Survivorship bias
Drawing conclusions from the cases that made it, while the ones that did the same thing and failed are invisible.
Survivorship bias is reasoning from a sample that only contains winners. The businesses that folded do not publish retrospectives, the creators who posted daily for three years to no audience stopped posting, and the failed version of every strategy is systematically missing from the evidence you can see.
The effect is that visible advice describes what successful people did, not what caused their success. If a hundred businesses posted daily and three grew, you will read three posts about the power of consistency and none from the ninety-seven. The advice is not dishonest; it is just a survey of survivors presented as a method.
The correction is not cynicism. It is asking one additional question of any strategy: how many people did this and it did not work, and can I find one of them? If you cannot find a single failed case, you are not looking at strong evidence, you are looking at a filtered sample. And when the failures are findable, the difference between them and the survivors is usually the actual lesson.
Worked: five case studies say a launch sequence produced $50,000. The vendor ran it with 400 customers. Forty at that scale is the number missing from the page, and the median result — not the five you were shown — is what your own plan should be built on.
Also known as
- survivor bias
- selection bias
Relevant for
- Founders
- Before copying a strategy, go and find someone it failed for; if you cannot find one, you are reading a filtered sample rather than evidence.
- Creators
- Every "post daily and it will compound" story is told by someone it worked for — the ninety-seven who did the same thing simply stopped posting and never wrote it up.