Quick answer
PMF stands for product-market fit: the degree to which a product satisfies strong demand in a specific market, so customers keep using it, pay for it and recommend it. It is not a single moment but a set of signals, such as a flattening retention curve and growing organic demand.
What does PMF mean?
In business, PMF stands for product-market fit: the degree to which a product satisfies a real, strong demand in a specific market, so that customers keep using it, pay for it and bring others without being pushed. (In statistics, PMF also means probability mass function; this entry is about the startup term.)
The common mistake is to treat product-market fit as a moment, a day when a startup "hits PMF". In practice it shows up as a set of signals you can keep observing: retention curves that flatten instead of sliding to zero, a growing share of new users arriving organically, and people paying without discounts. Survey scores are useful context, but on their own they tend to overstate fit. Fit can also weaken: a market moves, a competitor arrives, or you expand to a segment the product does not serve.
Where the term comes from
The phrase was popularised by Marc Andreessen in his 2007 essay "The only thing that matters", where he defined it as "being in a good market with a product that can satisfy that market." He credited the formulation to investor Andy Rachleff and argued that market matters more than team or product.
Andreessen described the difference in behaviour, not in a score. Without fit, customers do not quite get value, word of mouth is weak, usage grows slowly and sales cycles drag. With fit, demand outruns the company's ability to serve it: customers buy as fast as you can deliver, money piles up and the company hires sales and support as fast as it can. Those are observable signals, which is why the data-driven definition below still follows his essay.
Signals of product-market fit
| Signal | Type | What to look at | How it misleads |
|---|---|---|---|
| Retention curve flattens | Quantitative | Share of a signup cohort still active in week 8, 12, 26 | A curve that is still falling looks "flat" if you read absolute points instead of relative drops |
| Organic share of new users rises | Quantitative | New users from direct, organic search and referral as a share of all new users | Paid campaigns paused at the same time inflate the organic share |
| Paying without discounts | Quantitative | Share of paid conversions at list price; renewals | Launch discounts and lifetime deals buy revenue, not fit |
| Referral and word of mouth | Quantitative | New users referred by existing ones, mentions, branded search | One viral post spikes it once |
| Usage depth | Quantitative | Key action done weekly by retained users | Counting logins instead of the core action |
| "Very disappointed" survey | Qualitative | Share of users who would be very disappointed without the product | Survey only engaged users and the score rises |
| Sales and support pull | Qualitative | Inbound demand exceeding capacity; feature requests converge | Loud requests from a few large customers |
The detailed metric definitions and benchmarks are in product-market fit metrics; this entry focuses on how to read the signals together.
Worked example: does the retention curve flatten by week 8?
Sample data: two products each sign up a weekly cohort of 1,000 users. The table shows the share of each cohort active in week N.
| Week | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|
| Product A | 42% | 33% | 28% | 25% | 23% | 22% | 21.5% | 21.2% |
| Product B | 38% | 27% | 20% | 15% | 11% | 8% | 6% | 4.5% |
Decision rule: the curve has flattened if each of the last three week-over-week drops is under 5% of the previous week's value.
- Product A, week 6: (23 − 22) ÷ 23 = 4.3%; week 7: (22 − 21.5) ÷ 22 = 2.3%; week 8: (21.5 − 21.2) ÷ 21.5 = 1.4%. All under 5%: flattened. Of 1,000 signups, about 212 are still active in week 8, and the curve suggests most of them will stay.
- Product B, week 6: (11 − 8) ÷ 11 = 27.3%; week 7: (8 − 6) ÷ 8 = 25.0%; week 8: (6 − 4.5) ÷ 6 = 25.0%. Not flattened. About 45 of 1,000 are left, and if the 25% weekly decay continues, week 12 is 4.5 × 0.75⁴ = 1.4%.
The trap is in Product B's absolute drops: 3, 2 and then 1.5 points look like the curve is calming down. In relative terms it still loses a quarter of its remaining users every week. A product with B's curve has no fit in this segment, however well its signup numbers grow, because every cohort heads towards zero. Read the retention curve in relative drops, and confirm with the other signals in the table.
Before and after product-market fit: what changes
| Before PMF | After PMF | |
|---|---|---|
| Main question | Who loves this and why? | How do we reach more people like them? |
| Metric to watch | Retention of each cohort; activation | Acquisition cost, payback, expansion |
| Marketing spend | Small, to find signal | Scaled on channels that already convert |
| Product work | Change the core loop, cut features, narrow the segment | Reliability, onboarding at scale, adjacent features |
| Team | Small; founders sell and support | Hire sales, support and growth roles |
| Risk | Scaling too early burns cash on users who leave | Losing fit as the market or segment shifts |
Spending heavily on acquisition before the curve flattens fills a leaky bucket: each new cohort decays like the last. Growth models such as AARRR and product-led growth assume the retention stage already works; for the playbook after fit, see SaaS growth strategy.
Common mistakes
- Calling PMF from one survey. The Sean Ellis test asks "how would you feel if you could no longer use the product?", and 40% "very disappointed" is the benchmark Ellis drew from comparing about a hundred startups. Superhuman's founder reported starting at 22%. The score is useful for finding your most engaged segment, but it depends heavily on who you survey; pair it with retention.
- Reading growth as fit. Signups can grow on paid spend while every cohort churns.
- Averaging segments. A product can fit one segment (small agencies) and not another (enterprises); a blended curve hides both.
- Treating fit as permanent. Re-check the signals after pricing changes, new markets and big competitor moves.
- Counting discounted revenue. Lifetime deals and steep launch discounts show price sensitivity, not demand at your real price.
Checking PMF signals in TapCub
Here is the week-8 test from the example, run in TapCub. TapCub does not run surveys, so the Sean Ellis question needs a separate survey tool; the behavioural signals are all here.
- Build the curve. In All-platform analytics › Behavior › Behavior analysis, choose retention, set the period to week and show 12 weeks. Use "returned on period N exactly" to read each week's share, and set the start event to sign-up and the return event to your core action, not any activity. Each row is a cohort with its size and weekly rates; the line under the table is the size-weighted average. Apply the 5% rule to that average and to the newest cohorts. See build a retention report.
- Watch the mix of users. Switch to lifecycle, weekly. The four series, New, Returning, Resurrecting and Dormant, show whether returning users grow as a base or whether each week's new users simply replace those going dormant. See read the lifecycle report.
- Check organic and referral growth. In All-platform analytics › Business › AARRR growth, acquisition shows new users by channel, and referral shows referred new users (those whose channel is social or referral) with K = referred new users ÷ active users. Web analytics › Sources splits visitors by channel if you want the organic share week by week.
- Segment before you average. Define the segment you think fits (for example, sign-ups from agencies) in All-platform analytics › Users › Segments & tags and apply it to the retention report, so one segment's flat curve is not hidden in a blended average.
TapCub does not do billing either; payment signals come from the purchase events you send, which feed All-platform analytics › Business › Commerce & revenue.
See also: What is product differentiation? Definition, four kinds and how to check it is real.
Sources
- Marc Andreessen, The only thing that matters, 2007.
- Rahul Vohra, First Round Review, How Superhuman Built an Engine to Find Product/Market Fit, 2018.
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Product Adoption Consultant
Explores customer onboarding, live conversations, and self-service, turning common usage questions into practical guidance and improvements.