You might have heard that the rate of failure for startups is high. If you google "startup failure rate", you will be immediately confronted with the statistic that 90% of startups fail. There are lots of sites that parrot this number.

Jason Feifer, who is the editor in chief of Entrepreneur magazine, disputes this claim in his 2022 book Building for Tomorrow:

Don't believe any of that. The statistic is wrong - by the numbers and more important by the definition of fail...

First, the facts: About half of businesses survive their first four years according to the U.S. Census Bureau. Does that mean the other half failed? No. At the time that they closed, about a third of those businesses were successful. "It appears that many owners may have executed a planned exit strategy, closed a business without excess debt, sold a viable business, or retired from the work force," the study says. In other words: Just because a business ends, that doesn't mean it ends badly....

Even if the stat were true [i.e., Nine out of ten new businesses fail ], we would need to define the word fail. If a business truly does fail - if it ran out of money, laid off its staff, and totally crashed and burned - is that a failure? Maybe .. but maybe not! Great businesses have been built out of the lessons from failed ones. YouTube began as a failed dating site; Twitter began as a failed podcast platform called Odeo; Instagram began as a failed app to help plan meetups called Burbn.

This is not just true for companies. The most successful people in the world have told me that they attribute their successes to their failures. pp. 134-137

If you research some of the sites that report that 9 out of 10 startups fail you will notice that they often use a restrictive definition of what constitutes a startup. They often restrict the word startup to refer only to companies doing innovative work in some high tech space that are funded by private capital. If the people reporting the statistic want to restrict the use of the word startup to these types of businesses then they are arguably misleading the general public when claiming 9 out of 10 startups fail. I suspect most people believe that those who open salons, restaurants, retail stores and other types of businesses are startups also and assume the failure rate applies to them also.

Whenever anybody quotes a statistic on how common startup failure is, they should make an effort to be clear about what they mean by a startup and what source they are relying upon to make their failure rate claim. Depending on the definition of what a startup is, who the source of the statistic is, over what time horizon, and how they define what failure consists of, you could be looking at a huge range of possible statistics for what the "startup failure" rate is. Talking in general terms about the rate of "startup failure" is not likely to be that useful.

Is starting a business a risky proposition? Yes, but probably not as risky as most people have been lead to be believe and the reasons for closing down a business may not be as dire as we imagine.