What Is Product-Market Fit, Really?

Product-market fit is what you have when users would be upset to lose your product and keep coming back on their own. The most cited proxy is the Sean Ellis survey, where 40% or more of users say they would be “very disappointed” without the product. Stronger evidence is a retention curve that flattens instead of bleeding to zero.

The 40% test became famous because it gives a tired founder one number to chase. But fit is a behaviour, not an opinion. People tell a survey they would be “very disappointed” and then stop logging in.

Product-market fit, defined

Product-market fit = a defensible match between what your product does and a market that demonstrably wants it. The proof is users who retain (come back on their own) and, ideally, expand and refer. The Sean Ellis 40% test is one input. The harder evidence is:

1. A retention curve that flattens: cohorts reach a stable plateau instead of decaying to zero.
2. Cohort behaviour that holds or improves: newer cohorts retain at least as well as older ones.
3. Metrics that reinforce each other: retention, unit economics and any viral loop point the same way.

Where the Sean Ellis 40% test comes from, and what it misses

Sean Ellis ran growth at Dropbox, LogMeIn and Eventbrite. He noticed that the companies that took off shared a survey pattern: roughly 40% of users said they would be “very disappointed” without the product. Below that line, growth efforts mostly fizzled. Above it, they compounded.

That origin tells you where the test is weak:

• It is a stated preference. What people say in a survey and what they do with their calendar are different data sets.
• It surveys survivors. The users who already churned are not there to say they would not miss you.
• It is a snapshot. A one-time survey cannot show whether last quarter’s cohort is decaying.
• The 40% line is directional. It came from a handful of consumer-ish products. Your number depends on your market, your switching costs and how you sampled.

So run the survey. It is cheap, and the open-text answers show you your most-loved use case.

What product-market fit looks like in the data

1. The retention curve flattens

Plot the percentage of a cohort still active over time. Three shapes are possible:

• Decay to zero: retention keeps dropping month after month and never stops. No fit.
• Decay, then a plateau: retention drops at first, then flattens onto a stable line. That flat tail is a base of users who found durable value. This is the signature of product-market fit.
• Smiling curve: retention drops, flattens, then ticks up as dormant users return and the core deepens. It is rare enough that you should check your instrumentation before you celebrate.

What matters is whether the curve goes flat, not the height of any single point. A 25% plateau that holds is healthier than a 45% that is still sliding.

What “flat” looks like by month depends on your context. Read your own curve against these sourced reference ranges.

Retention benchmarks: B2B SaaS vs Consumer

B2B figures are cohort retention (Amplitude for Day 90; Day 1, Day 7 and Day 14 are estimates). Pendo’s returning-user rates (50 to 70% Day 1, 40 to 60% Day 7, 25 to 35% at 90 days) measure something else and are not comparable. Consumer sources: Adjust, Statista (Day 1); AppsFlyer, Amplitude (Day 7); Mixpanel, Amplitude (Day 14); AppsFlyer, Adjust (Day 90).

Check which metric a benchmark reports before you read the gap. Pendo’s B2B numbers look ten times higher because they count already-active users who come back, not a signup cohort.

On cohort retention the B2B and consumer Day-90 ranges overlap at the low end. What separates them is the shape of the curve and what “active” means. A workflow tool is used on workdays. A consumer app is used when the user feels like it.

Read the curve against the right reference class and the right metric. A B2B number judged against a consumer range will flatter or terrify you for no reason.

By Day 14 the drop should be flattening. Our benchmark tool reads a steep decline that has not levelled by then as a weak core loop. (For the curve-by-curve breakdown, see retention-rate-benchmarks.md.)

2. Cohort behaviour holds, or gets better

The stronger signal is comparing cohorts over time. Group users by the month they joined and plot each cohort’s retention separately.

• If newer cohorts retain as well as or better than older ones, your fit is real and improving.
• If each new cohort retains worse than the last, you are scaling acquisition faster than fit. Often you nailed an early niche and are now buying users from adjacent segments with a different job-to-be-done.

A single aggregate retention figure averages a loyal early cohort with a churning recent one and hides the divergence. Ask the fit question of each cohort, and keep asking it.

3. The metrics interact: fit is a system

Product-market fit shows up in how your metrics relate. The interactions:

• Retention is the foundation of unit economics. Lifetime value is mostly retention with a price tag; margin and expansion do the rest. A healthy LTV:CAC ratio is 3:1 to 5:1 for B2B (Bessemer, OpenView and a16z) and 2:1 to 4:1 for consumer (Adjust and AppsFlyer). If LTV looks great while the curve is still sliding, your LTV is a forecast resting on churn that has not finished happening. Check retention quality before you trust CAC efficiency. (More on the ratio and its traps in ltv-cac-ratio.md.)
• A “great” ratio can be a warning sign. An LTV:CAC above 5:1 often means you are underinvesting in growth. Read the ratio next to your CAC payback period. The range is 6 to 12 months for SMB and self-serve B2B (OpenView and KeyBanc). Longer is fine only with high retention and expansion.
• Virality without retention is a leak. K-factor (how many new users each user brings) is weak in B2B by nature, 0.1 to 0.3 per Reforge and Andrew Chen. In consumer it is stronger but rarely above 1, at 0.3 to 0.7. Andrew Chen has argued for years that the way to drive viral growth is to improve retention first.

When the signals disagree (great survey, leaking curve; great ratio, decaying cohorts), you have a flattering number and no fit yet.

Why benchmarks are context, not targets

A benchmark is a reference class, not a goal line. Good Day-90 cohort retention is 2.5% to 15.6% for B2B (Amplitude) and 1% to 4% for consumer. Your “good” depends on your business model, your contract length, your switching costs, and the job your product is hired to do.

You optimise to hit 40% on the survey, or drag Day-7 up to the middle of the range, and you are gaming the proxy instead of building the product.

The benchmark’s job is to tell you which questions to ask. Is my curve flattening at all? Are my newer cohorts holding? Do my economics rest on real retention or forecast retention?

How to assess your product-market fit

In order of how much each step tells you:

1. Plot your retention curve and look for the plateau. Does it flatten, or decay to zero?
2. Split it by cohort. Are newer cohorts holding up against older ones? Divergence here is an early warning the aggregate hides.
3. Compare against the right reference class. B2B to B2B, consumer to consumer. The slope (is it flattening by Day 14?) matters more than any single point.
4. Check whether your economics rest on that retention. Is your LTV:CAC built on the flat part of the curve, or on churn that has not finished? Is payback sane?
5. Then run the Sean Ellis survey, for the open-text answers as much as the 40%. It tells you who loves you and why, and that sharpens your ICP.

If the diagnosis comes back ugly, fix the curve before you fix the funnel. Sharpen the ICP so you stop acquiring users who were never going to stay. Deepen the core loop that drives repeat use. Only then spend on acquisition or virality.

See where your numbers land

The hard part is reading your own curve against the right reference class without fooling yourself. Our free benchmark tool does that. Paste in your retention by month and your unit economics.

It charts your curve against the B2B and consumer ranges and flags whether you are plateauing or still sliding. It also shows how your retention, K-factor and LTV:CAC relate, as context rather than targets.

Check your product-market fit signals → benchmark.scilla.studio

For the wider picture, see our full B2B SaaS growth benchmarks for 2026.