Metric
PMF survey score ("very disappointed" share)
Share of recent core users who would be very disappointed if they could no longer use the product, read per segment. Directional from about 40 responses. The 40% line is a heuristic, not a statistical cut-off.
Plain label: must-have shareAlso known as: Sean Ellis score, PMF score, 40% test
How to run it. Survey only users who used the product at least twice in the last two weeks; a one-off emailed survey works at any analytics level. Ask the disappointment question plus who benefits most, the main benefit and what to improve, then segment the answers.
Read it. Report the share and the n per segment; with fewer than about 40 responses it is a direction, not a reading. Superhuman went from 22% to 58% (2018) over about three quarters; a survey of 731 Slack users found 51% (2015) (benchmark confidence B). Neither is a target for your product.
Triangulate. A strong score with a curve that still declines to zero is a conflict to resolve with a question, never an average.
Benchmarks
Every benchmark carries its year and a confidence grade for its source. Use them to calibrate, not as targets.
| Context | Median | Top 25% | Top 10% | Source |
|---|---|---|---|---|
| 'very disappointed' share of recent core users (2+ uses in 2 weeks); 40% is a heuristic line | 40% | — | — | Sean Ellis, via Superhuman (First Round Review), 2018Benchmark confidence B. A reputable dataset. |
Limits
What we don’t claim
- The threshold is a heuristic, not a measured cut-off.
Sources
How Superhuman built an engine to find product-market fit · First Round Review (Rahul Vohra), 2018
Benchmark confidence B. A reputable dataset.
Book content is paraphrased, never quoted at length. Last verified October 24, 2026.
Your product
Find the stage your product is stuck in
Gamify It reads your product, picks the one move that matters at its stage, and sizes the test to your traffic. Connecting is free; a Pro subscription starts it.