Take twenty companies that became enormously successful. Look for what they had in common. Write it up as principles. This is the standard methodology for essentially all popular business advice, and it has a flaw that invalidates most of the conclusions.
The flaw is that you never looked at the companies that did all the same things and failed. And there were far more of them.
The mechanics of the error
Suppose that ten thousand startups pursue an aggressive growth-before-profitability strategy. Twenty become dominant. Nine thousand nine hundred and eighty fail.
Study the twenty and you'll find aggressive growth strategy as a common factor. Conclude that it causes success and you've inverted the actual relationship, which is that it's a high-variance strategy where the winners are extremely visible and the losers disappear entirely.
This is the same structure as the classic wartime analysis of returning aircraft — armouring the areas with bullet holes, when the planes that were hit in the critical places never came back to be examined.
Business writing has an unusually severe version of this problem, because failed companies produce no books, no conference talks and no case studies. The sample isn't just biased; the counter-sample doesn't exist in any accessible form.
The specific advice this undermines
"Never give up." Persistence is a common feature of successful founders. It's also a common feature of founders who spent six years on something that was never going to work. The advice provides no way to distinguish the cases, which is the only thing that would make it useful.
"Ignore the doubters." Every successful contrarian had doubters. So did every unsuccessful one. Sometimes the doubters were right and the founder should have listened.
"Move fast and break things." Worked for companies in specific markets with specific characteristics — network effects, low regulatory exposure, forgiving customers. Applied in medical devices or financial services, it produces a different kind of story.
"Culture is everything." Every profiled company has a distinctive culture, described in glowing terms after they succeeded. The same practices at a failing company get described as dysfunction. The description follows the outcome.
What would actually be informative
The useful comparison is between companies that were similar at some earlier point and then diverged. That requires identifying a cohort before you know the outcomes and following it, which is expensive, slow and produces less exciting results.
Some of this research exists. Work on startup failure rates, on what distinguishes companies that reach various stages, on the actual base rates of outcomes. It's less widely read than founder memoirs, for obvious reasons.
What it tends to show is that the variables people focus on explain less than assumed, and that market timing, capital access and founder circumstances explain more. Those are less actionable and more accurate.
The advice that survives
Some things hold up better, generally because they're derived from failure analysis rather than success analysis.
Talk to customers before building. The most common documented failure mode is building something nobody wanted. This is derived from the failures, which is why it's more reliable.
Watch cash. Running out of money is the proximate cause of most closures. Boring, universal, correct.
Co-founder conflict is a major risk. Appears repeatedly in post-mortems and is systematically underweighted in advance because nobody wants to have that conversation early.
Distribution is usually harder than product. Documented repeatedly by founders who built something good and couldn't get it in front of anyone.
Notice these are all failure-derived. Learning from failure has a much better sampling structure than learning from success, and it's a fraction of what gets published.
The narrative problem
There's a deeper issue than sampling. Success stories are constructed retrospectively, and retrospective construction imposes coherence that wasn't there.
A founder recounting how they identified an opportunity and executed a plan is describing a story assembled after the fact from a sequence that felt considerably more confused at the time. This isn't dishonesty — human memory reconstructs narratives — but it means the account describes a process nobody actually followed.
The pivots get described as strategy. The lucky introduction becomes networking. The moment the market shifted becomes foresight.
How to read business writing
A few habits that help.
Ask what the comparison group was. If there isn't one, the claim is an anecdote regardless of how well it's told.
Ask whether the advice is falsifiable. "Be persistent, but know when to quit" cannot be wrong, which means it cannot be informative.
Weight failure accounts more heavily than success accounts. They're rarer, less flattering to write, and structurally more reliable.
And be particularly sceptical of advice from people whose success came in a different era or market. Conditions that made a strategy viable — cheap capital, an open channel, an unregulated space — frequently no longer exist, and the advice outlives the conditions.
None of this means there's nothing to learn from people who built things. It means treating their accounts as data points from a biased sample rather than as instructions, which is a different posture and produces better decisions.