Previous  |  Next

Tue

Nov 27
2007

Jimmy Guterman

Jimmy Guterman

New Edition of Facebook Application Platform Report

In October, in response to the great interest in Facebook opening up its platform to developers, we released The Facebook Application Platform, a report by Tim O'Reilly, Niall Kennedy, Dave McClure, and the Radar team that offered plenty of data on what was working for Facebook developers, both quantitatively (we charted the Top 200 Facebook applications, broke them down by category and much more) and qualitatively (we offered a history of widgets, surveyed the best practices of Facebook app marketing, and profiled two of the more interesting companies in the space, iLike and Slide).

When we published the report, we promised two updates. The first of those two is now done. (If you purchased the original report, you'll receive information shortly on how to get the update. Let me know if you don't.) The new version updates all the charts and graphs in the original report and follows the astonishing increase in the number of Facebook applications -- up nearly 50% since the original report.

What's the big news in the update? We track the continuing concentration of successful Facebook applications. Rather than a long tail, Facebook may exhibit an "empty tail." As the update notes:

"In our initial report, we found that the top one percent of applications generate 74 percent of active usage. The top has consolidated even more since then. Even among the top developers, you can see a steep drop in just the top 10, from more than six million active users to less than one million. As in other parts of the entertainment industry, a Facebook developer needs to build a huge viral hit to get anywhere near the top."

We're working on the third and final version now, and we will issue it in January. It will include a great deal of new data and interviews and it will be sent automatically to anyone who has purchased either the first or second edition. You can order the report here.



tags:   | comments: 3   | Sphere It
submit:

 

0 TrackBacks

TrackBack URL for this entry: http://orm3.managed.sonic.net/mt/mt-tb.cgi/7307

Comments: 3

CW [11.27.07 09:41 AM]

Broken link at the end there, bub.

Jimmy Guterman [11.27.07 10:07 AM]

Fixed, thanks!

Michael Sparks [11.27.07 05:04 PM]

Just a silly question - you are using a log-log scale for graphing your data? If you are and you're NOT getting a straight line that would be *extremely* interesting.

After all zipf distributions (aka power law & long tail) tend to turn up anywhere (and everywhere) people have choice.

If you're getting something that looks like a long tail style graph but on a log-log scale (as it sounds like you might be), two things spring to mind 1) the social graph aspect or user interface for choosing aspect or simply the meme aspect is severely distorting behaviour in a (as far as I know, but I'm not a statistician) fairly unprecedented way. 2) It may be worth re-graphing the data on a log(log)-log(log) graph or similar combinations to see if you can regain a linear relationship.

Post A Comment:

 (please be patient, comments may take awhile to post)






Type the characters you see in the picture above.

Subscribe to this Site

Radar RSS feed

RELEASE 2.0 BACK ISSUES

BUSINESS INTELLIGENCE

CURRENT CONFERENCES