Product-market fit is the moment when a product solves a problem well enough that a clear group of customers uses it, comes back, pays and recommends it to others — without constant persuasion. The term was popularised by Marc Andreessen in 2007, when he wrote that it is the only thing that really matters for a new product.
Before fit, every customer is fought for. After it, some customers arrive on their own, and the main problem becomes keeping up with demand.
What a product looks like without fit and with fit
| Without product-market fit | With product-market fit |
|---|---|
| Users sign up and vanish after a few days | A clear share of users come back week after week |
| Every sale takes long persuasion | Customers ask on their own, and sales conversations are shorter |
| Growth only when you pay for ads | Referrals and traffic without ads appear |
| Feedback is polite but generic ("nice idea") | People complain when something doesn't work, because they need the product |
| The team adds features, hoping one will work | The team can't keep up with requests for specific improvements |
How to measure it
There is no single number that settles it. There are several signals worth looking at together.
Retention in cohorts. Take the people who started using the product in a given week and check what share is still active after 1, 4 and 8 weeks. If the curve drops to zero, there is no fit. If it flattens out at some level, you have a group that needs the product. What level is good depends on the kind of product: a tool for daily work should have higher retention than an app used once a month.
The Sean Ellis test. You ask active users one question: "How would you feel if you could no longer use the product?" with the answers: very disappointed, somewhat disappointed, don't care. Ellis proposed a threshold: if at least 40% answer "very disappointed", it's a good sign. Ask only people who have really used the product, and treat the result as one of the signals, not a verdict.
Referrals and organic traffic. Where do new users come from? If more and more of them arrive through a recommendation or search for you by name, the product is starting to spread on its own.
Money. Do people pay, do they keep paying after the first period, and is the cost of acquiring a customer clearly lower than what the customer spends over their whole time as a user. Fit without a profitable model is only half of the answer.
What to measure after launching an MVP
After the MVP goes live, set up measurement from day one, before you start inviting people:
- Activation — define one action after which the user got value (for example, sent the first reminder, placed the first order) and measure what share of new people reach it.
- Retention — weekly cohorts, as above.
- Payment — how many people move from a trial to a paid plan and how many stay after the first month.
- Conversations — every week a few short conversations with active users and with those who left. The numbers tell you what is happening, the conversations tell you why.
You don't need elaborate analytics right away. Well-configured events and a simple spreadsheet with cohorts are enough. What matters is to measure behaviour, not page views.
What to do if you don't have fit yet
This is a normal state for a new product. Instead of adding features, check in turn:
Is the problem important enough? Maybe people have this problem, but not badly enough to change their habits. Then only a different problem or a different customer group will help.
Are you aiming at the right group? Among your users there is often a small group that uses the product intensively and a large one that drops off. Narrow the product down to the former. A narrow group of delighted customers is a better start than a broad group of indifferent ones.
Does the product deliver value quickly? If it takes ten steps to the first result, people will drop off before they see it. Shorten the path to the first value.
Does the payment model fit? A subscription where someone uses the product once a year, or a one-time fee for something used daily, can kill a good product.
Each of these changes is another turn of the loop in the lean startup methodology, and sometimes a deliberate change of direction. Write it down on the lean canvas so you can see what you've already tested.
When to scale
Only when the signals are confirmed: retention stable, people paying and recommending, and the cost of acquiring a customer adding up. Scaling before fit only speeds up spending: more ads for a product people leave means more people who will leave.
If you are at the stage of the first version and want to quickly reach the point where you can measure all of this, see how I build an MVP: 2–6 weeks, a fixed price, measurement set up from the start.