Survivorship Bias
Studying only the winners teaches you lessons that failed companies were following too.
Why PMs should care
Product books, podcasts, and conference talks lean heavily on successful companies — and specifically on the story those companies tell about what they did, which is already shaped to sound good in hindsight. The problem isn't that the lessons are wrong. It's that hundreds of companies that died did the same things, and you never hear from them.
'Founder-led sales' works — until it doesn't. 'Launch fast, iterate' works — until your first launch ships a bug that destroys user trust in a regulated business. 'Cut features ruthlessly' works — until you cut the one feature keeping your biggest customer on the platform.
The partial fix is to actively go looking for the failed cases that looked like the success you're trying to copy. Then ask: what would have made me mistake one for the other, in advance?
Example in product work
A team writes a retrospective titled 'Why Figma won'. It lists: multiplayer-first architecture, browser-native, generous free tier, design-system-centric, community-led growth.
Every one of these is also true of InVision, Framer, Sketch Cloud, and UXPin at various points. Framer had multiplayer. InVision had the community. Sketch Cloud was browser-native.
The actually-differentiating factor was some combination of timing, founder judgement on when to charge, and the specific quality of Figma's multiplayer — low-latency, no sync conflicts — which is a much harder lesson to transfer. The team reading the retro thinks they've learned something generalisable. They haven't.
What to do when you see it
- Success stories are told by the winners. The losers who did the same things don't get a book deal.
- 'What did the winners do?' is less useful than 'what did everyone who tried this do, and what separated the ones who made it?'
- For every widely-praised product decision, find a company that made the same decision and failed. Then ask what was different.
- Treat advice from successful people as hypotheses to test, not as rules to follow.
Sources & further reading
- A Study in the Estimation of the Vulnerability of Aircraft — Abraham WaldWald's wartime memoranda on armouring bombers — the canonical real-world illustration of survivorship bias.
- Survivorship Bias — The Decision LabA practical explainer with examples from finance, business, and history.