The Sean Ellis test doesn't care what you sell. One question, three answers, one formula: the share of engaged users who'd be very disappointed to lose the product. A B2B analytics tool and a B2C fitness app run the identical survey.
But two founders with a 40% score on the same benchmark can be looking at very different levels of confidence, because B2B and B2C differ on the two things that make a PMF score trustworthy: the size and shape of the sample, and who's actually answering.
The sample math is genuinely different
Most B2C products reach thousands of users before their first funding round; most B2B products don't. That single fact changes the entire exercise:
- B2C: the challenge is usually filtering, not counting. You likely have far more than 40 or 100 total users; the work is excluding one-time visitors and casual dabblers so the survey only reaches people who've genuinely engaged with the core product, exactly as covered in how to run a PMF survey.
- B2B: the challenge is usually volume. If your entire customer base is 60 companies, hitting 40 valid engaged-user responses might mean surveying nearly every account you have, which takes longer and makes the response-rate playbook non-optional rather than nice-to-have.
This is exactly why the PMF benchmarks by industry vary so much sector to sector: a benchmark built mostly from B2C consumer apps doesn't map cleanly onto a 40-account B2B tool, and vice versa. The 40% threshold generalizes; the practical difficulty of getting a clean read on it does not.
Who actually answers isn't the same person
In B2C, the person using the product and the person who decided to use it are almost always the same human. Their "very disappointed" answer is a direct, unmediated read on the product's value to them.
In B2B, that link frequently breaks. A 40-seat account might have one power user who lives in the tool daily, a manager who approved the purchase and opens it twice a month, and 30 people who were added and never logged in. If you survey by account instead of by actual usage, you'll hear from whoever happens to reply, not from the people whose answer actually reflects product value. That's why B2B PMF surveys have to filter on individual usage, not seat assignment or account status, before the question is even sent.
The consequence for segmentation: in B2B, weight your "very disappointed" segment by account diversity, not just respondent count. Ten very disappointed answers from ten different companies is a far stronger signal than ten from three companies where one account happened to have chatty users. In B2C, weight by behavioral cohort instead, since account-level clustering usually doesn't exist.
The feedback loop runs at different speeds
B2C products iterate on usage data measured in days. A change ships, retention shifts within a week or two, and the next PMF survey a month later reflects a genuinely different product experience.
B2B sales and adoption cycles run longer, often months. Survey too frequently and you're mostly re-measuring the same account relationships and the same unresolved procurement-driven friction, not fresh signal. That's the practical reason a quarterly cadence, discussed in how often to run the PMF survey, fits most B2B products better than the faster loop that works for consumer apps.
Two companies known for treating their survey score as an engine, Slack (51%, 731 users) and Superhuman (22% to 58%), are both fundamentally B2B or prosumer-B2B products: a defined, surveyable user base, and enough time between iterations for the number to mean something new each round. UX Pilot's 49% score came from 303 paid B2B SaaS users, a sample size that would be trivial for a consumer app with a million signups and genuinely hard-won for most B2B tools at that stage.
See what a good score looks like for your sector
PMF benchmarks vary by industry and business model. Check where your score sits against SaaS, fintech, marketplaces, and more before you compare yourself to a number that isn't measuring the same thing.
See PMF benchmarks by industry → Free · no signup requiredWhat to actually do differently
- B2B: define "engaged" at the individual-user level inside each account, survey nearly your whole engaged base if you have to, weight results by account diversity, and run the survey quarterly.
- B2C: spend your effort filtering out casual and dormant users rather than chasing volume, segment by behavior instead of account, and consider a monthly or per-release cadence if your iteration speed supports it.
- Both: the 40% line, the formula, and the discipline of tracking the score as a trend instead of a one-off number don't change. Only the mechanics of getting a sample worth trusting do.
Track PMF the way your business model actually works
PMFtracker segments respondents automatically, whether that means grouping by account for B2B or by behavior for B2C, so your ICP reflects how your business actually sells and grows.
Start Tracking PMF → Set up in 5 minutes · No credit card required
