| > the need to compare crashed planes to the survivors As someone who is both a pilot and a "survivor" of three failed startups (and a coattail-rider at one successful one), this is the wrong analogy. The reason it's the wrong analogy is that most airplanes don't crash, but most startups do. This makes a big difference. The reason that it is possible to make reliable aircraft but not reliable startups is that the success of a startup changes the dynamics of the environment to make that success non-reproducible. This is not the case when designing aircraft. If you replicate a reliable aircraft design, chances are the result will be another reliable aircraft. If you replicate a successful startup, almost certainly the result will be an also-ran rather than another successful startup. Every successful startup must necessarily distinguish itself from its predecessors (i.e. from its competition) somehow. But there cannot possibly be any reliable way to produce such distinguishing factors because every one is necessarily different from all such distinguishing factors that have come before. Anything that can be easily and reliably reproduced will already be reproduced by someone and so cannot be used as the distinguishing factor that differentiates the next successful startup. This is the reason that looking to the past is of limited value. All of the stories of, "Here is what I did to succeed" are useless because none of those can possibly be a model for future success. This is not to say that you should not study the past. You should, because that can help you avoid repeating someone else's mistakes. But in the startup world, you can make zero mistakes and still fail to succeed. In aviation, making zero mistakes is pretty much the definition of success. Building a successful startup is much more akin to making a new scientific discovery than designing a good airplane. |
One thing I disagree on is the failure rate. Surely the failure rate of all businesses is skewed by restaurants and the self employed.
It seems like building a $1B+ startup is as you say, but what about building a $1M-$10M startup?