Product-market fit is the most cited phrase in B2B and the least useful, because almost nobody uses it to mean what it measures.
Ask a founder if they have product-market fit and they will tell you whether people like the product. Whether the demos go well. Whether the market gets it. None of that is product-market fit. Product-market fit is retention. It is measured after the fact, in cohorts, in lifetime value — in whether people who paid you once keep paying you. It is a number you read off the past. It cannot be aimed at, because by the time it can be measured, every decision that produced it has already been made.
So when a company says it is searching for product-market fit, it is searching for a reflection. The thing it is actually searching for is one of two other fits — and knowing which one is the difference between fixing the company and burning the runway.
There are three fits. Two of them are things you build. The third is what happens when you got the first two right.
1. Problem-market fit
Does the market want the outcome — and will it say so out loud, before you have proven anything?
Problem-market fit asks whether the problem you are solving is one the market already carries. Not whether they would agree it is a problem if you explained it well. Whether they already have it, already recognize it, already tried to solve it, and already have scar tissue from the attempts.
It is the cheapest fit to test, and almost nobody tests it, because it can be tested without building anything at all. A phone and a list will do it. The signals are behavioral and they are loud: Do they take the call? Does the outcome land without translation — can you say it with no product nouns in the sentence? Can they name what they already tried, and what it cost them? Do they book? How far into the conversation do they go before the interest dies?
This is why problem-market fit failure is loud. Nobody answers. Nobody books. The calls die in the first ten seconds. It hurts, but it tells you something immediately, and it tells you cheaply.
And here is the part that matters: problem-market fit is not a verdict on your company. It is a location. Most companies arrive here backwards — they built a capability first and are now searching for someone who needs it. That is a recoverable position. The lens does not say your product is dead. It says: you are holding a capability with no named outcome. Go find the market where that capability produces an outcome someone already wants. That is not a death sentence. It is a work order.
2. Prototype-market fit
Can you actually deliver that outcome — reliably, at a cost below what it commands, across the customer lifecycle?
This is the fit nobody names, and it is where good ideas die silently.
Problem-market fit tells you the outcome is wanted. It tells you nothing about whether you can produce it. And producing it has three separate ways of failing: you cannot deliver it at all — a capability problem; you can deliver it but not below your cost of acquisition — a margin problem; you can deliver it once but not consistently over time — a consistency problem.
Every one of these is a prototype problem. Not a marketing problem, not a sales problem, not a we-lost-product-market-fit problem. The prototype — the actual mechanism by which the promised outcome is produced — cannot carry the promise the market already accepted.
And this is why it kills quietly. The calls go well. The demos go well. The deals close. Every leading indicator is green, because problem-market fit is intact — the market genuinely wants the outcome and is genuinely willing to pay for it. The failure only appears downstream, in churn, in support load, in a margin that never turns, in a delivery timeline that keeps slipping past the point where the customer expected to see what they bought.
By the time it shows up, it looks like a product-market fit problem. It is not. It is a prototype problem wearing product-market fit clothes.
3. Product-market fit
What happens when the first two are true.
Product-market fit is not a third thing to achieve. It is the consequence of solving problem-market fit and prototype-market fit — the outcome is wanted, and the outcome is delivered, so people stay, so retention exists, so lifetime value exists, so the numbers everyone quotes as PMF show up in the cohort chart.
Retention is a readout. You do not chase it. You produce the two conditions that generate it, and then you read it. This is why chasing product-market fit directly feels like fog. You are aiming at an effect and wondering why the cause will not move.
The lens, in practice
The three fits are not a ladder you climb once. They are a diagnostic — a way of locating where a company actually is, and therefore which problem it actually has.
I have a capability. I have no customers, and I do not know who this is for. That is a problem-market fit phase. Who does this serve? What outcome do they already want? Can you say it with no product nouns? If you cannot name the outcome, that is the diagnosis. Not a marketing problem — a problem-market fit problem.
I cannot get anyone to take the call. Nobody books. That is problem-market fit failure, loud version. The outcome is not recognized, or it is not wanted, or the market you picked does not carry the scar. Change the market or change the outcome. Do not change the pitch.
Selling is easy. They buy fast. They churn in a month. That is a prototype problem. The outcome sold; the outcome did not arrive. Ask: how long before a customer sees the outcome? If the honest answer is we have not seen anyone get it yet — that is the whole diagnosis, and no amount of sales work will fix it.
Customers come. Customers stay. We are still not profitable. That is a prototype problem, margin flavor. Delivery costs more than the outcome commands. This is not a pricing problem you can solve with a pricing page. It is a delivery-mechanism problem that happens to express itself as price.
They stay, they pay, the cohorts hold. You have product-market fit. Not because you chased it. Because you solved the other two.
Why the order is inverted almost everywhere
Problem-market fit is cheap to test. Prototype-market fit is expensive to test. Testing whether the market wants an outcome requires a phone. Testing whether you can deliver it requires you to actually build the thing.
And the industry does it in exactly the wrong order. Build first. Sell after. Discover the problem-market failure with the runway already spent — and then, because the vocabulary is wrong, call it a product-market fit problem and go looking for a better pitch.
Put the cheap test first. Validate that the outcome is wanted, in live conversation, against real scar tissue, before the expensive test begins. Then build the prototype knowing exactly what it has to deliver, how fast, and under what cost ceiling — because the market already told you. That is not a growth hack. It is just doing the cheap experiment before the expensive one.
The honest boundary
None of the three can be fully known in advance. You learn delivery cost by delivering. You learn retention by retaining. Even problem-market fit gives you confidence, not certainty — it predicts conversation quality, not conversion rate, because conversion carries factors the topic never touches.
That is not a weakness of the lens. It is the point of it. The lens does not promise foresight. It promises location — it tells you where you are standing, which failure you are actually looking at, and therefore which experiment is worth running next. Most companies die not because they hit a wall, but because they spent two years fixing the wrong one.
Three fits. Two you build, one you read. Know which one is broken before you spend anything fixing it.