Ten years ago, I listened to a talk on innovation by Vijay Govindarajan, a professor at the Tuck business school.  He was illustrating how innovation in industries tends to follow a particular pattern.  What he used to illustrate this pattern was a little unusual – the evolution of the high jump and the improvements in world record high jumps over time.

A summary of his illustration can be found on Vijay’s blog.  But the gist is the following: jumpers would incrementally improve the world record height over time by small margins.  These improvements would start to diminish as humans reached the full potential of what a given technique had to offer.  However, radically new techniques emerged over time that seemed strange initially, and might even significantly underperform existing techniques at first.  But once jumpers started to hone theses new techniques, they Were able to break through height barriers that could not be surpassed by the former conventional approach.  Thus, the high jump has evolved to the version known today as the Fosbury flop, a seemingly strange and unnatural contortion of the body that allows humans to clear a 2.4M high pole.

I think about this illustration often in a few contexts:

1.  I’m often asked in panels and elsewhere what I look for in investments that I make.  This is a pretty common question, and  most folks rattle off some combination of team, product, and market.  Those are all true.  But I am increasingly finding that the only parameter that gets me excited about a company initially is whether it is really a non-linear departure from what exists today.  Does it change the way I think about a process or product? Does it approach a market that no one has even thought about in years?  Does it enable customers to do something so absurd that it seem like magic or a hoax?  Crazy different – something that looks as weird as a fosbury flop.  Thats probably the main thing I look for these days.   

2. I’ve also been thinking a lot about the discussion of first vs. Best.   Much like the flop, the initial efforts to use a radically different approach has a slight initial advantage, but there is no stopping competitors from copying what seems to be working.  Companies like Twitter, Zynga, and Facebook will often launch products that look promising based on what some early stage company has built.  After all, capital efficiency make these efforts much easier, and these companies can leverage their installed base to start on first base.  Even startups can copy each other. So, I have to temper my fascination with the crazy different with the knowledge that execution wins.  That’s obviously why teams are so important, and why the best companies usually breakout from the pack based on amazing execution.  That’s why fast followers do have a chance (if they are leaders in disguise), some early leaders fail, and why some incumbents are so hard to displace.  

3. This tension between optimization/execution and crazy different is at play in all companies and product organizations large and small.  Here are two examples that I actually see quite a lot.  First, an early stage company might have followed the principles of a lean startup.  They’ve gotten an interesting product out, have some user traction, and are tracking tons of data.  It’s easy at this point for the product founder to get absorbed with optimizing their product.  How to improve conversions by moving this button or changing this user flow.  How can I grow from 100 new users last week to 125 by making the registration process more seamless.  These are all great and worthy efforts.   But sometimes, these changes just don’t make a real impact on the company’s potential for success.  Especially if the company is dependent on funding and still searching for a repeatable business model.  It might be that Prospective Investors are going to be way more concerned about a glaring hole in the team.  Or maybe there are other big unknowns about your business that haven’t been addressed, and reducing those risks are a better use of time than improving conversion rates by 7%.

The second related example occurs in larger companies.  At large scale, small tweaks do make huge differences, and so the product teams get obsessed with watching metrics very closely, and only launching changes that have a measurable and relatively predictable impact.  But in some cases, this leads to a sort of frankenstein’s monster of a user experience.  Pieces are bolted on one by one, each incrementally improving certain metrics in small ways.  But you really can’t optimize your way to really beautiful and elegant user experiences.  Sometimes, it’s important to paint from a blank canvas to make step function improvements.  And I think providing the opportunity to do this also helps to attract stronger talent to a company and keep them stimulated to perform.