# How to Run a Product-Market Fit Survey With Under 100 Users > Run a product-market fit survey with under 100 users: who to ask, the Sean Ellis question, a 5-question template, and how to read small results. ```yaml url: "https://makerhunt.io/studio/how-to-run-a-product-market-fit-survey-with-under-100-users" markdown: "https://makerhunt.io/studio/how-to-run-a-product-market-fit-survey-with-under-100-users.md" title: How to Run a Product-Market Fit Survey With Under 100 Users type: studio_article category: Growth tags: "product-market fit, PMF survey, Sean Ellis test, customer feedback, user research, positioning, early users, makers, indie hackers" featured_image: "https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9nI8XuFwwdTIgQle9zSK4iMy6CWEBbsfhtqP8" published_at: "2026-10-13T09:48:00.000Z" updated_at: "2026-10-13T09:52:37.381Z" ``` ## About this studio article Run a product-market fit survey with under 100 users: who to ask, the Sean Ellis question, a 5-question template, and how to read small results. ## Article To run a product-market fit survey with under 100 users, send one short survey to the people who actually used your core feature recently (a good filter is two or more uses in the last two weeks), ask the Sean Ellis question "How would you feel if you could no longer use [product]?", add three or four open follow-ups, and read the answers by segment instead of staring at one percentage. At small numbers the score is a rough signal, not a verdict. The real value is finding the people who would be very disappointed without you, learning their words, and building for them. Below: where the 40% rule comes from, who to survey, a copy-paste template, how to read results, a made-up example, a 14-day plan, and the traps. ## What is a product-market fit survey, and what is it not? A product-market fit survey asks current, active users how they would feel if they could no longer use your product. The share who answer "very disappointed" is your PMF score. The open questions around it tell you who loves the product, why, and what holds everyone else back. It is not a few things people confuse it with: - **It is not NPS.** "Would you recommend us?" measures goodwill. "Would you be very disappointed if this disappeared?" measures dependence. People happily recommend tools they could drop tomorrow. - **It is not idea validation.** Nobody can tell you how disappointed they would be to lose something they never used. If you have no product yet, talk to people instead, for example by following this SideHunt guide to [validate a side project idea with five conversations](https://sidehunt.io/studio/how-to-validate-a-side-project-idea-with-five-conversations). - **It is not a feature request form.** You will get requests. Most should not go straight onto the roadmap, and later sections explain which ones to take seriously. ## Where does the 40% rule come from? The Sean Ellis test comes from Sean Ellis, who ran early growth at companies like Dropbox and Eventbrite. In his July 2009 post [The Startup Pyramid](https://www.startup-marketing.com/2009/07/), he wrote that he asks existing users how they would feel if they could no longer use the product, and that in his experience product/market fit requires at least 40% of users saying they would be "very disappointed." He also called the threshold "a bit arbitrary," defined after comparing results across nearly 100 startups: the ones struggling for traction were under 40%, and most that gained strong traction were over it. So the 40% rule is one experienced operator's benchmark, not a law of nature. Use it as a direction. The method got far more popular after Superhuman founder Rahul Vohra described his team's process in First Round Review's piece on [how Superhuman built an engine to find product-market fit](https://review.firstround.com/how-superhuman-built-an-engine-to-find-product-market-fit/). According to that article, Superhuman's first survey came back at 22% "very disappointed." Segmenting down to the users who loved the product most gave 33%, and within three quarters of product work the score reached 58%. The same article cites Hiten Shah asking Slack users the question in 2015, with 51% of 731 respondents saying they would be very disappointed. Those are big-company numbers. Do not copy them as targets. Copy the process. ## Is a PMF survey worth running with under 100 users? Yes, as long as you treat it as a read, not a verdict. The arithmetic is simple. If 25 people answer, each person is 4 percentage points. If 20 answer, each is 5 points, so 8 "very disappointed" answers is 40% and 7 is 35%. A 38% from 21 people and a 43% from 23 people are, for practical purposes, the same result. My rule of thumb (an opinion, not a statistical standard): under about 20 responses, read the open answers and ignore the percentage. Between roughly 20 and 40, write the percentage down as a direction. In the First Round piece, Vohra says Superhuman had between 100 and 200 users to poll at the time, and that you start to get "directionally correct results around 40 respondents." Still a direction, not a verdict. The open answers are where small user bases win. Twelve detailed answers from people who rely on your product every week can teach you more than a month of dashboard watching. ## Who should you survey? Not every signup. If you email everyone who ever created an account, you mostly hear from people who poked around once and left. Their "not disappointed" is true, and it tells you almost nothing about the product they never really used. In his 2009 post announcing [his free customer development survey,](https://www.startup-marketing.com/free-customer-development-help-surveyio/)[Survey.io](http://Survey.io), Ellis recommended sending it to people who have experienced the core of your product, used it at least twice, and used it in the last two weeks. Superhuman used the same filter. It is a good default for small products too: 1. **Define your core action.** The one thing a user does that means they got value: sent the invoice, published the page, finished the lesson notes. If you have not defined it yet, start with [tracking product metrics before your first 100 users](https://indiehunt.io/studio/how-to-track-product-metrics-before-100-users). 2. **Pull everyone who did that action at least twice in the last 14 days.** With a tiny base, survey all of them rather than sampling. 3. **Remove people who bias the answer.** Co-founders, partners, friends you begged to try it, anyone you are paying. 4. **Write down the count.** A score means nothing without the denominator. If the filter leaves you with eight people, that is a finding too. Your bigger problem is activation and retention, and a survey score will not fix it. ## What questions should you ask in a PMF survey? Keep it short enough to finish in two minutes. Superhuman's survey, as described in the First Round article, used four questions: the disappointment question, then who would most benefit, the main benefit, and how to improve. Ellis's [Survey.io](http://Survey.io) post also said the survey helps with positioning against the most common alternative solutions, so I add one optional question about what people would use instead. Paste this into any form tool: # Question Type What it tells you 1 How would you feel if you could no longer use [product]? Single choice: Very disappointed / Somewhat disappointed / Not disappointed / N/A, I no longer use [product] Your PMF score, and which group each person is in 2 What type of people do you think would most benefit from [product]? Short text Your best persona, in their own words 3 What is the main benefit you receive from [product]? Short text Why people stay, which becomes your headline 4 How can we improve [product] for you? Short text What holds people back from loving it 5 (optional) What would you use instead if [product] were no longer available? Short text Your real competition, often a spreadsheet or "nothing" - **Keep the N/A option.** Ellis's version includes "N/A, I no longer use [product]," which stops lapsed users from muddying the other answers. - **Ask question 1 first.** My preference: open with improvement questions and people start answering like critics before they tell you how losing the product would feel. - **Record email or user ID.** You will want to segment by plan or usage later, and reply to a few people. ![image](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9bfPSrlFnXleN6Dujc01vO79HRMIZYUCyTtKf) ## How do you send it so people actually answer? With a small user base, a plain email from the founder beats any polished campaign. People answer people. Something like this: **Subject:** Two minutes to help me decide what to build next "Hi [first name], I'm [your name], I build [product]. You've used it a few times in the last couple of weeks, so your answer matters more than most. Five quick questions, about two minutes: [link]. Honest answers help more than nice ones, including 'I wouldn't miss it.' Thanks, [your name]." - **Send from your own address,** not no-reply. Some people will just reply with a paragraph. That is a gift. - **Pick a normal working day** (a rule of thumb, not data). Avoid the day after a big release, when goodwill is inflated, or after an outage, when it is not. - **Use an in-app prompt only for people on your list,** shown after they finish the core action, not on login. - **Send one reminder** three or four days later, only to non-responders. Then stop. - **Skip big incentives.** A gift card attracts people who want the gift card. If you want to say thanks, offer something small tied to the product and mention it after they finish. Leave the survey open for about a week so people who read email late still get a chance. ## How do you read the results? Start with the score, then spend most of your time on the words. ### 1. Calculate the score honestly Divide the "very disappointed" count by everyone who picked one of the three real options. I leave N/A answers out of the denominator and note them separately. Always write n next to the percentage. "36% (n=22)" is honest. "36%" alone is not. ### 2. Segment the "very disappointed" group Put every response in a sheet, one row per person, with their question 1 answer, role or use case, plan, and how long they have been a user. Filter to "very disappointed." What do they have in common? Same job, company size, use case, or way they found you? This is the step that moved Superhuman from 22% to 33%: the narrower segment was who the product was really for. ### 3. Steal their words Read the fans' answers to questions 2 and 3 out loud. People usually describe themselves when they say who would benefit most, and they describe your real value when they name the main benefit. Copy exact phrases into a doc. They beat anything you will write from scratch, and they slot straight into a [one-page positioning brief before your next homepage rewrite](https://indiehunt.io/studio/how-to-write-a-one-page-positioning-brief-before-homepage). ### 4. Split the "somewhat disappointed" group This is where useful product work hides, and where roadmaps get wrecked: - **Those who named the same main benefit as your fans** already get the core value. Something smaller holds them back. Their "how can we improve" answers are the requests worth building. - **Those who named a different benefit** want a different product. Thank them and park their requests. The First Round article describes Superhuman making exactly this split around speed, their main benefit. ### 5. Do not let the "not disappointed" group steer Read their answers, especially question 5, but do not plan features around them. They are telling you the product is not for them, and that is fine. ## What should you do at 10%, 25%, or 40% and up? These bands are my rules of thumb for small products, not thresholds from Ellis or anyone else. Score (with n noted) What it probably means What to do next Under about 15% Very few people depend on the product yet, or you surveyed the wrong people Check the list first. If it was right, interview the few fans you have and consider narrowing to the one job they mention. About 15% to 30% A real core of fans, buried inside a broad user base Segment hard. Rewrite homepage and onboarding for the best persona only. Build the top request from near-fans who share the main benefit. About 30% to 40% Close, and the gap is often something specific Split roadmap time between doubling down on the main benefit and fixing the most common blocker. Re-run next quarter. 40% and up A promising sign you are a must-have for this group Check it holds with the next batch of users, then shift more time to distribution. Two warnings. A high score from 12 people is encouraging, not proof. And a high score from a segment too small to build a business on is a different problem; Ellis made the same point in his 2009 post, that beyond early traction those users need to represent a large enough market. ## How do you track your PMF score over time? Run it again every quarter, or after a big change to the product or audience. Per the First Round piece, Superhuman kept surveying new users and was careful not to survey anyone more than once. With a small base you run out of fresh people fast, so a quarterly round to users who joined since the last one is a sensible default. Keep one sheet, one row per round: date, surveyed and answered counts, the four answer counts, score with n, best-segment score with n, top main benefit in users' words, top blocker from near-fans, and what you shipped because of it. Watch the trend across three or four rounds, not one, and check it against behaviour. If the score climbs but people still stop using the product after a month, trust the usage data. ## A made-up example: Tutorloop Everything in this example is made up: the product, the person, and every number. Ana runs **Tutorloop**, a small tool that turns a tutor's quick lesson notes into a tidy recap email and homework list after each session. She has 84 signups. Her core action is "sent a lesson recap." Two or more recaps in the last 14 days leaves 31 people. She removes two friends and her cousin, leaving 28. She sends the five-question survey from her own email on a Tuesday morning, with one reminder on Friday. 21 people answer: 7 very disappointed, 9 somewhat, 4 not disappointed, 1 N/A. Leaving the N/A out, her score is 7 out of 20, so 35% (n=20). Of the 7 "very disappointed" answers, 6 are independent language tutors who teach adults online, and most describe the main benefit the same way: they send the recap in about a minute instead of writing it at the end of the day. Only 1 is a music teacher. Of the 9 "somewhat disappointed" users, 4 named that same time-saving benefit, and 3 of those 4 asked for recaps in the student's language, not only English. The other 5 asked for scheduling, payments, and a student portal: a different product. Ana rewrites her homepage headline around language tutors, using their phrase about sending recaps in a minute. She builds recap translation for the near-fans and parks scheduling and payments. Next quarter she surveys the 26 new active users who joined since. Still small, but the direction is what she tracks. ![image](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9bP1iSWFnXleN6Dujc01vO79HRMIZYUCyTtKf) ## A 14-day plan for your first product-market fit survey Times below are planning guesses, not measurements. Day Time Action 1 30 min Define your core action in one sentence. Pull users who did it at least twice in the last 14 days. 2 20 min Clean the list: remove team, friends, paid testers. Write down the final count. 3 30 min Build the five-question survey from the template. Test it on your phone. Make sure it records email or user ID. 4 20 min Write the founder email in your own words and send it in the morning from your own address. 5-6 10 min/day Reply to anyone who answers by email. Do not peek at the score and start changing things. 7 10 min Send one reminder to non-responders only. 8-10 10 min/day Let answers come in. Close the survey at the end of day 10. 11 45 min Calculate the score with n. Segment the "very disappointed" group. Note the persona and their exact words. 12 45 min Split the "somewhat disappointed" group by main benefit. List the near-fans' top requests. 13 30 min Pick one positioning change and one product change. Park the rest in writing. 14 20 min Start the tracking sheet with row one. Put the next round on your calendar three months out. ## Common traps - **Surveying every signup.** You get a low score that mostly measures your activation problem. - **Reporting the score without n.** 50% of 6 people is a conversation starter, not a result. - **Surveying right after a launch.** Goodwill spikes, and so does the score. - **Building for the loudest "not disappointed" users.** Their requests feel urgent and lead to a different product. - **Treating 40% as a finish line.** Vohra notes the score may drop as you reach less forgiving users. That is why you re-run it. - **Leading wording.** "How much would you miss our amazing time-saving tool?" is not the Sean Ellis question. Keep question 1 word for word if you want to compare rounds. ## FAQ ### Is the Sean Ellis test the same as NPS? No. NPS asks how likely someone is to recommend you. The Sean Ellis test asks how disappointed they would be to lose you. For an early product I think dependence is the more useful thing to know, because it separates a must-have from a nice-to-have, which is the distinction Ellis built the question around. ### Can I run a PMF survey before launch? Only with people already using a working version regularly, like beta users. If nobody uses the product yet, the question has no real answer. Do customer conversations first. ### What if only 10 or 15 people pass the active-user filter? Send it anyway, ignore the percentage, and read every answer. Better still, turn it into short calls with the people who say "very disappointed." At that size a conversation teaches you more than a score. ## Where to start this week Write down your core action today. Pull everyone who did it twice in the last two weeks. Send five honest questions from your own email, and when answers come in, spend more time on the words of people who would be very disappointed than on the percentage. That is a product-market fit survey that works at under 100 users. When the product is ready to meet more makers and early adopters, you can also list it on [MakerHunt](https://makerhunt.io/). ## Links - Article (HTML): https://makerhunt.io/studio/how-to-run-a-product-market-fit-survey-with-under-100-users - AI-friendly Markdown: https://makerhunt.io/studio/how-to-run-a-product-market-fit-survey-with-under-100-users.md ## Explore - [Studio index](https://makerhunt.io/studio) - [Blog](https://makerhunt.io/blog) - [Browse projects](https://makerhunt.io/projects) _HTML version: https://makerhunt.io/studio/how-to-run-a-product-market-fit-survey-with-under-100-users_