Survivorship Bias

Survivorship Bias is a cognitive error where you focus on the people or things that "survived" a selection process and overlook those that didn't — leading to false conclusions. The classic example: WWII engineers wanted to armor the parts of returning bombers that had the most bullet holes. Statistician Abraham Wald pointed out they should armor the parts with no holes — because planes hit there didn't return at all. We see this bias everywhere: studying successful companies, copying "what top performers do," and following advice from winners without considering the losers who did the same thing.

1. Ask "What am I not seeing?" — Whenever you're studying success stories or best practices, ask: "Where are the failures that tried the same approach?"
2. Seek out the failures — Actively look for examples of people, companies, or strategies that did the same thing but didn't succeed. The failure data is usually harder to find but far more informative.
3. Check your sample — Is your data set only made up of survivors? If you're studying "what successful startups have in common," you need to also study what unsuccessful startups had in common.
4. Be skeptical of success advice — "I dropped out of college and became a billionaire" ignores the millions who dropped out and didn't. The advice might be incidental to the success, not causal.
5. Look for base rates — Before following a strategy, ask: "Of everyone who tried this, what percentage succeeded?" The overall success rate matters more than cherry-picked examples.
6. Use pre-registration — When running experiments or studies, define what you're measuring before you start, so you don't unconsciously filter for favorable results.

• What you see: Spotify uses a squad model and is successful → squads must work!
• What you don't see: Dozens of companies adopted the Spotify model and failed or abandoned it
• Even Spotify's caveat: Spotify's own engineers have said the model didn't work as described in the famous blog post

• Study companies that tried the squad model and failed — what went wrong?
• Look at the base rate: of all companies adopting this model, what % saw improvement?
• Consider what actually made Spotify successful (might be talent, market timing, culture — not the org chart)

• "Successful people wake up at 5 AM" (ignoring unsuccessful 5 AM risers)
• "This company grew by focusing on culture" (ignoring companies with great culture that still failed)
• "The best products are simple" (ignoring simple products that flopped)

Survivorship Bias teaches you to look at the full picture, including the failures you can't easily see. The most valuable data often comes from studying what went wrong, not just what went right.