FREE CALCULATOR · SEAN ELLIS 40% TEST · SURVEY KIT
Product market fit calculator: score your startup with the 40% test
Product market fit is the point where a clear group of people wants your product so much that they would be upset to lose it, keep coming back without reminders and pay for it. This free calculator measures it the way Sean Ellis does: copy the 4-question survey kit, send it to active users, type in the answer counts and get your PMF score, a band, a sample-size check and three next steps. A second check sets the survey against what customers actually do. Below it: the signs of fit, 5 measures compared, the Superhuman case and a worked example.
Product market fit calculator
Copy the survey, send it to active users, then type in the answer counts. You get your PMF score, a band, a sample-size check, a behaviour check and three next steps.
Step 1
Copy the survey kit
Who to ask:
- People who used your product at least twice in the last 2 weeks
- People who reached the main experience at least once (a payment, a finished plan, a sent file)
- Each person once. Aim for 40 or more answers; around 30 is Sean Ellis's minimum
Step 2
Type in the answer counts for question 1
Example numbers loaded (40 responses). Replace them with your own.
Step 3 (optional)
Check what customers do
0%PMF score
Not sure what your score means for your idea? Paste the result and your follow-up answers into a chat with Elona, your AI co-founder, or test demand quest by quest in the Fe/male Switch startup game.
Ask Elona about my score Play the startup gameSources and date: Sean Ellis (Venture Hacks interview, 2009), Rahul Vohra (First Round Review, 2018), Kromatic's Real Startup Book; checked 7 November 2026. The 40% line is Ellis's; the 25% line is our reading, and the margin of error is our calculation (95%). Nothing you type leaves your browser.
What is product market fit, in plain words?
Marc Andreessen, co-founder of the venture firm Andreessen Horowitz, gave startups their working definition in a June 2007 essay: "Product/market fit means being in a good market with a product that can satisfy that market." You need both halves. A good market is a group of people with a problem urgent enough that they already spend time or money on it. A product that satisfies them solves that problem well enough that they keep using it and tell others.
Without fit, Andreessen wrote, customers do not quite get value, word of mouth does not spread and usage grows slowly. With fit, in his words, "the customers are buying the product just as fast as you can make it." He credits investor Andy Rachleff with the rule behind it: "The #1 company-killer is lack of market."
Recent data still backs him. CB Insights studied 431 venture-backed companies that shut down since 2023 (report published 5 March 2026). Among the 385 where a reason could be found, 43% cited poor product-market fit. Running out of money topped the list at 70%, and the report calls cash the final cause of death more than the root problem.
1. Problem-solution fit
A real group has the problem, and your idea solves it on paper. Proof: interviews, pre-orders, a waitlist with a price on it. Our guide to problem-solution fit covers this stage.
2. Product market fit
People use the real product, come back and would be very disappointed to lose it. Proof: a survey score of 40% or more, retention that levels off, referrals.
3. Growth
You put money and people into reaching more of the same customers. Doing this before stage 2 burns cash on users who leave in week one.
How does this product market fit calculator work?
The tool follows the survey Sean Ellis built while he ran early growth at Dropbox, LogMeIn and Eventbrite. In a December 2009 interview with Venture Hacks he put the rule simply: "If you're above 40% of the people saying they'd be very disappointed, I tend to say you've found product/market fit." Rahul Vohra, founder of the email app Superhuman, wrote in First Round Review (November 2018) that Ellis benchmarked nearly a hundred startups this way, and companies under 40% struggled to grow.
- The question set: Ellis's "How would you feel if you could no longer use [product]?" with four answers, plus the three follow-up questions Superhuman published: who benefits most, the main benefit and how to improve. Paste it into Google Forms, Typeform or Tally.
- Who to ask: people who used the product at least twice in the last two weeks, the filter Vohra quotes from Ellis. Ellis's 2009 version: people who came back at least once and reached the main experience.
- Your PMF score: very disappointed ÷ (all responses minus N/A) × 100. N/A means "I no longer use it", and those people fall outside the active-user filter, so the tool leaves them out of the base. Superhuman's article does not spell out the base, so the result also shows the score with N/A counted in.
- The band: 40% and above is a strong signal, Ellis's threshold. We call 25% to 39% close and anything under 25% not yet; those two lower bands are our reading. Superhuman started at 22% and still found fit.
- The sample check: Ellis looks for around 30 responses as a minimum, and Vohra writes that results become directionally correct at around 40. The tool warns under 30 and shows a margin of error for every score.
- The behaviour check: three yes/no questions about what customers do: come back after 4 weeks, pay, refer others. The verdict sets the survey against behaviour, because words and actions can disagree.
The margin of error is our own calculation, the standard 95% formula for a share. At 30 responses a score of 40% could sit anywhere between about 22% and 58%. At 100 responses the same score narrows to about 30% to 50%. That spread explains why 38% and 42% from a small survey tell you the same thing.
Prefer a printable version? Here is a step-by-step Sean Ellis survey method card.
What are the signs of product market fit?
A survey records what people say. Behaviour shows what they do, and three behaviours separate fit from hope.
1. Your retention curve flattens
Group users by the month they started (a cohort) and track what share is still active after 1, 2 and 3 months. Sequoia Capital's data science team puts it bluntly: "When a product has not achieved product-market fit, the retention curve will continuously decline." A curve that drops and then levels off means a share of users found lasting value. Sequoia adds that the higher the level where it flattens, the healthier the product.
2. Customers bring customers
Andreessen names word of mouth as the first thing that stalls without fit. When a growing share of new users arrives through a friend or a search for your name, people are selling the product for you. Our guide to referral programs shows how to make sharing easy once it starts on its own.
3. The market pulls the product out of you
Customers ask for the next version, complain loudly when it breaks and pay without long negotiations. Before fit, every sale feels like pushing a boulder uphill. After fit, your problem becomes keeping up.
How high should retention be? Lenny Rachitsky and Casey Winters asked about 20 growth practitioners in 2020 and published these six-month retention benchmarks. Treat them as a rough yardstick from expert opinion. SaaS means software as a service, SMB means small and medium-sized businesses.
| Business type | Good 6-month retention | Great 6-month retention | Counted from |
|---|---|---|---|
| Consumer social | about 25% | about 45% | registered users |
| Consumer transactional (marketplaces, booking) | about 30% | about 50% | users with at least one transaction |
| Consumer subscription (SaaS) | about 40% | about 70% | users who started a paid subscription |
| SMB and mid-market SaaS | about 60% | about 80% | companies with a paid subscription |
| Enterprise SaaS | about 70% | about 90% | companies with a paid subscription |
Source: Lenny Rachitsky, "What is good retention?", Lenny's Newsletter, 9 June 2020, based on answers from about 20 growth experts. The article gives both 70% and 75% as good enterprise retention; the table uses 70%.
How do you measure product market fit, step by step?
Use three methods together: the survey tells you who loves the product, cohorts tell you whether they stay, and interviews tell you why. Let's break it down.
- List your active users. Everyone who used the product at least twice in the last two weeks. With fewer than 30, run interviews first.
- Send the survey once per person. Copy the question set from the tool. Superhuman tracked its score weekly, monthly and quarterly, and never asked the same user twice.
- Calculate the score in the calculator above, and write down the band and the margin of error.
- Segment. Sort answers by the first question. Read what the very disappointed group wrote about who benefits and the main benefit, describe that person, then recalculate for respondents who match.
- Build cohorts. In a spreadsheet, put each signup month in a row and the share still active after month 1, 2 and 3 in the columns. Look for the curve to level off.
- Interview five people from each group. Very disappointed users tell you what to protect. Somewhat disappointed users who name the same main benefit tell you what to fix. Use the 15-minute problem interview script.
- Rerun in 4 to 6 weeks with new active users and compare. One score is a snapshot, and three scores show a direction.
Ellis's advice in that 2009 interview for anyone under 40%: watch the very disappointed number closely, keep your burn low and wait before you try to grow.
Which product market fit measures should a first-time founder trust?
Each measure answers a different question, and each one can mislead you on its own. The table compares the five you can run without a data team.
| Measure | What it tells you | When it lies | Effort |
|---|---|---|---|
| Sean Ellis test (40% survey) | How many active users would be very disappointed to lose the product, and who they are | Small samples, or surveys sent to fans and non-users | Low: one form, 30 to 40 answers |
| Retention cohort | Whether people keep using the product after week 1, month 1 and month 3 | Tiny cohorts, or a loose definition of active | Medium: needs a few months of usage data |
| Net Promoter Score (NPS) | How likely people are to recommend you on a 0 to 10 scale: % promoters (9 to 10) minus % detractors (0 to 6) | It measures goodwill, and polite users give a 9 to a product they will never open again | Low: one question |
| Paying customers and revenue | Whether people value the product enough to pay, and keep paying | One large customer, deep discounts or a launch-week spike can hide churn | Low once you charge |
| Organic share of new users | What share of new users came through word of mouth, referrals or a search for your name | Hard to read unless you ask every new user how they found you | Low: one "how did you hear about us?" question at signup |
Our comparison, 7 November 2026. NPS definition from Bain & Company's Net Promoter System site.
The safest reading combines a survey with one behaviour measure. When both agree, believe them. When they disagree, believe the behaviour, then find out why people said something different. The behaviour check in the calculator does exactly this. If retention is your weak spot, start with our guide to churn analysis.
How did Superhuman raise its PMF score from 22% to 58%?
Superhuman is an email app, and founder Rahul Vohra published the full method in First Round Review in November 2018. In summer 2017 the team had 14 people and 100 to 200 users. The first survey came back at 22% very disappointed, far below 40%.
- Segment the fans. The team grouped answers by the first question and gave each respondent a persona. The very disappointed group was mostly founders, managers, executives and business development people. Looking only at those personas lifted the score to 33%.
- Describe the person who loves it most. Using Julie Supan's high-expectation customer idea, they profiled a busy professional who reads 100 to 200 emails a day.
- Find what the fans love. Very disappointed users named speed as the main benefit.
- Split the build plan in half. Half the work went to what fans already loved: more speed, more shortcuts, more automation. The other half fixed what held back somewhat disappointed users who also valued speed, such as a mobile app, attachments, calendaring and better search.
- Track the score every week. Within three quarters, the score reached 58%.
The lesson for a first-time founder: your first score describes your current mix of users, so the right segment can move it before you build anything. For comparison, Vohra cites Hiten Shah's 2015 survey of 731 Slack users, where 51% answered very disappointed.
Which signals fake product market fit?
First-time founders mistake attention for demand. Each signal below feels like proof and costs the customer nothing to give, which is why it proves so little.
| Signal | Why it feels like fit | What to check instead |
|---|---|---|
| Friends and family love it | They are honest about liking you | Ask strangers in your target group for money or a pre-order |
| Hundreds of signups or a long waitlist | A big number on launch day | How many signups used it twice in two weeks, and how many paid |
| Press coverage or a viral post | Reporters and strangers are talking about you | Retention of the users who arrived that week, 4 and 8 weeks later |
| Nice survey answers from everyone | High average scores | The very disappointed share among active users only |
| One big customer | Real revenue and a real logo | Whether a second and third similar customer buy without custom work |
| Growth from paid ads | Charts go up and to the right | What happens to signups and retention when you pause ads for two weeks |
Our list, based on the methods on this page.
Andreessen's essay notes that reporters call once you have fit. The order matters: press follows fit and never creates it. A spike of visitors who leave in week one tells you the headline works, and the product itself is still untested.
What should you do before you have product market fit?
Keep spending low, learn fast and narrow the audience until a small group loves the product. Here is why: Ellis's 2009 advice was to wait before trying to grow, and CB Insights found that 70% of the failed startups it studied ran out of money. Every euro you keep before fit pays for another round of tests.
- Pick one narrow customer. "Working parents of toddlers in Amsterdam" beats "busy people". The user persona generator turns interview notes into a profile you can test.
- Test demand before you build. A smoke test with a price on it measures intent for very little money.
- Get paid early. A pre-order or a paid pilot is the strongest early signal there is. See pre-launch validation: getting paid before you build.
- Work in 30-day cycles. The 30-day validation plan for first-time founders breaks a test cycle into weekly tasks.Size the segment. The market size calculator tells you whether the group that loves you is big enough to build a business on.
- Change direction on evidence. If three survey rounds stay under 25% and retention keeps falling, read about the 7 types of startup pivots. If one round let you down, read when to hold your course first.
For a look at fit in a young market, our playbook article on product-market fit lessons from voice technology platforms walks through real examples.
How do Elona and the startup game help you test demand?
Fe/male Switch is a startup game for first-time founders, and Elona is the AI co-founder who builds the company with you. Both follow the logic of this page: prove that real people want the product before you spend months building it.
1. Ask Elona for an honest take
Paste your PMF score and the answers to the main benefit question into a chat with Elona, your AI co-founder. She points to the segment worth chasing and writes a 7-day plan.
2. Play the startup game
The Fe/male Switch startup game turns validation into quests: interviews, a landing page, a first sale, with feedback at every step.
3. Use the free generators
Find an angle with the startup ideas generator, check it with the idea validator and name it with the business name generator.
4. Check your founder habits
The entrepreneurial mindset test shows how you react to a no, which matters when your first score comes back at 15%.
Worked example: what does a PMF score look like for a small app?
This is an invented example with stated assumptions, to show the arithmetic. Mila runs a meal-planning app for parents. 61 of her 140 signups used it at least twice in the last two weeks, and 52 of them answer the survey.
| Answer | Responses | Share of the base |
|---|---|---|
| Very disappointed | 16 | 34.8% |
| Somewhat disappointed | 21 | 45.7% |
| Not disappointed | 9 | 19.6% |
| N/A, I no longer use it | 6 | left out of the base |
| Base (52 answers minus 6 N/A) | 46 | 100% |
Invented numbers for illustration.
Her PMF score is 16 ÷ 46 = 34.8%, in the close band. With N/A counted in the base it would be 16 ÷ 52 = 30.8%. Her 46 responses clear Ellis's minimum of around 30, and the margin of error is about ±14 points, so the true figure sits somewhere between about 21% and 49%.
Next she reads the follow-up answers. 12 of the 16 very disappointed users describe themselves as working parents with children under four, and they name the same main benefit: the weekly shopping list. Among all respondents in that segment, 12 of 22 answered very disappointed, which is 54.5%.
The segment score looks like fit, and the sample is small: 22 responses sits under Ellis's minimum, so the calculator shows it with a warning. Mila's next move: survey 20 more working parents of toddlers and answer the behaviour checks. If the segment holds above 40% with 40 or more answers and those parents keep coming back, she builds for them first.
Product market fit FAQ
What is product market fit?
Product market fit means a clear group of customers wants your product enough to keep using it, pay for it and recommend it. Marc Andreessen defined it in 2007 as being in a good market with a product that can satisfy that market.
What is the Sean Ellis test?
A one-question survey: "How would you feel if you could no longer use [product]?" with the answers very disappointed, somewhat disappointed, not disappointed and N/A. Sean Ellis says that above 40% very disappointed among active users, you have found product/market fit.
How do you calculate a PMF score?
Divide the very disappointed answers by the responses from active users and multiply by 100. This calculator leaves N/A answers (people who no longer use the product) out of the base and also shows the score with them included.
What is a good product market fit score?
40% or more very disappointed is Ellis's line for a strong signal. Superhuman began at 22% and reached 58% within three quarters, and Slack scored 51% in Hiten Shah's 2015 survey of 731 users.
How many responses do you need for a product market fit survey?
Sean Ellis looks for around 30 responses as a minimum, and Rahul Vohra of Superhuman writes that results become directionally correct at around 40. With fewer, treat the score as a hint and keep interviewing.
What are the signs of product market fit?
A retention curve that drops and then flattens, a growing share of new users who arrive through word of mouth, and customers who pay and ask for more without being pushed.
Is NPS a good measure of product market fit?
Net Promoter Score shows how likely people are to recommend you. It measures goodwill more than need, so pair it with the Sean Ellis test and a retention cohort.
How long does it take to find product market fit?
There is no fixed timeline. Superhuman needed three quarters of focused product work to move from 22% to 58%. Rerun the survey every 4 to 6 weeks and judge the direction.
What is the difference between problem-solution fit and product market fit?
Problem-solution fit means you have shown, through interviews or pre-orders, that a real group has the problem and wants your kind of solution. Product market fit comes later and needs a working product that people keep using. Our guide to problem-solution fit covers the first stage.
Read next
Made by Fe/male Switch, the startup game where first-time founders build a company with Elona, the AI co-founder.
