Guy’s post came on a day when I was engaged in a conversation with two separate startups about the truths and myths of startup strategy.

This is what I call the strategy paradox. That is, the same strategies that have the highest probability of extreme success also have the highest probability of extreme failure. In other words, everything we know about the linkage between strategy and success is true, but dangerously incomplete. Vision, commitment, focus…these are all in fact the defining elements of successful strategies, but they are also systematically connected with some of the greatest strategic disasters.

So, if what Guy says it true, why are there so many “proven strategies” for startup success? The short answer is that there aren’t specific proven strategies for startup success that go beyond the lessons found in first year MBA classes. There are, however, many myths about such strategies perpetuated by three forces:

• We tend to analyze only the success stories and some of the very prominent disasters as opposed to the entire (rather large) population of startups, the great majority of which fail. So the sample used to draw patterns is very skewed. In many cases, there is no data to suggest whether a certain pattern observed in successful startups is equally or more prevalent in failed ones. Many successful Bubble 1.0 startups bought Super Bowl ads. So did even more of the ones that failed.
• We tend to confuse correlation and causation. Buying a Super Bowl ad may correlate strongly with success for some type of startups but does it actually cause the success? If Super Bowl ad => Success, then Failure => no Super Bowl ad. That’s basic logic. Since there are plenty of failed startups that did spend a cool million and failed, it’s unlikely that Super Bowl ads cause startups to succeed.
• We love revisionist history because we are vain creatures with big egos. Success is due to our smarts and guts (and our competitors’ stupidity). Failure is due to random external events beyond our control. Since we can’t derive patterns from random external events, we derive them from the legends we create to commemorate our successes.

Guy’s point is spot on: instead of there being proven strategies for startup success, there are some rules of thumb for how a startup can assume insane amounts of risk to generate significant returns or fail spectacularly. Luck and timing play a huge role in determining the ultimate outcome. As a VC, one of my biggest turn-offs is a startup team which isn’t willing to give appropriate credit to randomness.

For those skeptical about the argument, consider a different ecosystem–money managers. Sometimes the ones that do best in a given market environment are the ones that take insane amounts of risk, e.g., making big currency or interest rate bets, going very long w/o cover, etc. As long as there is no major change in the environment they do well. The Long-Term Capital Management story–losing 4.6B in four months after the Russian government defaulted on their bonds–should be a lesson for everyone operating in complex environments with a lot of uncertainty. One interpretation is that what brought LTCM down is a completely unforseeable event. Another interpretation is that the principals, in their search for returns, took unbounded risks whose magnitude they weren’t aware of and eventually got burned. It’s OK to take these types of risks as long as both the principals and the investors (entrepreneurs and VCs in the startup case) are well-aware of the risk profile as opposed to fooling themselves to believe that certain events just won’t happen.