As with the previous cycle, we are once again seeing the formation of many new incubators (Om has a post up which inspired my post).   Having co-founded an incubator and having observed a bunch of others over the years, I have a theory about what works and what doesn’t.  I believe that the only two versions that work are: accelerators, such as YCombinator and Techstars, and “foundries” (my attempt at a neologism), such as Idealab and Alleycorp.

Let’s start with accelerators.  These are characterized by having fairly short, rigid programs with a well defined beginning and end, small and fixed investment amounts and ideally a highly selective admission process.   The best of these programs deliver on two premises: (1) take a process that might otherwise take a year and accelerate it down to a few months and (2) dramatically increase the likelihood of obtaining follow-on funding.  The rigidity of the programs is critical to their success.  Without it, they could not process as many applicants and they would rapidly get mired in trying to adapt to the individual circumstances of each company.

“Foundries” are almost the exact opposite. They are vehicles that let highly successful founders start multiple companies in parallel or rapid sequence.  These companies are the founders’ ideas and the founders start out with 100% of the equity.  They usually put in a relatively small amount of money themselves, quickly recruit a high class management team and build enough to be able to attract top-tier outside investors.  What is critical to the success of foundries is the strength and connectedness of the founders.  Without that it becomes impossible to attract strong teams and outside capital early in the process.  Very few people have enough of a track record (and are great at picking ideas) to pull this off.

So what is ruled out by this: any and all programs that fall in-between these two extremes (I learned this the hard way!).  If you try to customize your offering for each startup and work with outside teams and their ideas, you wind up with both moral hazard and adverse selection problems of huge magnitudes.  These incubators attract teams that are not strong enough to make it on their own (adverse selection) and expend too many resources for too long on concepts that are not getting traction (moral hazard).   On top of that you don’t even have the control to make the changes that would be required to fix things.

I believe that the ecosystem at any point can sustain only a handful of accelerators and foundries.  The fall off curve beyond those will be dramatic as by spreading out resources you diminish the likelihood of success (as we all know, starting just one successful business is super hard).