The Sean Ellis test (asking customers "how would you feel if you could no longer use this product?" and targeting 40%+ "very disappointed" response) is a useful qualitative signal. It's not sufficient for measuring product-market fit in a business that needs to make resource allocation decisions based on it.
The survey has known weaknesses: survey response bias, lack of predictive validity at the account level, and the measurement of stated rather than revealed preference. A customer who says "very disappointed" might still churn; a customer who says "somewhat disappointed" might be your best long-term account.
The behavioral signals that actually confirm PMF:
Organic growth without marketing investment. When customers are finding your product, signing up, and converting to paid without paid marketing driving the majority of that flow, someone is recommending you. This is the most reliable PMF signal.
Usage-driven expansion. When existing customers expand their usage — add seats, upgrade plans, adopt more features — not because your CS team sold them, but because the product was delivering more value than the initial purchase covered.
Retention without active customer success. Early-stage companies often maintain retention through heroic CS effort. Real PMF shows up when retention is strong even in accounts with limited CS engagement — the product itself is the retention driver.
Pulling behavior in sales cycles. When prospects come to you having already used the product (free trial or bottoms-up discovery) and are eager to buy, rather than needing to be convinced, this is pull behavior. Push behavior (heavy sales effort to close) indicates PMF uncertainty.
Unsolicited advocacy. Are customers telling peers about your product without being asked? This is the strongest single PMF signal because it requires positive emotion beyond satisfaction.
Survey for qualitative understanding. Measure behavior for conviction.