The Sean Ellis 40% Rule: Everything Founders Need to Know

Here's an uncomfortable truth: most founders can't tell you whether they have product-market fit. They have a feeling. A good week. A nice-sounding quote from a customer call. Then a churned account ruins the mood and the feeling flips.

Sean Ellis got tired of running on feelings. As the first marketer at Dropbox, LogMeIn, and Eventbrite, he kept being handed the same impossible job — "go make this grow" — and he noticed something. Sometimes growth tactics worked like magic. Sometimes the identical playbook did nothing. The difference wasn't the marketing. It was whether the product had already won people over.

So he went looking for a way to measure that before pouring money into growth. What he found became the most widely used product-market fit test in the startup world. And it comes down to a single question.

The one question that measures product-market fit

Ask your users this:

Give them three options: Very disappointed, Somewhat disappointed, and Not disappointed (a fourth — "N/A, I no longer use it" — keeps your data clean).

That's it. No 30-question NPS marathon. No focus group. One question that cuts straight to the thing that actually matters: would your users genuinely miss you if you disappeared?

Because that's what product-market fit really is. Not signups. Not vanity downloads. Not a polite "yeah, it's cool." It's a group of people who'd feel a real hole in their day if your product vanished. The "very disappointed" crowd are the people pulling your business forward — and this question finds them.

The 40% rule, and where the number actually came from

Ellis didn't pick 40% out of the air. He ran this survey across nearly a hundred startups and compared the results against which companies went on to grow sustainably. A pattern emerged with surprising consistency:

40% became the line. Above it, growth efforts compound — you're pouring fuel on a fire that's already lit. Below it, you're trying to light wet wood, and every dollar of acquisition leaks back out through the bottom as churn.

The genius of the threshold is what it does to your priorities. Below 40%, the answer is almost never "spend more on ads." It's "make the product more of a must-have for the people who already half-love it." The rule doesn't just measure fit — it tells you where to point your energy.

How to calculate your PMF score (the exact formula)

The math is deliberately simple. You want a number you can recalculate in your head and explain to an investor in one sentence.

Say you survey your active users and 70 of them give a valid answer. Of those, 28 say "very disappointed." Your score is 28 ÷ 70 = 40%. You're at the threshold.

Don't want to do the math by hand? Drop your numbers into the free PMF score calculator and see exactly where you land against the 40% line.

Two things founders constantly get wrong here:

• "Valid responses" excludes the people who no longer use the product. If someone picks "N/A," they don't belong in the denominator — they've already churned and they'd skew your score downward in a way that hides what's happening with active users.
• Only "very disappointed" counts in the numerator. "Somewhat disappointed" feels encouraging, but it's the most dangerous answer on the survey. More on that in a moment — it's where your roadmap hides.

Who to survey (this is where most scores go wrong)

A PMF score is only as honest as the people you ask. Survey the wrong crowd and you'll get a number that's either flattering nonsense or unfairly brutal.

The rule Ellis used, and the one we bake into PMFtracker: survey people who have experienced the core of your product at least twice in the last two weeks.

Why "experienced the core"? Because someone who signed up, poked around the dashboard, and never came back can't tell you anything useful about fit. They never reached the moment your product is supposed to deliver. Their disappointment (or lack of it) is noise. You want the verdict of people who actually used the thing for what it's for.

The two-sided risk

Survey only your superfans — your beta community, your Slack regulars — and you'll get an inflated 60% that falls apart the moment real users arrive. Survey everyone who ever signed up, including the tire-kickers who bounced in week one, and you'll bury your true signal under a pile of indifferent "not disappointed" answers. The engaged-user filter is what keeps the score honest in both directions.

How many responses do you actually need?

You don't need thousands. You need enough that one loud customer can't swing the result.

The practical takeaway: don't wait for a "big enough" user base to start. If you have 40 engaged users, survey them this week. A directional 32% measured today beats a perfect score you never get around to running. And once you're measuring, the number becomes something you watch climb — which is the entire point.

What your score actually means

A number on its own is just a number. Here's how to read where you land.

Below 40%: you're not there yet — and that's useful

Most startups live here, including ones that go on to win. A 25% score doesn't mean your product is bad. It means a quarter of your users would genuinely miss you, and your job is to figure out what they have in common and make more of those people. The score below 40% is a starting line, not a verdict. (We'll get to exactly how to climb it.)

Around 40%: the threshold

You've crossed into "fit." A meaningful slice of users depend on you. This is when growth investment starts to pay back instead of leaking away — and when the score becomes a credible data point in a fundraise.

Above 50%: strong, must-have territory

This is rare and powerful. Superhuman reached 58%. Products in this zone tend to grow by word of mouth because users can't shut up about them. If you're here, the question shifts from "do we have fit?" to "how do we not lose it as we scale?"

Real benchmarks: Slack, Superhuman, and the trajectory

The number gets a lot less abstract when you see what real companies scored.

Notice Superhuman's starting point. 22%. Just over half the threshold. By the "magic moment" theory of PMF, founder Rahul Vohra should have concluded he didn't have it and moved on. Instead he treated 22% as a baseline to improve — and walked it up to 58%. That story is the single best argument for why you measure: the score isn't a judgment, it's the first data point in a process.

We broke down exactly how Superhuman did it here →

The mistake hiding in "somewhat disappointed"

Here's the part almost everyone skips. Your score only counts the "very disappointed." But the "somewhat disappointed" group is the most valuable data on the whole survey — because they're your growth lever.

These people like your product. They use it. But something is holding them back from loving it. Buried in their open-ended answers is a roadmap: the missing feature, the rough edge, the use case you half-serve. Move even a chunk of "somewhat" into "very," and your score climbs toward 40% without acquiring a single new user.

Meanwhile, the "not disappointed" group is telling you who to ignore. They're often the wrong audience entirely — people who were never going to love what you built. Chasing them is how good products lose their focus. The survey, read properly, tells you both who to delight and who to stop trying to please.

Why a one-time score lies to you

The biggest misunderstanding about the 40% rule is that PMF is a moment you arrive at — a door you walk through once and you're done.

It isn't. Product-market fit is a number that moves. You can lose it. A pivot, a redesign, a flood of new users from the wrong channel, a competitor who raises the bar — any of these can quietly drag your score back down while your dashboard still looks fine. The founders who keep fit are the ones who keep measuring, the same way you'd never check your revenue once and assume it holds forever.

This is exactly why a survey buried in a Typeform export and a Google Sheet you ran once last quarter is worse than useless — it gives you a stale number and the false confidence that goes with it. A score is only an asset if you can see it move.

From score to story: what investors want to see

A single PMF percentage is good. A trend line is fundraising gold. When a VC sees "we went from 28% to 44% over two quarters, here's the ICP driving it, here's what the very-disappointed users say they'd miss," you've handed them evidence — not vibes — that you understand your own traction.

That's the difference between "we think people love it" and "here's the longitudinal data proving demand, segmented by who and why." One gets a polite nod. The other moves a term sheet. Our complete measurement guide walks through how the survey score pairs with retention to make that case airtight.

Putting it to work this week

You don't need a research team or a perfect dataset. You need to start:

1. Define your engaged users. Anyone who hit the core of your product 2× in the last two weeks.
2. Ask the one question with three options, plus an open-ended "What's the main benefit you get?"
3. Collect at least 40 responses for a read, push toward 100+ to trust it.
4. Calculate the score — very disappointed ÷ valid responses.
5. Read the "somewhat" answers for your roadmap, and the "very" answers for your ICP.
6. Run it again next month. The trend is the asset, not the snapshot.

That loop — measure, segment, improve, re-measure — is the entire game. The 40% rule just gives you a scoreboard so you can finally tell whether you're winning.