So he added more criteria, including the amount of equity an accelerator takes and the size of its alumni network. Finally he had enough data to come up with a ranking. At the top: TechStars, Y Combinator and Excelerate Labs. (Disclosure: Gilani’s firm is an investor in TechStars and Excelerate.)

//

:)

"The only two accelerators that had any meaningful exits were Y Combinator and TechStars."

Is it that there just aren't enough good companies?

"TechStars and Y Combinator typically sift through some 1,500 applications for a handful of spots each session."

To get a few decent exits you need to choose 500 startups from a pool of, say, 15,000 applications.

Wow those odds are low.

I have mixed feelings on accelerators. They seem gimmicky to me and aimed more at developing buzz amongst peers in the startup scene than enterprise campaign success.

The rest of the world has very little understanding of what goes on here and I think the valley tech scene has significantly less understanding of what goes on in the outside world than it thinks it does. This is why I believe there is so much concentration on tamagotchi social and mobile apps, and particularly, the rank attempt at a valley-holistic shotgun approach to hopefully find the next Instagram or hot game app.

They're interesting, but they don't solve any problem except boredom and the need to entertain our fickleness. This makes the entire genre weak-in-depth over the long term, since they will ALL die when they can no longer command the appropriate level of 'meme'share. In this environment, only the platforms, and other broadly based, highly 'pivotable' technologies will survive, and even then only after significant conglomeration.

And generally, I don't believe the hackathon/accelerator scene is geared towards the long game necessary to maintain success in that market, let alone the rest of the world desperate to build relationships with the tools that solve their problems.

The apparent flaw in my argument is, of course, that the costs to build new tech startups are dramatically lower these days, allowing the shotgun approach I mentioned above. After all, it's relatively cheap to build a new social site or mobile app, so why not try every permutation for the handful which stick? That's the value proposition, of course, for accelerators, the developer scene-kids, and the media buzz of the valley.

But how did it become so cheap to do a startup? The platforms. The companies which solve real world problems and figured out how to open their systems to innovation. Amazon opens the world to the material distribution channel and even went 'meta' by opening the world the cheap functional IT platform the distribution platform was built on. Apple figured out the need for clean interfaces and personalized mobile computing and opened the world to it. Google figured out how to monetize search and everyone can find info, things, and even paying customers on their platform. And Facebook seems to have figured out the power of online identity within virality, and their social graph is starting to open up the ability for us to take advantage of that.

It's the platforms which have the power. It's the platforms which have the real value. It's the platforms which survive even when the buzz has long moved on (e.g. Microsoft).

And of these accelerator graduates who have been able to maintain success, how many solve problems more useful than the Tamagotchi question? How many are platforms?

I honestly don't have a concrete answer to that question, since I'm pretty new to the scene, but coming from the outside, what I've seen seems to confirm my thesis.... (Ex: coming from Sendgrid, I know the VAST majority of Sendgrid customers are _not_ valley startups, and they put far more resources into business development with the rest of the world than in the Valley)

Could Dropbox, Heroku, and Airbnb have started without YCombinator?

Or did they need acceleration to help them survive the difficult early years?

I think there's an economies of scale argument in favor of acceleration, certainly. However, I think there exists an accelerator bubble, or at least perhaps a bubble within the accelerator scene. To what portions the bubble is caused by cheap money or by the dramatic reduction of barriers to entry remains to be seen.

uh oh. then someone has to create a strong alumni network for these accelerators. but case in point, it only shows 2 accelerators as part of the study. there should be more of those, especially around silicon valley

Well a lot of Accelerators are really just getting started.

500 Startups, AngelPad, and I/O Ventures really only began in 2010.

YCombinator started in 2005 and had a rough first few years, too.

Good things take time.

Btw I found the article here: http://www.avc.com/a_vc/2012/06/fast-followers-copy-cats-and-ripoffs.html

I wonder Fred Wilson would say about 500 Startups.