What is product-market fit?

The two canonical definitions

Sean Ellis (2009) — quantitative survey method: Run a PMF survey on users who have used your product at least twice. Single question: "How would you feel if you could no longer use [product]?" Options: Very disappointed / Somewhat disappointed / Not disappointed / N/A I no longer use it. 40% or more "Very disappointed" = product-market fit.

Andy Rachleff (Benchmark VC) — qualitative framework: "Product-market fit means being in a good market with a product that can satisfy that market." Two components: (1) the market is large enough + growing enough, (2) your product addresses a real need of that market.

The two definitions are complementary. Ellis is the measurable signal; Rachleff is the underlying causal frame.

The 6 observable PMF signals (when you have it):

1. Organic growth — users sign up without paid acquisition; word-of-mouth dominant
2. Low churn — monthly churn <5% (B2B) or <8% (consumer); users stay
3. Net Revenue Retention >100% — existing customers spend more over time
4. Customers pull product — they ask for it, recommend it, build workflows around it
5. Support load shifts from "help me get value" to "help me do more" — onboarding gap closes
6. You can't keep up with demand — features requested faster than you ship; positive problem

The opposite — pre-PMF — has the inverse: paid traffic dominant + high churn + pushed sales + support load on basic activation problems.

The Marc Andreessen canonical statement (2007):

"The only thing that matters is getting to product-market fit. You can always feel when product-market fit isn't happening. The customers aren't quite getting value out of the product, word of mouth isn't spreading, usage isn't growing that fast... And you can always feel product-market fit when it's happening. The customers are buying the product just as fast as you can make it — or usage is growing just as fast as you can add more servers. Money from customers is piling up in your company checking account."

This is the qualitative gut-check version. Combined with Ellis 40% measurement, you have both signal and feel.

The 40% threshold — where it came from:

Sean Ellis ran the survey across 200+ early-stage startups (Dropbox, LogMeIn, Eventbrite, others) and found: - Startups with ≥40% "very disappointed" had organic growth + low churn - Startups with <40% had to push hard via paid acquisition + fought churn - Threshold validated empirically across hundreds of subsequent companies

The 40% number isn't arbitrary — it's the inflection point where word-of-mouth becomes self-sustaining.

Pre-PMF reality (per YC + Ellis data):

Most early-stage startups are in the 10-30% range. The work is finding the 40%-segment, then ONLY targeting them.

Why "narrow to the segment" matters:

Aggregate 25% can hide: - Power-users segment at 60%+ → narrow to them, scale = PMF found - Casual users at 10% → ignore them; they're not your market

Run the Ellis survey + segment by: usage frequency, role, company size, use case. The 40%-segment is often surprisingly narrow at first.

Time to PMF (per YC + Pitchbook data):

The "60% of startups die before PMF" statistic comes from this data. Most quit before reaching 40% threshold.

The 4 PMF failure modes (per Ellis + Lessin + Rachleff convergence):

1. Wrong market — solve a real problem but not enough people have it. No amount of polish fixes this. PIVOT.
2. Right market, wrong product — people want SOMETHING in this space, but not exactly what you built. ITERATE within market.
3. Right product, wrong positioning — people would love it if they understood it. FIX MESSAGING.
4. Right product, right positioning, wrong segment — power users love it, but you're marketing to wrong group. RE-TARGET.

The "you'll know" myth (real but incomplete)

Andreessen says "you can always feel it." This is true but unhelpful pre-PMF — many founders feel they have PMF when they don't (confirmation bias on 5 happy customers). The Ellis survey provides the discipline: 40% threshold on actual user data.

Common PMF misconceptions:

• PMF is a one-time achievement — wrong; PMF erodes as markets shift, competitors enter, customer needs evolve. Constant work.
• More features = better PMF — wrong; PMF is about hitting the core need exactly. More features often dilute.
• Revenue = PMF — wrong; you can grow revenue via paid acquisition without organic PMF (and burn out when paid stops working)
• PMF means everyone loves you — wrong; PMF means 40%+ of your TARGET segment would be very disappointed. Most of the world is irrelevant.
• PMF threshold is fixed at 40% — Ellis's data; some practitioners use 50% in 2024-2025 (Lessin) due to higher competitive intensity.

The "fake PMF" trap

Common pattern: founder funds aggressive paid acquisition. Revenue grows. Looks like PMF. Then paid budget runs low or CAC rises → growth stops → reveals high churn → revealed-not-PMF. Painful pattern.

Antidote: Ellis survey + churn measurement + organic growth tracking. Honest data over revenue-vanity metrics.

This is NOT investment advice:

PMF is a framework for product strategy. It does not predict business outcomes — many products reach PMF and fail commercially (wrong pricing, wrong margins, hostile competitor moves). PMF is necessary but not sufficient.