How to Know If You Have Product-Market Fit (Not the Theory, the Actual Signals)
Product-market fit is one of the most referenced concepts in startup building and one of the most loosely defined in practice. Founders talk about "feeling" it. Investors talk about "seeing" it in the metrics. Neither framing is specific enough to be useful when you are inside a company trying to figure out whether you have it or are still searching.
In my experience working at early-stage companies, PMF is not a moment. It is a threshold crossing that shows up in a specific pattern of signals — some quantitative, some qualitative — that appear together and reinforce each other. Here is what those signals actually are.
The Three Quantitative Signals
Day 30 retention above the category baseline. The most reliable quantitative signal of PMF is cohort retention. Not average retention across all users — cohort retention for users who signed up in a specific month, measured at 30 days.
The threshold varies by category. For consumer products, Day 30 retention above 20 to 25% is considered strong. For B2B SaaS, Day 30 retention above 40% is the benchmark worth targeting. Andreessen Horowitz's benchmarks put top-quartile B2B SaaS retention at Day 30 in the 50 to 60% range, though these numbers are skewed by large enterprise products.
The reason retention is the best early PMF signal is that it measures whether users found enough value to come back, without coaching or prompting from the team. Activation metrics can be gamed with good onboarding. Retention is harder to manufacture.
Net Revenue Retention approaching or above 100%. If existing customers are spending more over time — through seat expansion, tier upgrades, or usage growth — the product is delivering increasing value as customers deepen their use. NRR above 100% means the revenue from your existing customer base is growing without adding new customers. That is only possible if the product is solving a real problem that grows in value as the customer gets more embedded in it.
NRR below 90% before PMF is a warning sign that retention is fragile and the product has not yet solved the problem durably enough for customers to stay and expand.
Organic inbound representing more than 30% of new pipeline. When people start finding the product without paid acquisition — through word of mouth, press, community, or search — it means something about the product is worth talking about unprompted. This is one of the clearest leading indicators that PMF is approaching. Paid acquisition can produce users. Only genuine value produces users who refer other users.
The Two Qualitative Signals That Matter More Than Surveys
What customers say when they cancel. Cancellations are the most honest customer feedback you will get. When a customer churns, they have nothing to gain by being diplomatic. What they tell you when they cancel — if you ask the right way, with a direct call rather than a cancellation survey — tells you the specific gap between what the product promised and what it delivered.
In my experience, early-stage teams with PMF will hear churn reasons that are circumstantial: budget cuts, team changes, company direction changes. Teams without PMF will hear product reasons: it did not do what we needed, we found something else that worked better, we went back to our previous process. The specificity and pattern of churn reasons is more diagnostic than the churn rate itself.
What customers say when they recommend. The Sean Ellis test — asking customers "how would you feel if you could no longer use this product?" and looking for 40% or more to say "very disappointed" — is a useful rough gauge. But the more useful signal is what customers say when they are telling a colleague about the product unprompted.
If your best customers describe the product with a specific outcome ("it cut our reporting time from two hours to fifteen minutes"), you have a tight value proposition that customers understand. If they describe it with a category ("it's a good data tool, you should check it out"), you have awareness but not differentiated value.
Ask your five most retained customers: what do you tell colleagues when you recommend us? If the answers are specific and outcome-oriented, PMF is close or present. If the answers are vague or category-level, the product has not yet crystallized into a specific, communicable value.
How PMF Looks Different for Bootstrapped vs. Funded Teams
Funded teams have a particular PMF trap: they can sustain growth through paid acquisition long enough to mistake traction for fit. When you have money to spend on ads or sales, you can generate users and revenue that look like PMF but are actually purchase-decision-dependent. The product is not pulling people in — the sales and marketing motion is pushing them in. When the spend stops, growth stops.
Bootstrapped teams face the opposite trap: they see every retained customer as validation, because they cannot afford to run the paid experiments that would tell them whether the value is real or the result of founder-led sales effort.
The shared discipline for both: separate acquisition-led metrics from product-led metrics. Track what percentage of your retained users came in through organic or word-of-mouth channels, versus paid or founder-direct channels. Product-market fit lives in the organic cohort. Everything else is a test of your distribution, not your product.
The Most Honest PMF Question
After looking at every framework and metric available, the most reliable single question I have found for diagnosing PMF is this: if you stopped all active selling and marketing tomorrow, would the product continue to grow?
Not rocket-ship growth. Just growth. New users finding it. Existing users staying. Some of them telling others.
If the answer is yes, or close to yes, PMF is real. If the answer is clearly no — if the entire top of funnel depends on founder effort or paid spend — the product has not yet found the thing that makes it worth finding.
That question cuts through a lot of metric noise. It is the one worth holding onto while you look at everything else.