How to Know If You Have Product-Market Fit (Feedback Data)
TL;DR
You have product-market fit when a meaningful slice of your users would be genuinely upset to lose your product - and you can measure that with feedback data, not vibes. The most reliable single test: survey your active users with "How would you feel if you could no longer use [product]?" If 40% or more say "very disappointed," you likely have it.
But one number isn't the whole story. This guide covers:
- What product-market fit actually means (from the people who defined it)
- The three signal layers: the survey test, the behavior, and the language
- The Sean Ellis 40% test - who to survey, how many, and when
- Whether NPS tells you anything about PMF (honestly, not much)
- What PMF sounds like in your feedback - the stage-by-stage framework
- How to run and analyze a PMF survey, and what to do if you're not there yet
- A diagnostic checklist you can score this week
What Does Product-Market Fit Actually Mean?
Product-market fit means you've built something a specific market genuinely wants - to the point where demand starts pulling the product out of your hands instead of you pushing it onto people. It's not a launch, a funding round, or a revenue number. It's the moment the market stops resisting.
The clearest definition still comes from the essay that popularized the term. As Marc Andreessen, co-founder of Andreessen Horowitz, wrote in 2007: "Product/market fit means being in a good market with a product that can satisfy that market." He argued it's the one milestone that matters more than anything else - "The only thing that matters is getting to product/market fit."
Here's why founders get this wrong: PMF feels binary in hindsight but fuzzy in the moment. Before fit, Andreessen wrote, customers aren't quite getting value, word of mouth isn't spreading, and deals drag on and rarely close. After, "you can always feel product/market fit when it's happening": "The customers are buying the product just as fast as you can make it - or usage is growing just as fast as you can add more servers. Money from customers is piling up in your company checking account."
The problem with "you can feel it": founders are the worst-positioned people to judge it objectively - you want it to be true. That's why you measure it with feedback data instead of trusting your gut, and why getting this wrong is so expensive. In CB Insights' analysis of 431 failed venture-backed startups, poor product-market fit was cited in 43% of post-mortems, second only to running out of capital (70%) - which CB Insights notes is usually "the final cause of death, not the root problem."
How Do You Measure Product-Market Fit With Feedback?
You measure product-market fit across three feedback layers that check each other: a survey signal (the Sean Ellis 40% test), a behavioral signal (retention and organic growth), and a language signal (what your feedback actually says). Any one of them can lie. Together, they're hard to fake.
Most founders reach for a single metric and stop - that's the trap. A survey can throw a false positive; a retention chart can flatter you during a spend spike; glowing feedback can come from a tiny fan club. Three layers give you triangulation: when all three agree, you can trust the verdict.
| Signal layer | What it measures | The tool | What "yes" looks like |
|---|---|---|---|
| Survey | Stated attachment to the product | Sean Ellis 40% test | 40%+ of active users "very disappointed" without it |
| Behavioral | Revealed preference over time | Retention curve + organic growth | Retention curve flattens; growth is partly word-of-mouth |
| Language | The content of feedback | Qualitative feedback analysis | Complaints move from "why would I use this?" to "don't break this" |
The rest of this guide walks through each layer, then hands you a checklist to score all three. The gap between the stated signal (surveys) and the revealed signal (behavior) is really a qualitative vs. quantitative feedback question - and PMF needs both.
What Is the Sean Ellis 40% Test, and How Do You Run It?
The Sean Ellis test asks your active users one question: "How would you feel if you could no longer use [product]?" - answered very disappointed, somewhat disappointed, or not disappointed. If 40% or more choose "very disappointed," you likely have product-market fit. Below that, you probably don't - yet.
The benchmark isn't arbitrary. Sean Ellis - the first marketer at Dropbox and Lookout, head of marketing at LogMeIn, and later founder of GrowthHackers and Qualaroo - derived it by benchmarking nearly 100 startups. Companies above 40% reliably grew; those below it stalled. As he put it: "I've seen several startups that were temporarily able to drive growth and sell product only to have growth crater within a few months. In each case they were well below the recommended 40% benchmark of people that would be very disappointed if they could no longer use the product."
The 40% test is the closest thing to a standard PMF instrument - the foundation of Ellis's original "Startup Pyramid." But running it correctly matters more than running it at all.
Who Should You Survey for the 40% Test?
Survey activated users only - people who have reached your product's core value at least once or twice and used it recently. Surveying signups, trial tourists, or churned users pollutes the result. You want opinions from people who experienced the product, not people who bounced off onboarding.
This is the most common way founders fool themselves. Survey your whole email list and you get a soft, meaningless average; survey people who completed your key workflow at least twice in the last two weeks, and you get a real read on whether the core resonates. Ellis recommends holding off until you've already gathered strong signal from direct customer conversations - the survey confirms a hypothesis, it doesn't replace talking to users.
How Many Responses Do You Need?
Aim for at least 100 responses from activated users before you trust the 40% number, though 30-40 gives you a rough directional read. Below ~100, the percentage swings wildly on a handful of answers, and you'll celebrate or panic over noise.
This is where the test gets abused. Tristan Kromer, a lean-startup coach at Kromatic, points out that a 40%+ result is a strong indicator but never a guarantee - his own startup once scored above 40% and clearly did not have fit. The rule of thumb: scoring below 40% is good evidence you don't have PMF, but scoring above 40% on a small or biased sample doesn't prove you do. One survey can't confirm fit by itself - which is why you cross-check it against behavior and language.
Is NPS a Good Product-Market Fit Signal?
Not really - at least not on its own. NPS (Net Promoter Score) measures stated loyalty and willingness to recommend, which correlates loosely with PMF but was never designed to detect it. Treat NPS as a secondary trend line, not your primary fit gauge - a rising NPS is reassuring, but the absolute number tells you little about fit.
Here's the honest problem. NPS asks "how likely are you to recommend us?" - a hypothetical about future behavior, flattened into a single score that hides the distribution underneath. For B2B software, Retently's 2026 benchmark data puts the average NPS around 41, but averages conceal huge variance, and a healthy NPS can coexist with weak PMF if your promoters are a small niche.
Where NPS earns its keep is as a trend line over time and a feedback generator. The score is noise; the open-text "why did you give that score?" is signal, so mine the comments, not the number. If you're deciding which metric to lean on, our NPS vs. CSAT vs. CES comparison breaks down what each measures - the short version: no satisfaction metric substitutes for the Sean Ellis test when your question is "do we have fit?"
What Does Product-Market Fit Sound Like in Your Feedback?
Product-market fit has a sound. As you approach it, your feedback shifts in a predictable way: complaints move from attacking the core idea, to asking for refinements, to begging you not to break the thing they now depend on. Reading that shift is the most underrated PMF signal there is - and it's free, because you already collect the feedback.
This is the part most PMF guides skip. The survey gives you a number; the language tells you where you actually are. Feedback moves through three recognizable stages.
| PMF stage | What feedback targets | What it's really telling you |
|---|---|---|
| Pre-PMF | The core value proposition itself | People don't get why your product should exist |
| Approaching PMF | Nice-to-haves, edge cases, polish | The core lands; now they want it to fit their workflow |
| Post-PMF | Scaling, integrations, reliability | They've built you into their life and fear losing it |
Let's make each stage concrete. The examples below are illustrative - they show the kind of language you'll see at each stage, not quotes from real FeedSense customers.
Pre-PMF: Feedback Attacks the Core Value Proposition
Before fit, feedback questions the fundamental reason your product exists. Users don't request features - they challenge the premise. The subtext is always "I don't understand why I'd use this instead of what I already do." When most feedback argues with the concept itself, you're pre-PMF, however polite it sounds.
Illustrative pre-PMF feedback:
- "I already do this in a spreadsheet - what does this add?"
- "I'm not sure who this is for."
- "It's neat, but I don't think I'd pay for it."
None of these are about a missing button - they're about the idea. No amount of feature work fixes a value-proposition objection; that's a signal to revisit the problem, not the roadmap.
Approaching PMF: Feedback Shifts to Nice-to-Haves and Edge Cases
As you approach fit, the argument about whether the product should exist ends, and feedback turns to how well it works. People stop questioning the premise and start requesting refinements, integrations, and edge-case fixes. That shift - from "why?" to "can you also?" - is the most encouraging pattern in early feedback.
Illustrative approaching-PMF feedback:
- "Love this - could it also export to CSV?"
- "This is great, but it doesn't handle [my specific edge case] yet."
- "I'd use this daily if it synced with [tool I already use]."
The tell is the implicit endorsement in every request: "I'd use this daily if..." means they already want it and are negotiating terms - a different conversation entirely from the pre-PMF stage.
Post-PMF: Feedback Is About Scaling, Integrations, and "Don't Break This"
After fit, feedback takes on a possessive, anxious tone. Users have woven your product into their workflow, and now they fear losing it. They ask for enterprise controls, deeper integrations, and reliability guarantees - and they get upset about downtime and changes. Anger about a change is a luxury only products with real fit receive.
Illustrative post-PMF feedback:
- "Please don't ever remove [feature] - my whole team relies on it."
- "We need SSO / audit logs / an API before we can roll this out company-wide."
- "The 20 minutes of downtime yesterday cost us - what's your uptime SLA?"
When your inbox fills with "don't break this" and "we need this to scale," congratulations - the market is holding on to you. Your job shifts from finding fit to not fumbling it.
How Do You Run a Product-Market Fit Survey?
Run a short survey to your activated users: the Sean Ellis question first, then three open-text questions that turn a score into a to-do list. Keep it under five questions, send it to people who've hit your core value recently, and aim for 100+ responses before drawing conclusions. It takes an afternoon to set up.
The magic isn't the score - it's the follow-up questions. Here's the exact set, adapted from the structure Ellis and others have used for years.
| # | Question | Why it's there |
|---|---|---|
| 1 | How would you feel if you could no longer use [product]? (Very / Somewhat / Not disappointed) | The core PMF score |
| 2 | What type of person do you think would most benefit from [product]? | Reveals your real target user in their words |
| 3 | What is the main benefit you receive from [product]? | Tells you your true value prop (often not the one you pitched) |
| 4 | How can we improve [product] for you? | Prioritized roadmap input from the people who matter most |
| 5 | (Optional) What would you likely use as an alternative if [product] were gone? | Maps your real competitive set |
A few rules that make or break the results:
- Timing: Trigger the survey after a user completes your key action the second time, not on signup. You want informed opinions, not first impressions.
- Channel: In-app prompts and targeted email both work. If your response rate is thin, our guide on getting user feedback when nobody responds covers the timing, incentives, and phrasing that lift reply rates.
- Segment before you average: Break the "very disappointed" percentage down by user type, plan, and channel. A 30% overall score that's 55% in one segment isn't "no PMF" - it's "PMF with a person you haven't focused on yet."
How Do You Analyze Open-Text PMF Responses at Scale?
Read every "very disappointed" user's open-text answers first - that cohort is your goldmine - then group the responses into themes to see what's driving love and hesitation. For a few dozen responses, do it by hand. Past a hundred or so, manual tagging breaks down and you need thematic analysis to find the patterns without drowning.
The structured way to think about this is theme extraction: turning a pile of free-text answers into a ranked list of recurring benefits, blockers, and requests. When 200 people answer "what's the main benefit?" in 200 different ways, the insight isn't in any single answer - it's in the cluster that all point at the same thing. That's the premise of feedback analysis as a discipline.
This is the point where reading responses one by one stops scaling and AI-assisted analysis earns its place. It's the kind of work FeedSense was built for: it reads open-text survey responses and surfaces the recurring themes and sentiment behind them, so 400 "how can we improve?" answers become a ranked list of themes instead of a spreadsheet you'll never finish. You still make the calls - the analysis just gets you there faster.
Those themes then feed your roadmap: separating must-fix blockers from nice-to-haves is a prioritization problem, and our guide on how to prioritize feature requests gives you a scoring framework for it.
What Should You Do If the Data Says You Don't Have PMF?
If you're below 40%, don't panic and don't scale. Segment down to the users who are "very disappointed," find what they have in common, and rebuild the product and messaging around that high-expectation core. You're not looking for fit with everyone - you're looking for intense fit with someone, then expanding out.
The canonical example is Superhuman. When founder and CEO Rahul Vohra ran the test, only 22% of users said they'd be very disappointed - well under the benchmark. Instead of giving up, he built a systematic engine around the feedback:
- Segment to find the users who love you, and define your "high-expectation customer."
- Analyze the feedback to understand what converts on-the-fence users into fanatics.
- Build the roadmap by, in his words, splitting effort - "spend half your time doubling down on what users already love and the other half on addressing what's holding others back."
- Repeat, and make the PMF score your most important metric.
The result: the score climbed to 33% after focusing, then 58% within three quarters. Vohra's most counterintuitive move was ignoring the wrong users on purpose: "Politely disregard those who would not be disappointed without your product. They are so far from loving you that they are essentially a lost cause." Pleasing them only dilutes the product for the people who love it.
The other half of "don't have PMF yet" is discipline: do not scale. Spending on growth before you have fit is the fastest way to burn your runway. The Startup Genome report, which analyzed more than 3,200 high-growth startups, found that roughly 74% failed from premature scaling - growing headcount and spend before the product stuck - and 93% of those never broke $100k in monthly revenue. Fit first, then fuel.
One more honesty check: PMF isn't permanent, and it isn't the finish line. As Brian Balfour, founder and CEO of Reforge and former VP of Growth at HubSpot, argues, product-market fit is necessary but not sufficient - you also need product-channel, channel-model, and model-market fit, all of them "always evolving, changing, and breaking." Even after you hit 40%, the work is to keep listening. Closing that loop with the users who gave you feedback keeps fit from quietly eroding - here's how to close the feedback loop without annoying the people who got you here.
The Product-Market Fit Diagnostic Checklist
Score your product across the six signals below this week. Give each 0 (weak), 1 (mixed), or 2 (strong), then total it out of 12. This forces you to weigh all three layers - survey, behavior, and language - instead of cherry-picking the metric that flatters you most.
| Signal | Weak (0) | Mixed (1) | Strong (2) |
|---|---|---|---|
| Sean Ellis 40% test | Under 25% "very disappointed" | 25-39% | 40%+ |
| Retention curve | Keeps declining toward zero | Decline is slowing | Flattens into a stable plateau |
| Organic / word-of-mouth growth | You push every signup | Occasional referrals | Referrals are a real acquisition channel |
| Feedback language | Attacks the core value prop | Asks for nice-to-haves | "Don't break this" / wants scale + integrations |
| Weekly growth rate | Under 1% | 1-4% | 5%+ |
| NPS trend (secondary) | Below category average and flat | Near average | Above ~40 and rising |
The weekly-growth benchmark comes straight from Y Combinator. As co-founder Paul Graham wrote, "A good growth rate during YC is 5-7% a week. If you can hit 10% a week you're doing exceptionally well. If you can only manage 1%, it's a sign you haven't yet figured out what you're doing." Growth is the revealed-preference check on what the survey claims.
Now read your score:
- 0-4 - Not yet. You're pre-PMF. Stop adding features and spend. Go back to the "very disappointed" cohort, find what they share, and rebuild around them. Talk to users before you survey again.
- 5-8 - Approaching. The core resonates with someone. Focus: define your high-expectation customer, double down on what they love, and clear the blockers keeping fence-sitters out. Re-survey each quarter.
- 9-12 - You likely have it. Fit is real. Shift from finding it to not fumbling it - protect reliability, build the integrations power users ask for, and find repeatable growth channels. Keep measuring; fit erodes if you stop listening.
No single row is destiny - a product can score an 8 with a mediocre NPS, or a 6 that hides a 55% "very disappointed" score in one killer segment. The checklist isn't a verdict machine; it forces you to see the whole picture honestly, which is the one thing founders reliably avoid.
Measure Fit With Feedback, Using FeedSense
FeedSense turns PMF survey responses into a clear read on where you stand - without hand-tagging hundreds of open-text answers.
What you get:
- Theme and sentiment analysis across open-text PMF responses
- Segmentation, so you can read the "very disappointed" cohort on its own
- A ranked view of the benefits, blockers, and requests hiding in your feedback
What it won't do: tell you that you have PMF (the 40% call is yours), run the survey for you (bring your own survey tool), or replace talking to users directly.
You can set up feedback collection in 5 minutes and start seeing the themes today. Start free - the free tier includes 100 feedback items/month, no credit card required.
FAQ
What percentage means you have product-market fit?
The widely used benchmark is 40%: if 40% or more of your active users say they'd be "very disappointed" to lose your product on the Sean Ellis survey, you likely have product-market fit. Below 40% suggests you don't yet. It's a strong indicator, not a guarantee - always cross-check it against retention and feedback language.
What is the Sean Ellis test for product-market fit?
The Sean Ellis test is a one-question survey: "How would you feel if you could no longer use [product]?" with options of very disappointed, somewhat disappointed, or not disappointed. Sean Ellis derived the 40% "very disappointed" benchmark by studying nearly 100 startups and finding it reliably separated companies that grew from those that stalled.
Who should I survey to measure product-market fit?
Survey activated users only - people who have reached your product's core value at least once or twice and used it recently. Do not survey signups, trial tourists, or churned users; they dilute the result. Aim for at least 100 responses from real users before trusting the percentage, since smaller samples swing wildly on a few answers.
Is a high NPS the same as product-market fit?
No. NPS measures stated willingness to recommend, which correlates loosely with product-market fit but wasn't designed to detect it. A healthy NPS can coexist with weak PMF if your promoters are a small niche. Use NPS as a secondary trend and mine its open-text comments, but rely on the Sean Ellis test to actually gauge fit.
What does feedback look like before you have product-market fit?
Before fit, feedback attacks the core value proposition rather than requesting features. You'll hear "I already do this in a spreadsheet," "I'm not sure who this is for," or "why would I switch?" People are arguing with the premise, not the polish. When most feedback questions why the product should exist, you're pre-PMF regardless of how polite it sounds.
What should I do if I don't have product-market fit yet?
Don't scale. Segment down to the users who would be "very disappointed" without you, identify what they share, and rebuild the product and messaging around that high-expectation core. This is the Superhuman playbook - Rahul Vohra raised their score from 22% to 58% by doubling down on what fans loved and ignoring users who were never going to care.
How long does it take to reach product-market fit?
There's no fixed timeline - it varies from months to years depending on the market. What matters is treating PMF as a metric you improve, not a moment you wait for. Superhuman moved from 22% to 58% "very disappointed" in about three quarters by systematically acting on feedback each cycle, rather than hoping fit would arrive on its own.
Can product-market fit be lost after you find it?
Yes. Product-market fit isn't permanent. Markets shift, competitors improve, and your product can drift from what made it stick. Brian Balfour of Reforge notes that the fits a company needs are "always evolving, changing, and breaking." Keep running the survey, keep reading feedback language, and close the loop with users so erosion shows up early enough to fix.
Tags:
Stay in the loop
Get the latest insights on customer feedback, product management, and building better SaaS products. No spam, unsubscribe anytime.
By subscribing, you agree to our privacy policy. We respect your inbox.