The skill: Knowing whether you actually have product-market fit instead of guessing. The specific measurement toolkit: Sean Ellis's "very disappointed" survey, retention curve analysis, organic pull rate, and what to do at each stage.
Benchmarks
Run the Sean Ellis survey
Ask users: "How would you feel if you could no longer use this product?" If 40%+ answer "very disappointed," you have PMF. You need 40+ responses minimum for meaningful signal — anything less and a handful of enthusiastic early users can skew the number.
Before you run it on your own product, calibrate. Hiten Shah surveyed 731 Slack users and found 51% would be "very disappointed" — that's what strong PMF looks like in practice. He then segmented responses into must-have users (very disappointed), users who see value but aren't there yet (somewhat disappointed), and users who don't find it useful (not disappointed). Studying the gap between those groups revealed exactly what Slack needed to convert the next wave. You can do this with any product in your space to understand what "good" looks like before measuring yourself.
Segment before you celebrate (or panic)
Your PMF score for "everyone" is meaningless. Rahul Vohra's team at Superhuman scored 22% overall. Terrible. But when they filtered to their target persona, 33%. Still low, but a foundation to build on.
Here's what they did next: (1) Identified who the "very disappointed" users were and studied what they loved — then doubled down on it. (2) Studied what held "somewhat disappointed" users back — then fixed those barriers. (3) Split the roadmap 50/50 between deepening love and removing barriers. The result: 22% → 58% in three quarters. No pivot. No new product. Just relentless focus on the right segment with the right split between strengthening what works and removing what doesn't.
The lesson: find the segment that loves you. Measure them separately. Build for them first.
Check your retention curves and acquisition mix
The Sean Ellis survey tells you how users feel. Two other signals tell you what they actually do.
Retention curves: plot by monthly cohort. A curve that flattens = users who stay, stay. That's PMF. A curve that keeps declining and never flattens = no PMF, no matter what your survey says. The survey captures sentiment. Retention captures behavior. When they disagree, trust behavior.
Organic acquisition share: if more than 50% of new users come from word-of-mouth or organic channels, that's a strong PMF signal. When people voluntarily tell other people about your product, something is working. Paid-only growth can mask the absence of PMF entirely — you're buying attention, not earning it.
When PMF feels off, isolate the broken piece
PMF isn't one thing — it's five hypotheses working together. The problem you're solving, your target audience, your value proposition, your competitive advantage, and your growth strategy. When PMF feels off, most teams tweak the solution. But Sachin Rekhi's insight is to ask which hypothesis is actually wrong. Maybe the problem is real but your audience is wrong. Maybe the audience is right but your value proposition doesn't beat the alternative. Maybe everything is right but your growth strategy can't reach the people who need you.
Isolating the broken hypothesis saves you from pivoting everything when only one piece needs fixing. Before you overhaul the product, check whether the problem is upstream.
PMF isn't a one-time event. Product changes, market shifts, and new competitors can erode it. Re-measure after every major change — quarterly at minimum.
Do's and Don'ts
Written with ❤️ by a human (still)