a 25% overall score can hide a 50% score in one customer type, which is where to focus.

Surveys measure what people say. Retention measures what they do, which makes it the harder evidence.

Plot a retention curve for each monthly sign-up cohort: the share of the cohort still active in week 1, 2, 3 and onwards. Every product loses people early. The question is what happens next.

• Curve slides towards zero. Nearly everyone eventually leaves. Growth only continues while you pour new users in at the top. This is the clearest sign fit is not there yet.
• Curve flattens. After an early drop, a stable share of each cohort keeps using the product month after month. That flat section is a group of people for whom the product works.
• Curve rises slightly over time. Some users come back after lapsing, or usage deepens. This is rare and a strong signal.

Two practical notes. Define "active" as doing the valuable thing, not just logging in; a login can be a user looking for the cancel button. And compare cohorts: if newer cohorts flatten at a higher level than older ones, the product is moving towards fit even if the overall numbers look flat.

Retention tells you whether people stay. Activation tells you whether they get far enough to find out. If only a small share of sign-ups reach the product's core action, a weak retention curve may be an onboarding problem rather than a product problem. Find your activation event (see the aha moment guide) and check what share of new users reach it before judging fit.

The qualitative signals are harder to chart but often arrive first:

• Customers describe the product to others in words you did not give them.
• Support messages shift from "how do I" to "can it also".
• Users build workarounds to make the product do more than it was designed for.
• An outage produces a flood of messages rather than silence.
• Sales conversations get shorter because prospects already know what they want.
• People who churned come back without being chased.

Write these down as they happen, with dates. A log of real customer quotes is far more convincing to you and to anyone you are raising from than a general sense that things are going well.

Several things look like fit and are not.

• A launch spike. A burst of sign-ups from a launch or a viral post measures curiosity. Check the retention curve for those cohorts a few months later.
• Revenue bought with discounts or paid acquisition. If growth stops the moment spend stops, you have a marketing channel, not pull.
• One large customer. A single enthusiastic account can carry the numbers, and its needs may not be anyone else's.
• Praise from friends, investors or other founders. They are not the people who will pay.
• Feature requests alone. People asking for more can mean they love it or that it does not yet do enough to be useful. Retention tells you which.
• A permanent state. Markets move, competitors arrive and customer needs change. Fit you had two years ago needs checking again.

Fit is also specific. It is fit between a product and a particular group of customers, which is why segmenting both the survey and the retention curve matters more than any single headline number.

The hardest part of judging fit is that the evidence goes stale. A survey from last spring and a retention chart someone built once in a spreadsheet will not tell you whether this month's cohorts are doing better.

OpenHelm's journey analyst maps your product's key user journeys from the product itself and its GA4 events, and files a task with the evidence wherever users stall. Funnels, retention and audience appear as charts and data tables that the agents keep updated, so you are looking at current cohorts rather than an old export. If the product has no analytics yet, OpenHelm files a task and can create the GA4 property and data stream and wire the measurement ID into your deployment. The survey and the qualitative log are still yours to run, but the behavioural half of the evidence stays up to date without anyone rebuilding a spreadsheet. See user funnels.