In World War II, the US Military examined damaged aircraft and concluded that they should add armor in the most-hit areas of the plane. Abraham Wald at Columbia University proved this was the wrong conclusion, that instead, adding armor to the least hit areas of the aircraft is more effective. Wald reasoned that the military was only considering aircraft that had survived the missions; any shot-down or destroyed aircraft wasn't available to be studied.

Survivorship bias is where we only consider things that pass a selection process, i.e., survivors, and not failures. This cognitive bias can lead us to be overly optimistic. It can lead us to mistake correlation for causality by extracting common traits in successful data without considering it was also present in failures.

You can find survivorship bias in almost every business book. For example, in Good to Great: Why Some Companies Make the Leap and Others Don't, Jim Collins looked at 11 companies out of 1435 that outperformed the stock market over 40 years, looking for common traits that he believed accounted for their success. The problem with this method is that these traits could have existed in failed companies.

We can learn a lot from failure stories.