Why is most startup advice affected by survivorship bias?
Because almost all of it comes from companies that succeeded, and the failures that did exactly the same things are not available to be interviewed.
The classic illustration. During the Second World War, analysts examined returning aircraft and proposed reinforcing the areas with most bullet holes. Abraham Wald pointed out the error: those were the planes that came back. The undamaged areas on survivors were where the lost aircraft had been hit. The critical data was missing by definition.
Startup advice has exactly this shape. The founders available to write books and give talks are the ones whose companies worked.
How it distorts specific advice:
"Never give up." Every successful founder persisted through a bad period. So did an enormous number of founders who lost years and their savings persisting with something that was not going to work. Persistence is visible in success stories and invisible in failures, so it appears more reliably useful than it is.
"Drop out of university." Drawn from a handful of extraordinary cases, against a very large number of people for whom it went badly and who were never asked.
"Move fast and break things", and every other maxim that worked for a company with an unusually forgiving market position.
"Trust your vision over research." Survivors who ignored evidence and were right are memorable; those who ignored evidence and were wrong are not.
The deeper problem: luck is systematically underweighted. Successful founders reconstruct their history as a series of good decisions, because humans build coherent narratives. Timing, market conditions, a chance introduction and macroeconomic luck are real and enormous, and they do not make good stories.
What to do about it:
Seek out failures. Post-mortems from founders who closed companies are far more informative than success stories, and there is a genuine body of them.
Ask about the base rate. Not "did this work for them" but "of everyone who did this, what proportion succeeded?"
Prefer advice with a mechanism, not just an outcome.
Weight process over result. A good decision can produce a bad outcome, and copying the outcome teaches nothing.
General information, not business advice.