Funding the best company from each accelerator is a better strategy than funding all from the top program

Yesterday, I had the chance to talk to a friend who is an angel investor and an entrepreneur. Over the last 1 year he had a chance to view our work at the Microsoft Accelerator and he was keen to come and talk to me about funding companies from our program.

Over the last 3 years, we have helped over 350 companies at Microsoft Accelerator programs worldwide, which has resulted in 15 acquisitions. Over 85% of the companies got funded after our program, largely from angel investors. Looking at these numbers, he was keen to participate in the upside.

He was considering joining a syndicate which allowed investors to buy a chunk of a startup fund, which invests in all startups coming out of YCombinator. He was asking my advice on why he should consider investing in other accelerator companies at all instead of putting his weight behind YC companies.

Angel list syndicates are the new and efficient way to invest in a lot of startups using a single instrument. Think of them like a Mutual fund for privately held technology companies.

If you buy into the construct that accelerators are the new MBA programs for entrepreneurs, then it makes more sense to invest in the top graduates from every program than all the graduates from a single program.

Here is some evidence as to why based on research and in this piece it is presented by Malcom Galdwell.

The research so far suggests that the best performing students from even the worst colleges perform better than the worst students at the best colleges.

That’s even when you consider that the students at the top college performed better in standardized tests than the top performers in the not-so-good colleges.

The question is then does this hold true for tech startups as well? There is very little data to suggest that it is possible to make a completely data driven argument for the case.

Considering that companies going through accelerators made up < 10% of all Venture funded companies worldwide last year, there will be good data only in a few years.

If you look at the historical data and compare 3 programs – YC, 500 startups and TechStars, the data seems to indicate that it might make sense.

YC has funded 800+, 500 startups over 700 and TechStars over 400. The top 10 companies from each of these programs, has outperformed the remaining 95% of the startups from the other programs.

Again, the data set is limited, but it proves the theory.

The second question is how do you know which are the top startups from each cohort or program before or at demo day?

If you look at traction as a proxy or mentions in the press ( I am going to use TechCrunch and VentureBeat as examples), during the demo day, then traction “seems” to be a better indicator. In fact, if you bet against the press, you’d do very well. So there might be a contrarian investment thesis around not investing in companies that the press anoints as the “best companies from the demo day for each accelerator”.

The final question is does this hold true for all accelerator or just the top 10 accelerators?

That means should you just fund the top 10 companies from the top 10 accelerators or the top 1 company from every accelerator?

That’s also a question that’s not easy to answer directly, but you can make some educated guess via proxies.

If you look at the 1500 companies that have gone through accelerators over the last 2 years (check out SeedDB), you will notice that 21% of them have gone through multiple accelerator programs. So the best from each program (or the worst, not sure) make it to other programs as well to eventually be among the top in the top programs.

So I think the best strategy is to fund the top 1 or 2 companies from each of the top 100 accelerators, instead of all the top program or the top 10 from the top 10 accelerators.