Product-Market Fit: What It Is, How to Measure It and the 40% Test
Product-market fit means, in Marc Andreessen's words of 2007, “being in a good market with a product that can satisfy that market.” Sean Ellis's test puts a number on it: ask active users how they would feel if they could no longer use the product, and call it fit when at least 40% answer “very disappointed”. On 19 November 2026 we read the product-market fit survey templates of 11 survey tools: 8 ask Ellis's question word for word, 8 print the 40% line and 3 set a minimum number of answers, from 40 to 100.
What follows covers what product-market fit means, read on Andreessen's and Ellis's own pages, how to run and score the product-market fit survey with a worked example, how to measure fit beyond one question, its stages and how problem-solution fit differs from it.
Fit is a product a good market wants, and its usual measure is Ellis's survey: 40% of active users “very disappointed” to lose it. Eight of 11 survey tools ask his question word for word, and their minimums run from 40 answers to 100; at 40 a score carries a margin of about 15 points. Superhuman went from 22% to 58% by acting on the answers. None of it works without users to ask. 10words, a startup newsletter we submit to, prices one sponsored week at $150. For $29, Submitator files your product with 40 directories you pick among the 202 directories and launch sites in our catalog, and keeps live any badge those sites require.
What is product-market fit? Definition, meaning and who coined it
Product-market fit is the state in which a product satisfies a market that wants it badly enough to pay, return and tell others. The product-market fit definition most guides repeat comes from Andreessen's essay posted on 25 June 2007, and it has two halves: a good market, meaning enough people with the problem, and a product that satisfies them. Andreessen put the market first, under a law he credits to Andy Rachleff: “When a great team meets a lousy market, market wins.”
The essay is explicit about the debt. It presents that law as “Rachleff’s Law of Startup Success”, introduced “In honor of Andy Rachleff, formerly of Benchmark Capital, who crystallized this formulation for me”, and the sentence about fit as “Rachleff’s Corollary of Startup Success”. a16z's piece of 18 February 2017, by Tren Griffin, goes further: “The product/market fit (PMF) concept was developed and named by Andy Rachleff.” The spelling varies (product/market fit, product-market fit, market product fit), the shorthand is PMF, and Andreessen's essay calls the two halves of a startup's life BPMF and APMF, before and after fit.
What does product-market fit mean once you have to measure it? Andreessen describes a feeling, and feelings arrive late. Sean Ellis gave the term a number. On his own blog, in a post of 2 September 2009, he wrote: “I define product/market fit as at least 40% of your active users saying they would be “very disappointed” if they could no longer use your product.” That sentence is the 40% test, and it is what most tools mean by product-market fit.
Marc Andreessen on product-market fit: the 2007 essay
The essay's title is its claim, stated as Rachleff's corollary: “The only thing that matters is getting to product/market fit.” It describes both sides of the line. On the first: “You can always feel when product/market fit isn’t happening.” Its symptoms: “The customers aren’t quite getting value out of the product, word of mouth isn’t spreading, usage isn’t growing that fast”. On the other side, “The customers are buying the product just as fast as you can make it”, and “Money from customers is piling up in your company checking account.” It also states what is at stake: “Lots of startups fail before product/market fit ever happens.” The original sat on Andreessen's blog and lives on at pmarchive.com, an archive of those posts.
Product-market fit quotes, read where they were published
Six lines worth keeping, all read on 19 Nov 2026:
- Marc Andreessen, 2007: “When you are BPMF, focus obsessively on getting to product/market fit.”
- Andy Rachleff, as a16z quotes him in 2017: “You know you have fit if your product grows exponentially with no marketing.”
- Fred Wilson, in the same piece: “Getting product right means finding product/market fit. It does not mean launching the product.”
- Sean Ellis, 2009: “When surveys tell you that you’ve reached product/market fit, the final validation is charging for your product (or a version of it).”
- Rahul Vohra of Superhuman, 2018: “Politely disregard those who would not be disappointed without your product.”
- First Round Capital, 2024: “Product-market fit progresses in levels.”
The product-market fit survey: the Sean Ellis test and the 40% rule
The product market fit survey, also called the Sean Ellis test or the Sean Ellis survey, is one question with fixed answers. Rahul Vohra, the founder of Superhuman, summed it up in First Round Review on 12 November 2018: ask users “how would you feel if you could no longer use the product?” and measure the share who answer very disappointed. Where the 40% came from is in the same piece: “After benchmarking nearly a hundred startups with his customer development survey”, “Ellis found that the magic number was 40%.”
The survey tool that carries Ellis's name keeps the answers short. PMFsurvey.com, signed “Product/Market fit survey by Sean Ellis and GoPractice”, describes itself as free, and its demo survey offers “Very disappointed”, “Somewhat disappointed”, “N/A - No longer use” and “Not disappointed”. Who answers matters as much as the wording: Ellis wrote “active users”, and Vohra surveyed Superhuman's users “following Ellis’ recommendation to focus on those who used the product at least twice in the last two weeks.”
Run as a product market fit survey, the test is six steps:
- Pick the people. Users who have used the product recently and more than once, not every signup. People who tried it once cannot miss it.
- Ask the question first, with the answers fixed. Very disappointed, somewhat disappointed and not disappointed, plus a way out for those who stopped using it.
- Add three or four open questions. They explain the score (next section).
- Decide how no-longer-use answers count before you open the results. The choice moves the score, as the worked example below shows.
- Ask each person once. Vohra: “We were careful to ensure that we didn’t survey users more than once, so as to not throw off the 40% benchmark.”
- Read the very disappointed answers before the total. Who they are tells you where fit already exists.
Product-market fit questions: what to ask after the first one
The first question gives the score; the product market fit questions after it give the reasons. Superhuman's survey had four:
- “How would you feel if you could no longer use Superhuman?”
- “What type of people do you think would most benefit from Superhuman?”
- “What is the main benefit you receive from Superhuman?”
- “How can we improve Superhuman for you?”
PMFsurvey's demo runs seven, headed PMF Metric, Recommendation, Referrer, Alternatives, Benefits, Audience and Improvement. The referrer question deserves its place: crossed with the first answer, it shows which channel brought the people who would be very disappointed, and for a new product those channels are often the launch sites and directories it was listed on.
Product-market fit survey templates: 11 tools compared on 19 Nov 2026
Most product-market fit surveys start from a template, so the template's wording is the test founders run. We swept the sites of survey and product analytics tools for a product-market fit template and read the 11 that publish one, recording the question as each ships it, its answers, the threshold it prints, the fewest answers it asks for and who it says to ask:
Product market fit survey templates compared: 11 tools, their question and threshold (19 Nov 2026)
| Tool | The question, as shipped | Answers | Threshold | Fewest answers | Who to ask | Free use |
|---|---|---|---|---|---|---|
| Chameleon | “How disappointed would you be if you could no longer use [productname]?”Reworded | “Very disappointed”“Somewhat disappointed”“Not disappointed”Ellis's three | “Ellis discovered that companies achieving sustainable growth typically see at least 40% of respondents select "very disappointed."” | “Aim for at least 100 responses to ensure statistical significance, though 200-500 responses provide more reliable insights.” | “Focus on active users who have experienced your product's core value.” | Free template, account or plan |
| Formbricks | “How would you feel if you could no longer use [product]?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed”Ellis's three | “If 40% or more of your users say they would be "very disappointed" without your product, you have product-market fit.” | None printed | “Wait until users have completed onboarding and had at least one to two weeks of active usage.” | Free template, account or plan |
| PostHog | “How would you feel if you could no longer use us?”Ellis's wording | “Not disappointed”“Somewhat disappointed”“Very disappointed”Ellis's three | “40% of users saying "very disappointed" is a strong sign of product-market fit.” | None printed | “Use the PMF survey with existing users and run it monthly/quarterly.” | Free template, account or plan |
| RefinerAnswers from the interactive template; its text speaks of four options | “How would you feel if you could no longer use [product name]?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed”Ellis's three | “you can assume that you’ve achieved product market fit if at least 40% of users responded with “very disappointed” to your survey.” | None printed | None printed | Free template, account or plan |
| SprigQuestion from the template preview the page embeds | “How disappointed would you be if you could no longer use {product}?”Reworded | None printed | None printed | None printed | None printed | Free template, account or plan |
| SurveyMonkeyA page of its learn section, not a template gallery | “How would you feel if you could no longer use (product name)?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed (it isn’t useful)”Ellis's three | “The 40% rule states that if 40% of survey respondents say they would be “very disappointed” if they could no longer use the product, Product-Market Fit has been achieved, and you can expect sustainable growth.” | “We recommend a sample size of 40-50 responses to obtain meaningful results.” | “the respondent panel you choose must be diverse and consist primarily of potential customers.” | Free template, account or plan |
| SurveySparrow | “How will it affect you if our product is no longer available in the market?”Reworded | None printed | None printed | None printed | None printed | Free template, account or plan |
| SurvicateQuestion and answers from the template preview the page embeds | “How would you feel if you could no longer use our product or service?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed”“I no longer use your product”Ellis's three | “If it’s over 40%, congratulations” | None printed | None printed | Free template, account or plan |
| Tally | “How would you feel if you could no longer use [company]?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed (it isn't really that useful)”Ellis's three | None printed | None printed | None printed | Free template, account or plan |
| TypeformQuestion and answers from the template preview the page embeds | “How would you feel if you could no longer use our product?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“No reaction/neutral”Changed | “The industry benchmark for strong product market fit is 40%, meaning that at least 40% of respondents should say they'd be very disappointed if the product disappeared.” | None printed | “Target users who have been active within the last two weeks.” | Free template, account or plan |
| Zonka Feedback | “How would you feel if you could no longer use [Product Name]?”Ellis's wording | “Very disappointed”“Somewhat disappointed”“Not disappointed”“I no longer use it”Ellis's three | “If 40%+ of respondents say "Very disappointed," you have PMF.” | “At least 100 responses from users with meaningful product usage.” | “Send via email survey to users who've been active for 14+ days and completed at least one core workflow.” | Free trial |
Each row is one tool's own product-market fit template or learn page, read on 19 November 2026 and quoted as printed (curly and straight quotes kept). The question and answers come from the template preview the page embeds where it has one (Sprig, Survicate, Tally and Typeform), else from the page; Sprig, Survicate and Tally were also read in a headless browser, which shows each preview as a visitor sees it. Ellis's wording: "How would you feel if you could no longer use [product]?", whatever fills the product slot, as Vohra's piece and the PMFsurvey.com demo print it. Ellis's three: answers beginning "Very disappointed", "Somewhat disappointed" and "Not disappointed", plus at most a no-longer-use answer. Threshold, fewest answers and who to ask: the page's own line, or "None printed". Free use: our summary of the page's line (a free template, account or plan, or a free trial). Set aside by rule: Contentsquare, since Hotjar's template address now opens Contentsquare's page, which lists what the template finds out and prints no question wording; SatisMeter, since its PMF page paraphrases the question as asking how disappointed users would be if you stopped existing. Userpilot, Qualaroo, UserGuiding, Appcues, Attest and Pendo had posts about the survey in their sitemaps and no template page. A second template is not counted: Formbricks' "Product Market Fit (Superhuman)" and Zonka Feedback's "Sean Ellis Product Market Fit (PMF) Survey Template". Naming a tool endorses none, and no link here earns us anything.
Eight of 11 ask Ellis's question word for word, and seven keep his three answers with it. The other three reword it: Sprig and Chameleon ask “How disappointed would you be if you could no longer use [productname]?” (Sprig with {product} in the slot), which puts the answer's key word into the question, and SurveySparrow asks “How will it affect you if our product is no longer available in the market?” with no answer scale on its page. Typeform keeps the question and changes the scale: its template's third answer is “No reaction/neutral” where Ellis's scale has “Not disappointed”. Two of the nine that print answers add one for people who no longer use the product, as PMFsurvey does.
Eight of 11 print a threshold, and every one of them is 40%. They differ on where it comes from. Ellis's figure rests on nearly a hundred startups, as Vohra reports it; Formbricks' template says “The benchmark comes from analyzing hundreds of startups.” Typeform calls 40% “The industry benchmark”. Survicate's FAQ goes further and describes a second version under Ellis's name: “Ask your survey respondents to rate on a scale of 0-10 how disappointed they would be if they could no longer use your product or service”, scored at 7 or higher, with this rule: “According to the Sean Ellis method, if at least 40% of your respondents report experiencing a "must-have" experience, and your NPS is above 50, then you have achieved product-market fit.” Ellis's own definition of 2009 names one number, the 40% of active users, and Survicate's own template ships his question and his answers.
On who to ask, the templates part ways. Six of 11 say, and five of those point at users already in the product, from PostHog's “existing users” to Typeform's users “active within the last two weeks”. SurveyMonkey's page says the panel “must be diverse and consist primarily of potential customers” and suggests SurveyMonkey Audience to assemble one. Ellis's question asks people to imagine losing the product, which presumes they have it. As for cost, 10 of 11 offer the template, an account or a plan for free; Zonka Feedback offers a 14-day trial.
How many responses a product-market fit survey needs
Three of the 11 set a floor. SurveyMonkey: “We recommend a sample size of 40-50 responses to obtain meaningful results.” Zonka Feedback: “At least 100 responses from users with meaningful product usage.” Chameleon: “Aim for at least 100 responses to ensure statistical significance, though 200-500 responses provide more reliable insights.” Vohra, writing from Superhuman, argued for fewer than founders expect: “you start to get directionally correct results around 40 respondents, which is much less than most people think.” Ellis's own blog sets the bar lower still: under his 2 September 2009 post on charging, a reply signed “Sean” reads “I generally look for at least 30 responses, which is where I start to see results stabilize.” Survicate's free use, meanwhile, stops at 100 responses, the floor two other tools set.
The arithmetic settles the disagreement. A 40% score carries a margin, at 95% confidence, of about 18 percentage points either way from 30 answers, 15 from 40, 10 from 100 and 7 from 200. So 30 or 40 answers are enough when the score is far from the line, as Superhuman's first 22% was, and not enough when it sits at 35% or 45%. Survey more users of the same kind rather than lowering the bar to people who barely used the product.
10words: $150 for one sponsored week. Submitator: $29 for 40 directories.
10words is one of the 81 launch sites in our catalog: put it among the 40 picks of a $29 order and we file its free route for you, with a description written for the site and every badge kept live. 70 directories cost $49 and 110 cost $79, one payment each.
How to measure product-market fit: the score and the metrics around it
How to measure product market fit comes down to two kinds of evidence: what users say, which the survey captures, and what they do, which retention, usage and growth show. The survey is the faster read; behaviour is the slower one that confirms it.
The product-market fit score: a worked example
Fernquill is a made-up meeting-notes app. It emails the survey to the 150 users who opened it at least twice in the last two weeks, and 90 answer (60%). Scored over every answer, its product-market fit score is 36%; with the 8 people who no longer use it left out, as Zonka Feedback advises, it is 39%. Both sit below the line:
Product market fit score: a made-up survey of 90 answers, scored two ways
| Answer | Answers | Share of all answers | Share of those still using it |
|---|---|---|---|
| Very disappointed | 32 | 36% | 39% |
| Somewhat disappointed | 35 | 39% | 43% |
| Not disappointed | 15 | 17% | 18% |
| N/A, I no longer use it | 8 | 9% | Left out |
| Score | 32 | 36% | 39% |
Fernquill is made up, and so are its numbers. Shares are rounded to whole percent; the score is the "Very disappointed" share. All answers: 32 of 90. Still using: 32 of 82, the 8 no-longer-use answers left out.
Then Fernquill does what Vohra did and looks at who the 32 very disappointed users are. Team leads gave 40 of the 82 answers from people still using it and 19 of the very disappointed ones: 48% among leads, 31% among everyone else. That is a signal, not a verdict. At 40 answers the leads' 48% carries a margin of about 15 points, a range of 32% to 63%, so the next step is more team leads in the survey and a product built for them.
Product-market fit metrics beyond the survey
The survey asks for a prediction; product market fit metrics record behaviour. Three carry the most weight early on:
- Retention. Of the users who start in a given week, how many are still active weeks later. With fit, the curve flattens instead of sliding to zero. Churn and retention formulas, with benchmarks, are worked through in our post on what to measure in a SaaS.
- Usage. How often active users come back and whether they reach the core action. Ellis's survey assumes this: a person who barely used the product cannot be very disappointed to lose it.
- Growth you did not pay for. Rachleff's test, as a16z quotes it: “You know you have fit if your product grows exponentially with no marketing.” Referrals and repeat mentions are the early form of it.
Ellis put the order bluntly in 2009: “Metrics don’t matter until you achieve product/market fit”, and after it, “then they are critical to your success.” First Round adds a dimension most definitions skip, efficiency: whether each new customer can be found and served at a cost that works. Its thought experiment: “say you open up a stand on the side of the road where anyone who lines up and gives you $1 gets $100 in return.” Lines out the door, very happy customers. “But it would be a mirage.”
Signs of product-market fit, and how to know you have it
How to know if you have product market fit, in the order the evidence usually arrives:
- At least 40% of active users would be very disappointed to lose the product, on enough answers to trust the number.
- A group of users keeps using it week after week without reminders.
- New users arrive through people who already use it.
- People pay, and keep paying. Ellis called charging “the final validation”.
- Demand outruns you. Andreessen's version: “The customers are buying the product just as fast as you can make it”.
And the signs of its absence, from the same essay: “The customers aren’t quite getting value out of the product, word of mouth isn’t spreading, usage isn’t growing that fast”. First Round lists among its Level 1 yellow flags customers saying that if the product disappeared tomorrow “they wouldn’t be disappointed.”
Product-market fit stages: before, after and four levels
Pre product-market fit and after: Andreessen's two halves
Andreessen divides a startup's life in two, before product/market fit and after, and gives the first half one instruction: “When you are BPMF, focus obsessively on getting to product/market fit.” Pre product market fit, everything else (hiring, process, marketing spend) waits on that. Ellis's 2009 milestones say the same in sequence: validate the need, find who loves the product, then charge, then drive growth. In what stage does one find their product/market fit, then? In the first, search stage, before the company scales.
The four levels of product-market fit
First Round Capital argues that fit is not a single moment: “Product-market fit progresses in levels.” Its PMF Method names four product market fit stages, judged on satisfaction, demand and efficiency, and says the journey to the top usually takes anywhere from two to six years. The benchmarks it prints are for B2B software:
Product market fit stages: First Round's four levels and the B2B benchmarks it prints
| Level | Customers | ARR |
|---|---|---|
| 1. Nascent | “3-5 customers.” | “$0-$500K ARR.” |
| 2. Developing | “5-25 customers.” | “$500K to $5M ARR.” |
| 3. Strong | “25-100 customers.” | “$5-25M ARR.” |
| 4. Extreme | “100+ customers.” | “$25M+ ARR.” |
From First Round Capital's "Levels of PMF", the first session of its PMF Method (dated 4.8.2024 on the page), read on 19 November 2026. The customer and revenue lines are quoted as printed from each level's benchmarks, which First Round draws for enterprise companies "based on what we’ve seen from the hundreds of early-stage startups we’ve backed". Each level also prints team size, stage, demand source, conversion, churn, gross margin and burn multiple; ARR is annual recurring revenue.
The ladder runs both ways. First Round: “the journey to PMF is a two-way ladder, meaning levels aren’t permanently unlocked”. For a consumer or self-serve product the customer counts do not transfer, but the order does: satisfy a few, then find demand you can repeat, then make it efficient. As a product market fit framework for the first level, Vohra's engine is a concrete one, and it is the subject of the section after next.
Problem-solution fit vs product-market fit
Problem-solution fit comes first. It asks whether a real group of people has the problem, feels it enough to act and would take your proposed solution, and it can be checked before the product exists: through interviews, a landing page, a waiting list. Product-market fit asks whether the product, in use, holds a market, and it can be measured on people who use it, and on nobody else. Ellis's milestones start at the first one: “Before any coding begins it is important to validate that the problem/need you are trying to solve actually exists”. First Round's version of the same check is a question about the problem: “It’s the classic painkiller versus vitamin question.”
The difference in product market fit vs problem solution fit is the evidence. Problem-solution fit rests on what people say about a problem; product-market fit rests on what they do with a product, which is why the survey goes to active users and not to prospects. For testing a software idea before writing code, our census of small SaaS launches has a section on validating an idea.
How to validate product-market fit once the problem is real
Product market fit validation needs three things in order: first users who keep coming back, then a survey of them that clears 40% on enough answers, then payment. Ellis: “When surveys tell you that you’ve reached product/market fit, the final validation is charging for your product (or a version of it).” Which provider takes that first payment, and what it keeps of each sale, is laid out in our comparison of payment providers.
How to find product-market fit: the loop Superhuman ran
How to find product market fit is a loop, not a launch; Fred Wilson's line from the a16z piece makes the distinction: “Getting product right means finding product/market fit. It does not mean launching the product.” One loop published in full is Vohra's, which turned the survey into what he called an engine, in four steps:
- “Segment to find your supporters and paint a picture of your high-expectation customers.”
- “Analyze feedback to convert on-the-fence users into fanatics.”
- “Build your roadmap by doubling down on what users love and addressing what holds others back.”
- “Repeat the process and make the product-market fit score the most important metric.”
Product-market fit examples: Superhuman and Slack
Superhuman product market fit is a well-documented case, because Vohra published the numbers: “When we started this journey in the summer of 2017, our product-market fit score was 22%. After segmenting to focus on the very disappointed set of users, we were at 33%. Within just three quarters of our work to improve the product, the score nearly doubled to 58%.” The same piece cites a second example: Hiten Shah “posed Ellis’ question to 731 Slack users in a 2015 open research project”, and “51% of these users responded that they would be very disappointed without Slack”. Both are surveys of people who already used the product, which is where the loop has to start. Finding those first people is the subject of our guide to the first people who try a new product, and the launch that brings them is planned in a week-by-week launch plan.
Where Submitator fits: being listed where new products are found
Every method above starts with people who use the product, and a product nobody has found has nobody to ask. A new product's first visitors often come from the places people browse for new products, launch sites and directories, and the referrer question in the survey tells you which of them brought the users who would miss it. Ellis listed the same channel in his 2009 milestones, at the point where a startup turns to growth: “Start by building out free channels such as listing in directories and basic SEO.”
Our catalog holds 202 such sites, each with its Domain Rating in the catalog table: 81 launch sites and startup directories, 62 SaaS directories and 59 AI tool directories. 10words is one of the launch sites. Its home page sums it up as “Discover new apps and startups in 10 words or less” and invites makers to “Get your 10 words ready and submit it for free.” Its sponsorship page prices the other route: “At just $150 for a full week sponsorship”. That one payment covers “Premium placement of your 10 word project description in all emails” (the Friday edition goes to 11,047 subscribers) plus “Premium placement on our homepage”: one week of premium placement for $150. When we read the page on 19 Nov 2026, its availability block listed no open week and offered “Get notified when new slots open up”. Choose 10words as one of the 40 sites of a $29 order and it goes in through the free route (our guide to submitting there has the details).
The $29 order works like this. You pick the directories in a dashboard that prints each one's Domain Rating next to its name, from a catalog that takes in a site after one of our submissions has run through its form from start to finish. Filling those forms is on us: every site gets a description of its own, within the length limit it sets, and CAPTCHAs never reach you. Where a directory makes a badge the price of its listing, the badge comes from a single component on your site that serves all of them; we keep each badge live, and you never commit code for one directory. As each listing goes up, we push its page through IndexMachine, the free indexing that comes with every plan ($25 value), with nothing for you to connect; whether Google indexes a page stays Google's call, as our post on how Google picks up new links explains. As for ads and outreach, we run neither.
Starter costs $29 for 40 directories, Growth $49 for 70 and Scale $79 for 110, paid once each; All Directories, $149, takes the whole catalog, later additions included. Most directories are delivered in 24 to 48 hours, every directory within 21 days at most. A rejected submission earns you, by default, one more directory of your choice (All Directories excepted, since it holds every one already). A refund can be requested within 30 days of purchase, on the conditions in our terms. We have submitted 580+ startups so far, and the figures from our own pipeline follow what became of those submissions.
All reads behind this page were GET requests made on 19 November 2026 between 20:59 and 21:44 UTC: curl for most pages, plus a headless browser for the Sprig, Survicate and Tally previews, PMFsurvey.com with its demo and the 10words sponsorship page, whose text arrives by script. Nothing was submitted, signed into or answered. Quotations keep their page's wording and punctuation and stop where a quote reaches a long dash. The 11 tools of the census were found by searching each vendor's sitemap for product-market fit pages and sorted by rules fixed before counting (see the census table footnote), and the counts in the copy are worked out from those rows when the site is built: 8 with Ellis's wording, 7 with his wording and his three answers, 8 printing 40%, 3 setting a minimum and 6 saying who to ask. Margins of error are 95% intervals by the normal approximation, 1.96 × the square root of p × (1 - p) ÷ n. Sources: Marc Andreessen, "The only thing that matters" (25 June 2007), at pmarchive.com; Sean Ellis, "When Should a Startup Start Charging?" (2 September 2009) and "Milestones to Startup Success" (30 November 2009), on startup-marketing.com, read on its Product-Market Fit category page; Rahul Vohra in First Round Review (12 November 2018); a16z, "12 Things About Product-Market Fit" by Tren Griffin (18 February 2017); First Round Capital, "Levels of PMF" (dated 4.8.2024); PMFsurvey.com and its demo. 10words: its home page and its sponsorship page, the latter read at 20:59 UTC, a minute later after a longer wait and again at 21:31 UTC, with identical text each time: "At just $150 for a full week sponsorship" for "a week-long exclusive sponsorship", with "Premium placement of your 10 word project description in all emails" and "Premium placement on our homepage"; under "Current Availability" no week was listed. The comparison sets one payment against one payment and says nothing about when a week is free; a Submitator plan does not buy a 10words sponsorship, and 10words goes in through its free route, as its guide on this site describes. Fernquill and every number about it are invented. The catalog counts are taken from the export of 17 November 2026, with intercepts excluded. What this page says about Submitator rests on three sources. The owner's word on how orders run: descriptions written for each site within its limits, CAPTCHAs kept off the customer's desk, badges served from one component and kept live, listing pages pushed through IndexMachine on every order, delivery of most directories within 24 to 48 hours and catalog admission after one full submission. The pricing data: prices, plan sizes, the IndexMachine bonus and the 580+ startups. The terms: one-time payment, the reach of All Directories, the 21-day ceiling, the extra pick when a directory rejects and the 30-day refund window. No sentence here promises fit, signups, traffic, revenue, indexing or a ranking.
FAQ
What does PMF stand for in a startup?
Product-market fit, which Marc Andreessen spelled product/market fit in his 2007 essay: being in a good market with a product that can satisfy it. The same essay calls a startup's life before and after it BPMF and APMF. A "PMF company" or a post-PMF startup is one past that line; by Sean Ellis's measure, at least 40% of its active users would be very disappointed to lose the product.
How do you calculate a product-market fit score?
Count the "very disappointed" answers to Ellis's question, divide by the answers and multiply by 100. PostHog's template puts it as “Divide the number of users who responded "very disappointed" by the total responses.” Settle one thing before counting: people who say they no longer use the product. Zonka Feedback's template says “Filter out users who selected "I no longer use"”, which shrinks the base and lifts the score; in the worked example on this page that choice moves a made-up app from 36% to 39%. Either way the line is 40%.
How many responses does a product-market fit survey need?
Rahul Vohra wrote in 2018 that “you start to get directionally correct results around 40 respondents, which is much less than most people think.” Of the 11 survey tools we read on 19 Nov 2026, SurveyMonkey asks for 40 to 50 and Zonka Feedback and Chameleon for at least 100. The arithmetic explains the gap: a 40% score from 40 answers carries a margin of about 15 points either way, from 100 answers about 10. Forty answers settle a score far from the line; a score near 40% needs a hundred or more.
In what stage does a startup find product-market fit?
Early, and before it scales. Andreessen splits a startup's life into before and after product/market fit and tells founders in the first half to “focus obsessively” on reaching it. Sean Ellis's 2009 milestones put fit after the minimum viable product and before driving growth, and First Round's first level, Nascent, is a B2B startup with 3 to 5 customers. Fit can also slip later, so it is a state to keep, not a stage to pass once.
Does product-market fit automatically result in high-volume sales?
No. Fit says a market wants the product; volume also needs a way to reach that market at a cost that pays. a16z's 2017 piece on fit says “It signifies an important milestone that is necessary but not sufficient for success.” First Round makes efficiency a dimension of fit in its own right and gives the example of a roadside stand that hands back $100 for every $1: lines out the door and very happy customers. “But it would be a mirage.”
What are the four levels of product-market fit?
First Round Capital's PMF Method names them Nascent, Developing, Strong and Extreme, each judged on satisfaction, demand and efficiency. For B2B companies it prints 3 to 5 customers at the first level, 5 to 25 at the second, 25 to 100 at the third and 100 or more at the fourth, and says the climb usually takes anywhere from two to six years.
Can a startup lose product-market fit?
Yes. In a16z's 2017 piece Ben Horowitz lists “Once you achieve product/market fit, you can’t lose it” among the myths about fit, and First Round says “the journey to PMF is a two-way ladder, meaning levels aren’t permanently unlocked”. Markets move, rivals ship and new users arrive with other needs. Superhuman, by Vohra's account, kept surveying new users to track its score as it climbed from 22% to 58%.
Keep reading
Their $150: one sponsored week. Our $29: 40 directories.
10words' $150 pays for one week of premium placement in its emails and on its homepage. Our $29 submits your product to 40 directories you choose, or 110 for $79, paid once, each with a description of its own and every badge they ask for kept live.
