Sean Ellis PMF Test Calculator
Score the 40% rule "very disappointed" test with a Wilson 95% confidence interval, ICP-weighted segmentation, sample-size confidence tier, and an auto-generated ICP paragraph.
Last reviewed: April 2026
What the Sean Ellis PMF Test Actually Measures
The sean ellis pmf test calculator scores one of the most durable product-market fit signals in SaaS: the "very disappointed" test. Sean Ellis — who led early growth at Dropbox, LogMeIn, and Eventbrite before founding GrowthHackers — noticed a pattern. Products with over 40% of users answering "Very disappointed" to "How would you feel if you could no longer use [product]?" were the ones that went on to scale sustainably. The very disappointed test calculator makes that math explicit: count the "Very" responses, divide by the valid sample (excluding N/A), and compare to the 40% threshold.
The test works because it measures loss aversion, not satisfaction. Users who would be very disappointed have formed a dependency — a habit, a workflow, a relationship. Satisfaction surveys (NPS, CSAT) measure fondness. The Sean Ellis test measures pain-of-loss, which is the only feeling strong enough to predict retention and viral growth.
The 40% Rule: Where It Came From and Why It Matters
Ellis first published the 40 percent rule pmf calculator framework on the Qualaroo blog in 2009 after running the survey across dozens of pre-PMF and post-PMF startups. The cutoff was empirical — 40% was the inflection point above which startups could sustainably grow organically, and below which paid growth masked a retention problem. The rule survived the decade because it captures a real threshold in user psychology: at 40%, enough users have formed dependencies that word-of-mouth compounds faster than churn.
The pmf score calculator in this tool treats the 40% line as the headline but adds statistical rigor around it. A 42% score with n=30 is reported as "Low Confidence" because the 95% confidence interval is 27–58% — statistically indistinguishable from a 38%. A 42% score with n=200 is reported as "PMF Confirmed" with a ±7pp CI.
How to Run the Survey: The Exact 5 Questions
The sean ellis survey questions template uses one primary question and four open-ended follow-ups:
- How would you feel if you could no longer use [product]? — Very disappointed / Somewhat disappointed / Not disappointed / N/A — I no longer use it
- What type of people do you think would benefit most from [product]?
- What is the primary benefit you have received from [product]?
- What would you use as an alternative if [product] were no longer available?
- How can we improve [product] for you?
Send to users who have completed at least one core action — the full population pollutes the N/A bucket with inactive users. Response targets: 100+ responses is reliable, 200+ is investor-grade. Surveys fielded via Sprig, Typeform, Maze, or SurveyMonkey can collect this volume inside a week for most SaaS products.
Scoring the Results Correctly (Exclude N/A, Weight by ICP)
The how to score sean ellis pmf survey convention excludes N/A responses from the denominator. Including them would systematically penalize every product with an onboarding dropout, making the score uncomparable across companies. The sean ellis survey analysis also benefits from segmentation — scoring ICP and non-ICP cohorts separately. A product with 45% ICP score and 15% non-ICP score has clear PMF with the target audience; the blended number would misleadingly show ~30%.
This tool lets you enter ICP and non-ICP counts separately, then weights the composite by ICP share (default 80%). The pmf score confidence interval around the final number uses Wilson's score formula, which is more accurate than the Wald approximation at small samples or scores near 0% or 100%.
Sample Size and Confidence: Why n=40 Is a Minimum
The sample size for pmf survey threshold is not arbitrary. Below n=20 the 95% CI on a 40% score is roughly ±22pp — the "score" could be anywhere from 18% to 62%. At n=40 it tightens to ±15pp (directional). At n=100 it is ±10pp (reliable for shipping decisions). At n=200 it reaches ±7pp (investor-grade). At n=400 you hit academic rigor with ±5pp.
The sample-size ladder in this tool lights up your current rung so you know the confidence class of your number before you report it. If you are prepping a Series A deck, get to n=200 before citing the score in the investor materials.
Benchmarks by Startup Stage
The pmf benchmark score by stage distribution — calibrated from First Round Review's founder surveys and the Superhuman case study data — is: pre-seed median 18%, seed median 24%, Series A median 33%, Series B+ median 40%. The 40% threshold aligns roughly with Series A readiness, but does not require it — many great Series A companies raise at 32–38% with strong ICP lift. The top-quartile ("p75") at each stage is: pre-seed 28%, seed 34%, Series A 42%, Series B+ 48%. This tool positions your score against the stage you select so you can tell whether you are leading, on-pace, or lagging the median.
The Superhuman Case Study
The superhuman pmf engine calculator framework comes from Rahul Vohra's 2018 First Round Review essay. Superhuman's first Sean Ellis survey scored 22% — clearly pre-PMF. Vohra's team then identified the traits of the "very disappointed" segment, built personas around them, and shipped features targeted to that cohort exclusively while ignoring the "not disappointed" segment entirely. Within 12 months the score crossed 40%; within 24 months it reached 58%. The lesson embedded in this tool's ICP Extractor: the traits of your "very disappointed" users ARE your ICP, and doubling down on them is the fastest path to crossing the 40% line.
Using the "Very Disappointed" Cohort as Your ICP
The icp extractor from pmf survey approach treats the "very disappointed" respondents as your highest-signal population. Pull their self-described roles, company sizes, alternatives, and must-have features into a concentrated ideal-customer paragraph. This paragraph becomes the North Star for positioning, acquisition targeting, and product prioritization. The calculator above generates this paragraph automatically when you fill the four open-ended fields — top benefit, primary alternative, must-have feature, and ideal customer description.
The ICP paragraph is the single most-valuable artifact from the Sean Ellis process. It beats brainstorming personas in a whiteboard session because it is sourced directly from users who already love you — no speculation required.
Frequently Asked Questions
What is the Sean Ellis PMF test?
The Sean Ellis PMF test is a single-question survey: "How would you feel if you could no longer use [product]?" with four response options. If at least 40% of active users answer "Very disappointed", the product is considered to have product-market fit. Ellis identified this threshold across dozens of early-stage startups and popularized it via Qualaroo and GrowthHackers around 2009.
How do you calculate the 40% rule for product-market fit?
Score = (very disappointed count) ÷ (very + somewhat + not disappointed count). N/A is excluded from the denominator because it represents users who never activated. A score at or above 40% is the PMF signal. This calculator also applies a Wilson 95% confidence interval so you can tell whether a 42% with n=30 is actually distinguishable from a 38%.
How many survey responses do I need for a reliable PMF score?
Below n=20 the score is noise. At n=40 it is directional. At n=100 it becomes reliable and shippable. At n=200 you hit investor-grade confidence with roughly ±7pp Wilson CI. At n=400 you reach academic rigor (±5pp). The sample-size ladder in the calculator highlights your current rung so you know whether your number is signal or variance.
What is the "very disappointed" test?
The test scores the Sean Ellis question using only the "Very disappointed" response as the numerator. The intuition: people who would be very disappointed to lose you have formed a dependency — they are your PMF signal. Somewhat-disappointed users are at-risk and fence-sitting. Not-disappointed users never really adopted you.
How do you score a Sean Ellis survey with segments?
Run the analysis across ICP and non-ICP segments separately. Score each segment, then blend using an ICP-weighted composite — default 80% ICP / 20% non-ICP in this tool. A strong ICP lift (15pp or more above non-ICP) means you have a tight fit with your target users and should narrow your acquisition channels. The ICP Extractor pulls the traits of the "very disappointed" cohort into a ready-to-use ideal customer paragraph.
What is a good Sean Ellis score by startup stage?
Pre-seed median 18% (top-quartile above 28%); seed median 24% (top-quartile above 34%); Series A median 33% (top-quartile above 42%); Series B+ median 40%. The 40% threshold is where most Series A-ready companies land. Below 30% is pre-PMF territory. Above 50% is rare-air — Superhuman-tier signal.
How did Superhuman use the Sean Ellis test?
Rahul Vohra (CEO of Superhuman) scored 22% on the first survey, then iterated product positioning and features toward the "very disappointed" cohort until the score reached 58%. His 2018 First Round Review essay made the methodology famous. The core trick: use the "very disappointed" users as your ICP and deliberately ignore the "not disappointed" segment — they will not become fans regardless of what you ship.
What survey questions should I ask alongside the 40% rule?
Add four follow-ups to the core disappointment question: (1) What is the primary benefit you get from [product]? (2) What would you use instead if we went away? (3) What is the one feature you cannot live without? (4) Describe the person you would recommend this to. The first three sharpen positioning. The fourth builds your ICP. This calculator accepts all four fields and assembles them into an ICP paragraph automatically.
Why exclude N/A responses from the PMF score?
N/A ("I no longer use it") is excluded because those users never formed a habit with the product — they cannot tell you whether losing it would disappoint them. Including them in the denominator would systematically understate the score for any product with an onboarding dropout. If more than 25% of your responses are N/A, you are surveying too broadly — filter to users who have completed at least one core action before asking.
Can the 40% rule be wrong? When is PMF a false positive?
Yes. A 42% score with n=30 has a 95% confidence interval of roughly 27–58% — you genuinely cannot distinguish it from a 32% score. Survey selection bias (only surveying power users), N/A exclusion bias (sweeping inactive users out of the denominator), and ICP conflation (high non-ICP scores masking a weak ICP) all create false positives. Superhuman scored 22% initially, then realized they were surveying the wrong segment — once they narrowed to their ICP, the "true" score was much higher.