Three Product-Market Fit Signals You Should Actually Track

"You'll know it when you feel it" is the most commonly repeated and least useful piece of advice in startup building. Product-market fit signals are real, they're measurable, and there's no good reason to rely on intuition when you can track the actual indicators.

This post covers the three signals that matter most, how to measure them, what the Sean Ellis test adds to the picture, and the common mistake that leads teams to misread where they actually stand.

Why Most PMF Assessments Are Wrong

The standard PMF assessment looks at growth rate and NPS. Both are misleading in isolation.

Growth rate can be inflated by outbound sales, aggressive discounting, or a single large customer. It's a lagging indicator that tells you what happened, not whether customers are finding genuine value. A company growing fast through discounting is not at product-market fit.

NPS measures satisfaction. Satisfaction is not the same as fit. A customer can be happy with your product and still not be retained, referring, or paying full price. NPS above a certain threshold feels like a green light when it's really just the floor for acceptable customer experience.

Product-market fit signals need to be behavioral, not attitudinal. What are customers doing, not just saying?

Signal One: Pay

Customers at true product-market fit pay without heavy discounting and without extended negotiation.

Watch your deal velocity and discount rate together. If every deal requires a discount above a certain percentage to close, that's a signal that the perceived value of the product doesn't justify the price at face value. Buyers are factoring in uncertainty about whether the product will deliver.

Watch your trial-to-paid conversion rate and the time it takes. At product-market fit, users who try the product see the value quickly and convert without extensive sales nurturing. The product is doing the selling. When the conversion rate requires heavy sales intervention, the product isn't making the case on its own.

Pay is the floor. If customers aren't paying without friction, the other two signals don't matter yet.

Signal Two: Stay

Customers at product-market fit renew without constant success management intervention.

The signal to watch is churn rate relative to customer success effort. If retention requires a dedicated CSM for every account and retention drops when coverage decreases, that's the product failing to deliver standalone value. Customers are staying because of the relationship, not the product. That's a services business, not a SaaS business.

Healthy retention at product-market fit looks like customers renewing before their contract ends, expanding usage without being asked, and doing so across accounts with minimal CS involvement. The product is embedded in a workflow that the customer can't easily abandon.

Look at your retention curve by cohort. A curve that flattens over time means you have a retained core — users who found the product indispensable. A curve that never flattens means something is wrong with the core product loop, and no amount of customer success can fix it sustainably.

Signal Three: Refer

Unprompted referrals are the ceiling signal for product-market fit.

This is the hardest signal to manufacture and the most reliable one to trust. When customers recommend your product to peers without being prompted by a formal referral program, they're staking their professional credibility on the recommendation. People only do that for products they genuinely believe in.

Track your new logo sources over time. What percentage of new customers were referred by an existing customer without any formal incentive? When that number starts climbing consistently, something has shifted in how customers experience the product.

Referral is harder to measure than pay or stay, but it's worth instrumenting. Add a "how did you hear about us" question to your sales process and log it consistently. When referral becomes a top-three source of new pipeline, you're at or near product-market fit.

The Sean Ellis Test

The Sean Ellis test adds a leading indicator on top of the three behavioral signals. Survey your active users with a single question: "How would you feel if you could no longer use this product?" with four response options: very disappointed, somewhat disappointed, not disappointed, and not applicable.

When a meaningful portion of respondents answer "very disappointed," users have built enough of a dependency on the product that its absence would be felt. This is an attitudinal measure, not behavioral, but it precedes behavioral retention and referral in the customer journey. You can use it as an early warning system before churn data tells you something is wrong.

A useful tool for running the Sean Ellis survey is Delighted or a simple Typeform. Send it to users who have been active for at least two weeks so you're measuring users who've had enough experience to form a real opinion. Run it quarterly to track the trend over time.

Common Mistake: Treating NPS as a PMF Signal

The most common PMF measurement mistake is relying on NPS as the primary indicator of product-market fit.

NPS measures willingness to recommend. It does not measure whether customers are actually recommending, whether they would be disappointed if the product went away, whether they're renewing without discounts, or whether they're staying without hand-holding. A company can have a strong NPS and none of the three behavioral signals.

This matters because NPS is easy to game. White-glove onboarding, aggressive customer success, and responsive support all improve NPS without improving the product. You can build a high-NPS, low-PMF company — and many teams do, without realizing the distinction.

The second version of this mistake: checking one signal and calling it done. Pay is strong, so the team declares PMF and scales. But retention is dependent on CS coverage, and referral rate is near zero. The product hasn't actually found fit — it's found a customer base that is satisfied enough to stay but not compelled enough to stay or share. Scaling into that situation accelerates the problem.

Start Here

Map all three signals for your current customer base. What's your close-without-discount rate? What percentage of renewals happen with minimal CS intervention? What percentage of new logos came from unprompted referrals in the last 90 days?

If one of the three is significantly weaker than the others, that's your PMF gap. It points directly at where the product or the customer experience needs work before you scale.