Users would be very disappointed
34%
20 of 58 responses
Benchmark 40%Approaching
▲ from 28% four weeks ago
New: the PMF Monitor and the 3W loop
PMF analytics for founders without an analyst.
See where users leave before they pay, what your paying users did differently, and how to test your next change. Set up with one script and a read-only Stripe key.
- Revenue and churn from Stripe on day one
- 2-week free trial, no credit card
- Pay once, own it forever
Verdict
Getting closer
The survey score and free-to-paid conversion improved over the last 4 weeks, beyond their noise band. Nothing got worse.
Revenue compounds
+3.1%a month
MRR $2,940 · 17 paying customers
10% a month on trackBelow
Free users become paying
3.8%
12 of 312 signups paid in 30 days
Freemium 3-5% goodAt benchmark
▲ from 2.6%
Paying customers stay
Not enough data yet
Needs 20 paying customers. You have 17.
Shown at 20+Waiting
Growth without pushing
39%
121 of 312 signups organic or referred
Against your own trendRising
▲ from 31%
Revenue growth is below 10% a month. Retention needs 3 more paying customers, so the loop starts with free-to-paid conversion.
The problem
Most startups fail because nobody needs them.
0%
of failed startups cite "no market need"
The #1 killer in CB Insights' post-mortem data. Not funding. Not the team. Nobody wanted the thing.
0%
of tech startups scale prematurely
Startup Genome's data: hiring and spending on growth before fit is proven. Burning runway to fill a leaky bucket.
The PMF Monitor
One survey can't measure product-market fit. Five signals can.
Product-market fit shows up in more than one place. Each signal is read against its own benchmark, with the counts behind it, and stays hidden until there is enough data to mean something. No blended score.
Signal
Source
Benchmark
Shown only with
01
In-app survey
Users would be very disappointed
The Sean Ellis survey score, asked inside your product.
40% very disappointedThe 40% rule →
40+ responses
02
Stripe, read-only
Revenue compounds
MRR growth a month, compounded over the last 4+ weeks.
10% a month on track, 15% strong. PMFtracker's own bar.
10+ paying customers
03
Tracking or CSV, plus Stripe
Free users become paying
Free-to-paid conversion by signup cohort, in a 30-day window.
Freemium 3-5% good, 6-8% great. Free trial 8-12% good, 15-25% great.Lenny Rachitsky and Kyle Poyar, 2023 survey →
30+ signups in the window
04
Stripe, read-only
Paying customers stay
Paid churn: the share of paying customers who stopped in the last 30 days.
Lower is better, month by monthRetention and PMF →
20+ paying customers
05
Tracking, first touch
Growth without pushing
The share of new signups that came organically or by referral.
Your own trend
30+ signups in the window
The verdict
Not enough data yet
Slipping
Getting closer
Holding
Not yet
One of five, picked by fixed rules from how the signals moved over the last 4 weeks. Every signal has Improve this, which starts the loop below.
How it works: the 3W loop
Find the leak. Ask why. Test the fix.
Improve this on free-to-paid conversion opens a loop that runs the way a good analyst would, in four steps. Numbers below are from the same example project.
01Live
Where
Where do people stop?
The leak map from signup to paid, with counts at every step. Steps are picked from your most-viewed pages and most-clicked buttons, no code per step.
Leak map · 30 days
- Signed up 312
- Created a project 212
- Used the core feature 94 · −118
- Paid 12
Used Templates in week 1: 8 of 61 paid (13.1%) vs 4 of 251 (1.6%)
This is a correlation. Test it first.
This is a correlation. Test it first.
02Live
Why
Why do they stop there?
Three competing explanations, drafted from your numbers. Then you ask the users themselves: a short survey shown only to the people who stopped, and an interview list with an invite to copy. The drafts, the survey and the interview list are in PMF Startup Kit.
Three explanations
- AThey never reached the moment it pays off
- BThe core feature needs data they don't have yet
- CThey signed up to look, not to use it
"Tell me about the last time you opened your project. What happened next?"
03Live
Whether
Will the fix work?
The fix cut to its smallest version, its riskiest assumption, and a test card sized to your traffic. If an A/B test would take a year, the card says so and picks a method that can finish. In PMF Startup Kit.
Test card
- Change
- Templates on the first screen
- Metric
- Paid within 30 days
- A/B, +45%
- about 6,100 users
- Your traffic
- 75 signups a week
- So
- Ship it, compare cohorts
04Coming
Learn
Did it work?
When the window closes, the learning card fills in from your data: weekly signup cohorts before and after the change, and a call to keep, kill or iterate.
Learning card
- 4 cohorts before
- 3.8% paid
- 4 cohorts after
- 5.1% paid
- Call
- Keep
- Next Where
- Paid churn
The method behind it. In a ProductTrio sprint at UX Pilot, users who tried Deep Design in their first 7 days paid at 12.3%, everyone else at 1.3%. That gap became a test on 120,000 users, which measured +44.7% free-to-paid. Real, and far smaller than the gap suggested: that's why every gap here is labeled a correlation. Read the case →
Stuck on a draft? In PMF Startup Kit, every AI draft has Request expert review. It goes to Olek Uznanski, who has trained 300+ startups on product-market fit, and his reply lands in the card.
Built for honest answers
Analytics that won't flatter you.
Small startups have small samples. Most dashboards hide that. PMFtracker shows it, on every number.
Every rate shows its counts
A percentage without its sample size hides how far to trust it. Every rate comes with the counts behind it.
Thin data stays hidden
A signal shows only once it can mean something: 40 survey responses, 30 signups, 10 paying customers. Until then it says how much more it needs, never a misleading 0%.
Gaps are labeled, then tested
Users who do X may pay more because of who they are, not because of X. Every behavior gap carries that label, and appears only with 250+ signups and 10+ payers.
The verdict follows fixed rules
Not enough data yet, Slipping, Getting closer, Holding or Not yet: picked by rules checked in order. The AI only writes one sentence that explains it.
AI drafts can't invent figures
Every number in a drafted explanation has to come from your data. A draft that adds one is rejected and written again.
Product analytics tools are good at showing where a funnel leaks. PMFtracker is built for the next two questions: why people stop, and whether your fix worked.
Setup
Useful before you install anything.
Setup is where analytics tools usually stall. So PMFtracker starts with data you already have, then asks for one script.
Connect Stripe
Paste a restricted read-only key. PMFtracker can read your billing history and can never change anything in your account.
- Revenue growth from your Stripe history
- Paid churn, before a single event arrives
- Add a user CSV (email, signup date) and conversion by signup month appears too
Add one script and one line
Page views and clicks record themselves. You pick the funnel steps in the app.
<!-- before </body> --> <script async src="https://app.pmftracker.com/pmf.js" data-key="pk_..."></script> // after login window.pmf?.identify(user.id, { email: user.email });
- HTML, Google Tag Manager, React, or a ready prompt for Lovable and Bolt
- A live check confirms the first page view and the first identify
Ask the right users
The PMF survey widget, shown inside your product to engaged users, or only to the people who stopped at one step (PMF Startup Kit).
- AI analysis of every open answer (PMF Startup Kit)
- Who loves the product, and what holds the rest back
- Your PMF score, tracked over time
Page paths only, never query strings Never records typed input Emails and long numbers masked in click text Cookieless mode available
Signal one, in your hands
Feel your PMF score move.
The survey score is still the signal founders know best. Drag the sliders to see the Sean Ellis score against the 40% line, and how much a small sample can swing it. In PMFtracker the score stays hidden until 40 responses, and it is always shown with its counts.
33%
± 13 pts
Below the line, for now
40 to 99: emerging read
could be 20% to 46%
The range in this demo is a 95% margin of error, to show how sample size matters. Below 40 responses, treat the score as directional. At 100+, it is investor-grade.
Get your PMF scoreAlready have answers? Score them in the free PMF calculator → No survey yet? Copy the Sean Ellis survey template →
The methodology
We didn't invent the method.
We built its instrument.
PMFtracker didn't invent the product-market fit survey. Sean Ellis did, and Superhuman's Rahul Vohra built a product-market fit engine around it. PMFtracker automates the method.
"It is hard to overstate the impact of this product-market fit engine on our company."
"I've trained 300+ startups on this exact method. PMFtracker is the tool I always wished they had."Olek Uznanski, Founder of PMFtracker
Who this is for
Worth your time, or worth skipping.
The signals only tell you something real under the right conditions. Better to know that now than after you've set it up.
The full suite is for you if
- People sign up for your product on their own, every week
- You bill through Stripe
- Nobody on the team has the time to read the data
- You want the next step, not another dashboard
Start with the survey if
- You've launched and people use it, but signups are still a trickle
- You sell through demos and contracts, not self-serve signup
- You need a PMF score your investors will trust
This is not for you if
- You're pre-launch or sitting on a waitlist. Measure fit after people use the thing. Do this instead →
- You have users but no way to contact them
- You want a number for a deck more than you want the truth
What founders ask
Sound familiar?
Four questions founders ask once real people sign up and the numbers still say nothing clear. Each one has an answer below.
Where they leave
See the step where people stop.
The leak map runs from signup to paid through your own steps, with the count at every step and the biggest drop marked.
- You pick the steps in the app, from your most-viewed pages and most-clicked buttons. No code per step.
- Every drop shows its counts, so a big percentage from a handful of people can't fool you.
LiveOn every plan
Where people stopSample data
People who signed up 30 to 58 days ago
- Signed up312
- Created a project212-32.1% from the previous step (312 to 212)
- Used the core feature64-69.8% from the previous step (212 to 64) Biggest drop
- Opened pricing29-54.7% from the previous step (64 to 29)
- Paid12-58.6% from the previous step (29 to 12)
In all, 12 of 312 signups (3.8%) paid within 30 days of signing up.
They sign up, they don't pay
What your paying users did differently.
Your analytics can count paying users. PMFtracker describes them: the pages and clicks of their first 7 days, next to everyone who didn't pay.
- A behavior shows only when at least 30 people did it and the difference is unlikely to be chance.
- It needs about 250 signups and 10 payers in the window. Below that, the app shows how close you are instead of guessing.
- Every gap is labeled a correlation. Test it before you build on it.
LiveOn every plan
What paying users didSample data
First 7 days after signing up. 312 people who signed up 30 to 58 days ago, 12 of them paid.
BehaviorDid itDid notp
Clicked Use template on /projects/new8 of 61 paid (13.1%)4 of 251 paid (1.6%)<0.001
Visited /integrations5 of 48 paid (10.4%)7 of 264 paid (2.7%)0.024
With 14 behaviors tested, about 1.4 can pass p < 0.10 by chance.
This is a correlation. Test it before you change the product.Why, in their words
Ask the people who stopped.
Two questions, shown inside your product only to the people who stopped at the drop and to the people who got through. You read both groups side by side, with counts.
- The answer options start as three explanations drafted from your numbers. You edit them before the survey goes out.
- An option stands out only when the two groups differ by more than chance.
- Then talk to them: an interview list with an invite to copy.
LivePMF Startup Kit. On PMF Starter you see this step with an upgrade note. See what each plan includes
Ask the usersSample data
- Stopped at Created a project
- 14 answered of 41 asked
- Got through to Used the core feature
- 11 answered of 23 asked
Question 1When you reached Created a project, which of these comes closest to your experience?
AnswerStoppedGot through
I didn't get to the point where it paid off3 of 14, 21%3 of 11, 27%
I didn't have the data it needs yet Stands out8 of 14, 57%1 of 11, 9%
I was only looking around2 of 14, 14%4 of 11, 36%
None of these1 of 14, 7%3 of 11, 27%
Self-reported answers. Compare them with what people do before you act.
Did it work?
Size the test before you ship.
At small traffic, most changes ship without a fair test. PMFtracker plans one around your own traffic before you ship, so you know what to compare and when.
- Pick a fix, find its riskiest assumption, and get a test card: the baseline, the target, how many people it needs, how many weeks that takes, and when to read it.
- When an A/B test could not finish at your traffic, the card says so and picks a method that can.
LiveTest cards are in PMF Startup Kit. See what each plan includes
ComingReading the result in the app is coming next. Until then, the card tells you what to compare and when to look.
Test cardSample data
Your traffic: about 78 signups a week. Method: Interviews first, then ship and compare cohorts.
- Now
- 30.2% 64 of 212
- Target
- 39.2% +30%
- People
- 433 before, and 433 after
- Read on
- 13 wk after you ship
- We believe that
- people who start from a template reach the core feature, because it needs data they don't have yet.
- To verify that, we will
- Ship templates on the first screen and compare the step for people who signed up before and after we ship.
- And measure
- People who reach Used the core feature after Created a project: 30.2% now (64 of 212).
- We are right if
- It reaches 39.2% or more.
- By
- After about 9 weeks of people and 30 days for them to act.
Noise, or a verdict?
"Not enough data yet" is a feature.
Small startups have small samples. PMFtracker would rather say so than hand you a number that is noise. Three rules it never breaks:
No score before 40 responses
Below 40 survey responses you see how many more you need, not a score. A number you show your team or investors should not be noise from 12 answers.
Counts next to every percentage
Every rate comes with the people behind it, so you can see how much it rests on before you act on it.
AI drafts can't invent figures
Every number in a drafted explanation has to come from your data. A draft that adds one is refused and written again.
Setup
Set up in one paste.
One read-only key, one script, one survey. No analyst, and no code for each step.
Paste a read-only Stripe key
Revenue growth and paid churn from your billing history, before a single event arrives. PMFtracker can never change anything in your account.
Paste one script
<script async src="https://app.pmftracker.com/pmf.js" data-key="pk_..."></script>
Plus one identify line after login. Works with HTML, Google Tag Manager, React, or a ready prompt for Lovable and Bolt.
Turn on the survey
The PMF survey, shown inside your product to engaged users. Your score appears at 40 responses, always with its counts.
Page paths only, never query strings Never records typed input Cookieless mode available
See your PMF signals
Pay once, own it forever · 2-week free trial · No credit card
Why I built it
A note from Olek.
I have trained 300+ startups on product-market fit. PMFtracker runs the same loop as the conversion sprints at ProductTrio, my other company: find where free users stop, ask them why, then test one fix.
Most founders I meet have analytics installed and nobody to read it. They don't want another dashboard. They want a straight answer about their own product, and the next thing to do. So I built PMFtracker to run that loop on your numbers.
It refuses things on purpose: no score before 40 responses, no rate without its counts, no gap called a cause without a test.
Lifetime access
Pay once. Track forever.
No subscription, no recurring fees. Pay once, own it forever.
PMF Starter
Measure PMF like a pro
$99$79
- PMF Monitor: five signals and the verdict
- Tracking script and Stripe revenue sync
- Conversion loop: Where, the leak map
- 1 active PMF survey, unlimited responses
- Advanced PMF scoring (40% test)
- Import from external survey tools
- PostHog Cohorts integration
- Custom branding
- AI-powered sentiment analysis
- Why and Whether: AI drafts, surveys to the users who stopped, test cards
- VC showcase access
2-week free trial included · No card needed
PMF Startup Kit
Find PMF faster with AI + expert guidance
$199$149
- Everything in PMF Starter, plus:
- Unlimited active surveys
- AI-powered sentiment analysis
- Why: three explanations drafted from your numbers, a survey to the users who stopped, an interview kit
- Whether: pick a fix, find its riskiest assumption, size the test
- Expert review on any AI draft
- VC showcase access
Bonus: 1-hour strategy sessionWith Olek Uznanski, product leader who's trained 300+ startups on PMF. Included free.
2-week free trial included · No card needed
30-day money-back guarantee
Secure payment via Stripe
Instant access after checkout
Would rather not run it yourself? We'll do the whole thing in four weeks →
FAQ
Fair questions.
What is PMF analytics?
Product analytics counts clicks. PMF analytics answers one question: do people need this, and is it getting better? PMFtracker reads five signals, the Sean Ellis survey score, revenue growth, free-to-paid conversion, retention and organic share, each against its own benchmark. When one is weak, Improve this starts a loop on it.
Do I need PostHog, Mixpanel or another analytics tool?
No. PMFtracker has its own tracking script: one tag, plus one identify line after login. Revenue comes from Stripe through a restricted read-only key. If you already use PostHog, you can still bring its cohorts in.
What does the tracking script collect?
Page views, and clicks on links and buttons with their visible text. It records page paths, never query strings, never records typed input, and masks email addresses and long numbers in click text. You choose the consent mode: cookieless, wait for consent, or remember visitors across visits.
What's live today, and what's coming?
Live on every plan: the PMF Monitor with its five signals and the verdict, the tracking script, Stripe sync, CSV import, and the Where step of the conversion loop. Also live in PMF Startup Kit: the Why step, with drafted explanations, surveys shown only to the users who drop and an interview kit, and the Whether step, which picks a fix, finds its riskiest assumption and sizes a test card to your traffic. Coming next: Learn, which reads the result of your test. Loops for retention come after that.
How is the PMF score calculated?
One question: "How would you feel if you could no longer use this product?" The share of engaged users answering "very disappointed" is your score. N/A answers from churned users are dropped so they never distort the result. 40% or more is the benchmark for product-market fit.
How many responses do I need?
40 valid responses gives a directional read, and the score stays hidden on the Monitor until you have them. 100+ makes it reliable enough for investors. The score is always shown with its counts, for example 20 of 58 responses, so you can see how much it rests on. More on sample size →
Is it really a one-time payment?
Yes. Pay once, own it forever. No subscription, no recurring fees, and a 2-week free trial with no credit card up front. Payment runs through Stripe.
What if it turns out it's not for me?
Start with the 2-week free trial, no card needed, so you'll know before paying anything. And if you buy and change your mind, there's a 30-day money-back guarantee: email hello@pmftracker.com and you get a full refund. No forms, no interrogation.
FAQ
The questions founders ask.
How do I know if I have product-market fit?
No single number settles it. Fit shows up in more than one place: people would be very disappointed without your product, free users start paying, paying customers stay, and new users arrive without you pushing. PMFtracker reads each of these against its own benchmark, for example 40% "very disappointed" on the Sean Ellis survey, and fixed rules turn them into one verdict: Not enough data yet, Slipping, Getting closer, Holding or Not yet. The 40% rule →
Is the 40% survey valid with only 30 or 40 users?
At about 40 valid responses it gives a directional read, and from 100 it is reliable enough for investors. With 30 answers, one person moves the score by more than 3 points. So PMFtracker keeps the score hidden until 40 responses, says how many more you need, and always shows the score with its counts, for example 20 of 58 responses. More on sample size →
How many users or how much MRR until I have PMF?
No user count or MRR figure means fit on its own. What matters is how people behave: whether they would miss the product, whether free users pay, whether paying customers stay. PMFtracker shows each signal only once there is enough data behind it, and until then tells you how much more it needs.
My users say they love it but don't pay. Do I have PMF?
Half of it, maybe. Saying and paying are separate signals: the survey score shows how much people would miss the product, free-to-paid conversion shows whether they pay for it. When the two disagree, start at the step where people stop before paying. The leak map shows that step on every plan. Asking the people who stopped why, and sizing a test for the fix, are in PMF Startup Kit.
Do I need an analyst or a developer to set this up?
No analyst. Stripe takes a restricted read-only key that you paste in. The tracking script is one tag plus one identify line after login, with instructions for HTML, Google Tag Manager and React, and a ready prompt for Lovable and Bolt. Funnel steps are picked in the app from your most-viewed pages and most-clicked buttons, so there is no code per step.
What if I don't have enough data yet?
Then PMFtracker says so, and says how much more each signal needs, never a misleading 0%. The survey works from your first users, so with only a trickle of signups, start there and talk to the people who answer. If you are pre-launch or on a waitlist, this is not for you yet: measure fit after people use the product. Do this instead →
What's live today, and what's coming?
Live on every plan: the PMF Monitor with its five signals and the verdict, the tracking script, Stripe sync, CSV import, and the Where step of the conversion loop. Also live in PMF Startup Kit: the Why step, with drafted explanations, surveys shown only to the users who drop and an interview kit, and the Whether step, which picks a fix, finds its riskiest assumption and sizes a test card to your traffic. Coming next: Learn, which reads the result of your test. Loops for retention come after that.
What does the tracking script collect?
Page views, and clicks on links and buttons with their visible text. It records page paths, never query strings, never records typed input, and masks email addresses and long numbers in click text. You choose the consent mode: cookieless, wait for consent, or remember visitors across visits.
Is it really a one-time payment?
Yes. Pay once, own it forever. No subscription, no recurring fees, and a 2-week free trial with no credit card up front. Payment runs through Stripe.
What if it turns out it's not for me?
Start with the 2-week free trial, no card needed, so you'll know before paying anything. And if you buy and change your mind, there's a 30-day money-back guarantee: email hello@pmftracker.com and you get a full refund. No forms, no interrogation.
Stop guessing. Start tracking PMF.
See your PMF signalsPay once, own it forever · 2-week free trial · No credit card