Product market fit: what it means and how to know you have it

The definition, from where it came from

The phrase was popularized by Marc Andreessen in a 2007 essay called The only thing that matters, and his definition is still the most useful one:

Product/market fit means being in a good market with a product that can satisfy that market.

Two halves, and founders usually only work on the second. A good product in a market that does not want it, or cannot pay, or cannot be reached, is not fit. Andreessen's argument in that essay is that the market is the part that dominates, which is uncomfortable, because the product is the part you control.

Practically, it means enough people want what you built that growth stops feeling like pushing. Before fit, every customer is won by effort. After fit, the market starts pulling, and your problems change from persuading people to keeping up with them.

What it feels like on each side

The same essay describes both states plainly, and the descriptions have held up for nearly twenty years.

Without it:

The customers aren't quite getting value out of the product, word of mouth isn't spreading, usage isn't growing that fast, press reviews are kind of "blah", the sales cycle takes too long, and lots of deals never close.

With it:

The customers are buying the product just as fast as you can make it, or usage is growing just as fast as you can add more servers.

The reason this matters more than it looks: the first description is not dramatic. Nothing is on fire. Everything is nearly working. That is exactly why founders stay in it for years, and why a measured read beats a felt one.

The signals that fool people

Most false positives come from a signal that is real but measures something other than fit.

• Signups. They measure your landing page and your channel, not your product. A signup who never returned tells you the promise worked, which is a different achievement.
• Compliments. People are kind, especially to founders they like. Enthusiasm in a conversation costs nothing.
• A big launch spike. A launch measures novelty and your network. The number that matters is what usage looks like three weeks later.
• One large customer. A single design partner can carry revenue and mask the fact that nobody else wants the generic version.
• Investor interest. Investors are assessing a market and a team, often before fit exists. Their enthusiasm is not evidence about your users.

The common thread is that every one of these can be true while the customers described above still are not getting value. The only signals worth much are retention and repeat use, because those are the ones a person cannot give you out of politeness.

Measure it instead of arguing about it

Because the felt version is unreliable, use a number that different people cannot interpret differently.

The most widely used is the one-question survey: ask active users how they would feel if they could no longer use the product, and read the share who answer very disappointed. The working bar is 40 percent. The full method, who to ask, and how to read the result by segment is in the product market fit survey guide.

Alongside it, watch two things the survey cannot tell you:

1. Retention that flattens. Every product loses users early. What matters is whether the curve levels off, which means a group has settled into using it, or continues toward zero, which means nobody has.
2. Unprompted repeat use. People coming back without an email telling them to.

None of the three is sufficient alone. Together they are hard to fool yourself with, which is the entire point.

How you get there: talk to people, in the right way

Fit is found more often than it is designed, and the finding is done in conversations.

Customer discovery is the practice of interviewing people in your target market about their problem rather than your solution. The discipline that separates useful interviews from flattering ones is asking about the past instead of the future.

Questions that work:

• What do you do about this today?
• Walk me through the last time it happened.
• What have you already tried, and why did you stop?
• What did it cost you when it went wrong?
• Who else is involved when you fix it?

Questions to avoid: would you use this, would you pay for this, does this sound useful. Every one invites a hypothetical answer, and hypothetical answers are consistently more positive than real behavior.

What you are listening for is evidence the problem is already costing them something: a workaround they built, money they already spend, or visible irritation. That is what a painkiller sounds like. "That would be nice" is a vitamin, and vitamins do not produce the pull described above.

The wider method for choosing which assumption to test first is in how to validate a business idea, and the market-side homework that stops you validating inside a market that cannot support you is in market research for startups.

What to do on each side of it

The practical value of knowing where you stand is that the right thing to do is almost opposite on either side.

Before fit, do not scale anything. This is the expensive mistake, and it is expensive precisely because it looks like progress. Spending on acquisition sends more people into a product that does not hold them, which burns money faster and uses up an audience you will want later. Hiring a sales team to push harder does the same with salaries. Every extra channel adds work that has to be redone once the product changes.

What to do instead is narrow. Pick the segment with the strongest signal, talk to more of exactly those people, and change the product or the audience until the number moves. Andreessen's advice in that essay was blunt about this: do whatever is required to get to fit, including changing out people, rewriting the product, moving into a different market, and telling customers no when you did not want to.

After fit, the bottleneck moves. The question stops being whether people want it and becomes whether you can reach more of them affordably and serve them without falling over. That is a channel and operations problem, and it is a genuinely better problem to have. The failure mode on this side is different too: it is scaling a channel that does not pay for itself, or letting quality slip while keeping up with demand.

The reason to measure rather than feel is that these two mistakes look similar from inside and cost very different amounts.

It is one segment at a time, and it does not stay put

Two things about fit that the phrase itself hides.

It is rarely universal. You do not have product market fit in general; you have it with a particular kind of customer doing a particular job. A blended survey score of 25 percent can be 55 percent in one segment and near zero elsewhere. That is not failure, it is a targeting instruction, and finding it is one of the most valuable outcomes of measuring at all.

It is not permanent. Markets move, competitors arrive, and the customers who loved the early version are not the ones who arrive next. Fit is a state you can lose, which is why the measurement is worth repeating rather than celebrating once.

If you are earlier than all of this and do not yet have a product to measure, the cheapest version of the same question is a real demand test: publish the offer, send the right traffic, and see who acts. How to run a fake-door test covers it, and it will tell you whether the market half of the definition is there before you build the product half.