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.
Five signals tell you whether you have product-market fit: your survey score, revenue growth, free-to-paid conversion, retention and organic growth, each read against its benchmark. When one is weak, PMFtracker finds where users drop, asks them why, and tells you what to test next.
- 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 week
MRR $2,940 · 17 paying customers
YC pace 5-7% a weekBelow
▲ from +1.8% a week
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 is below YC pace. 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.
0 in 3
seed startups ever reach Series A
The gate is fit, and investors know it. The ones who pass show a number, not a narrative.
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, compounded over 4 weeks, with its noise band.
5-7% a week is a good rate during YCPaul Graham, Startup = Growth →
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, then tracking
Paying customers stay
Monthly paid churn now, then weekly active retention by signup cohort as events build up.
Your own curve, cohort by cohortRetention 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.
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?"
03December
Whether
Will the fix work?
The fix cut to its smallest version, 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.
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
04December
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 card? Every card 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.
- AI analysis of every open answer
- 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: this is the same math PMFtracker runs on your real responses, the Sean Ellis score, the 40% line, and a 95% confidence band that tightens as your sample grows.
33%
± 13 pts
Below the line, for now
40 to 99: emerging read
could be 20% to 46%
95% interval. 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. We automated the one these operators and investors publicly champion. Here's what they say about the method behind it.
"The insights we gained from the Product Market Fit Survey in 2008 at Dropbox and Eventbrite were incredibly helpful for figuring out how to deliver the right experiences to the right customers."
"It's hard to overstate the impact of this product-market fit engine on our company. Everything we do at Superhuman, from hiring to selling and marketing to raising capital, has become significantly easier."
"The Sean Ellis score is a key element of how Nubank measures product-market fit and makes decisions about investing in new products."
"Back in April, we ran this very survey. In total, we got 138 responses. The feedback we received in the open-ended text fields has driven a lot of upcoming product features and improvements."
"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
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 and Why
- 1 active PMF survey, unlimited responses
- Advanced PMF scoring (40% test)
- Custom branding
- AI-powered sentiment analysis
- VC showcase access
- PostHog Cohorts integration
- Import from external survey tools
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:
- AI-powered sentiment analysis
- VC showcase access
- PostHog Cohorts integration
- Import from external survey tools
Bonus: 1-hour strategy session €150 valueWith 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: the PMF Monitor with its five signals and the verdict, the tracking script, Stripe sync, CSV import, and the Where and Why steps of the conversion loop, including surveys shown only to the users who drop. Rolling out in December: the Whether and Learn steps, meaning test cards sized to your traffic and learning cards. 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 confidence band shows exactly how much to trust the number at every sample size. 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.
Stop guessing. Start tracking PMF.
See your PMF signalsPay once, own it forever · 2-week free trial · No credit card