Also, I think the section regarding TechStars success makes an assumption mistake:
"The national rate of failure for startups is 75-90%. So lets say it was a good year and Techstars Chicago received a 1000 applicants/startups to choose from. You know based on national averages that 750-900 of these companies will fail. That leaves 100-250 that will succeed. Techstars is only taking 1% of applicants, which is 10 startups. Unless Techstars is picking poorly, they are selecting companies that are positioned for success from day one. These startups would most likely succeed with or without Techstars."
The common assumption that 75%-90% of startups fail is usually regarding venture backed startups, not people who fill out the TechStars applications. I would wager that the if you looked at the failure rate of anyone who had an idea they wanted run through an accelerator, the failure rate would be much, much higher. What you should benchmark TechStars against VC performance.
(Disclaimer: I went through TechStars.)
Also important to control for cohort size. These are 10 company classes. A 100 million+ breakout seems to be somewhere between a 1 in 100 and a 1 in 50 occurrence.
They've crowd-funded over 125 million dollars toward medical expenses for people in need.
http://betaboston.com/news/2014/09/16/product-design-startup...
And the real answer to this is that getting large exits takes a long time. Unless I'm mistaken YC, which is older than Techstars and has 50% more portfolio companies, has 3 exits over 100 million, OMGPOP, Heroku and Twitch. Granted the last one was a big one.
(Disclosure: Hackstar in 2009, Founder in 2010)
The exit analysis is the first thing I think of when I think of Techstars, too. Ostensibly, the reason to sign on with an "incubator" is that the connection you get would enable you to obtain rapid growth. But I can't think of good examples of companies where that's happened.
The equity analysis you pointed out as well as the description of the LoI stuff made me reconsider that position, and by the end of the article I was glad I read it. It wasn't quite the muckraking I was expecting, and instead was a decent exploration of when a fit with an accelerator just didn't exit.
Matasano never did the incubator/accelerator thing, right?