Product-market fit: what it is and how to measure it

Product-market fit is the point at which a big enough market needs your product badly enough to keep paying for it and to bring you new customers on its own. Before it, you push every sale up the hill yourself. After it, in Marc Andreessen’s description, demand runs ahead of your ability to deliver.

You can measure it, roughly, in two ways: a one-question survey and a retention curve. We’ll try both on a made-up accounting-software startup in Padua. Still at the idea stage? Our map from zero to a first round shows where fit comes in.

What is product-market fit?

The phrase comes from ‘The only thing that matters’, an essay Marc Andreessen published on 25 June 2007. His definition fits in one line: ‘Product/market fit means being in a good market with a product that can satisfy that market.’

He also described how it feels. Without fit, customers don’t quite get value, word of mouth doesn’t spread, the sales cycle drags and lots of deals never close. With it, the problem flips: you can’t add capacity fast enough, and customer money builds up in the bank. He thought you could always feel which side you were on.

Fit belongs to a market, not to a product on its own. Software that clicks with 20-person accounting firms in Padua can leave solo practitioners cold. And fit found in Italy doesn’t carry over to Germany by itself.

How do you measure product-market fit?

1. The 40% survey. Sean Ellis, who benchmarked nearly a hundred startups, asked users how they would feel if they could no longer use the product, and set the bar at 40% answering ‘very disappointed’. Ask only people who’ve really used it. Superhuman sent the question to users who had used the product at least twice in the previous two weeks, and found results turn directionally correct at around 40 responses.
2. A retention curve that flattens. For each monthly cohort, plot the share of users still active in month 1, 2, 3 and onward. Sequoia’s data science team is blunt: without fit, ‘the retention curve will continuously decline’. For a benchmark, Lenny Rachitsky and Casey Winters put good six-month retention at about 60% for SMB and mid-market SaaS, and about 40% for consumer SaaS.
3. Growth you don’t pay for. A rising share of new users arrives through referrals and word of mouth. Switch the ads off for a week and see what’s left.
4. Shorter sales cycles. In B2B, deals close faster, with fewer discounts and fewer pilots that go nowhere.

A worked example: a survey of 120 users

Say you run a startup in Padua that sells practice-management software to accounting firms. Of your 300 active users, 180 logged in at least twice in the past two weeks, and they’re the ones who get the Ellis question by email. Two in three answer: 120 replies, plenty to work with.

Segment	Replies	‘Very disappointed’	Share
Firms with 5+ staff	50	26	52%
Firms with 2 to 4 staff	40	12	30%
Solo practitioners	30	4	13%
All users	120	42	35%

Overall you score 35%, under the line. The first row tells a different story: firms with five or more staff are at 52%. The fit is there, in one segment.

Superhuman found itself in a similar spot with its 22%. It looked hard at the users who’d be very disappointed, asked what they valued most, and split the roadmap in two: half doubled down on what those users loved, half fixed what held back the ‘somewhat disappointed’ users who cared about the same thing, which was speed. Within three quarters the score reached 58%.

For your Padua startup, the next quarter might mean dropping solo practitioners from the ads, rewriting the homepage for firms of five or more, and running the survey again after 90 days. Plot those firms’ cohorts on their own, too. If their curve flattens while the others keep sliding, you know where to dig.

Signs you don’t have product-market fit yet

• Your Ellis score stays under 40% after several rounds, in every segment you try.
• Cohorts keep shrinking month after month, and not one has levelled off.
• In a week without search ads, new sign-ups fall to almost nothing.
• When a firm finally signs after a long pilot, it’s usually because you offered a discount.
• The accountants call the software useful when you talk to them, but the login data a few weeks later shows many of them have gone quiet.

When most of them still describe your company after months of honest work, the problem tends to sit in who the product serves or what it does, and adding features rarely fixes that. That’s usually when founders start weighing a startup pivot. Our piece on why startups fail covers the warning signs in more detail.

What seed investors ask about fit

A seed investor won’t necessarily expect proof of fit. They will want to see that you measure it honestly. Be ready for questions like these:

• What share of users would be very disappointed without you, and in which segment?
• Can I see retention by monthly cohort? Where does the curve flatten?
• How many new customers came last quarter without paid acquisition?
• Is a sale quicker to close now than it was six months ago, and by how many weeks?

The cohort table and the raw survey answers will do more for you in that meeting than a slide that just says ‘strong PMF’, which convinces nobody. Our lists of startup metrics investors want at pre-seed and seed and our guide to churn rate show how to present them. If you plan to raise in the next few months, the investor readiness test is a sensible first step: it asks 12 questions and emails you a score and an action plan.

Related terms

• You’ll usually start looking for fit with a minimum viable product, since you need a first version in real users’ hands before there are any survey answers or cohorts to read.
• When fit starts to show, it reaches investors as traction, which covers the revenue, retention and usage figures they’ll ask you for.
• Nothing built yet? Then you’re a step earlier, validating the startup idea by testing the problem rather than the product.
• If all you need is the short definition, it’s in our startup glossary under product-market fit, with 49 other terms.