Monday, October 08, 2007


I just got home from attending two conferences - CommunityNext (organized by Noah Kagan and Johnny Lam) and GraphingSocialPatterns (organized by Dave McClure).

The latter is still going on - I had to get back to Los Angeles for some prior commitments, alas.

At CommunityNext, in addition to meeting some exceptionally bright developers & listening to their ideas about and experiences from developing successful Facebook applications, I listened to various presentations & observed: every successful developer/entrepreneur was quick to market with a product and quick to iterate new releases based on user feedback. In addition, they "failed" quickly - if an application didn't gain traction as hoped, they'd move on to a new idea.

I was very lucky to have the opportunity to speak with, among others, the developers of Free Gifts, Top Friends, Honesty Box, Nicknames, Audio Poke, Social Moth, one as-yet-unlaunched application, the CTO of Hi5, and representatives from UserPlane, Compete, FaceReviews, both Noah Kagan and Johnny Lam, and others. (If I didn't mention you here, gimme a shout!)

It was humbling to sit at a table with developers whose Facebook applications have user bases in the millions while mine has barely 1000. Lesson: Move! Build something else quickly, and iterate. And simplify! While there are users who think what I've built is cool and fun, it needs to be re-worked to allow for more social engagement - and it still won't appeal to a general population, but rather only to the music&sound-geek-crowd. So we're moving on to building other applications & will see what takes.

And if you have a chip on your shoulder as large as the one I have on my shoulder, being humbled sucks. So I won't be blogging much for a while.. I have a lot of work to do.

At GraphingSocial, which I was only able to attend for one day, I took the technical session given by R. Tyler Ballance of He reviewed a Facebook application he created specifically for the class (Why Are You Awesome?) , gave insight into some ways to build a better application (E.g., use fql), and presented an excellent walkthrough of how to get started. If you're interested in building facebook applications, watch the video once it's on the web. I suggest keeping an eye on the GraphingSocial website.

I also got to meet Dave McClure, who truly rocks. I've been following his blog for a while, and his "Startup Metrics for Pirates" presentation is a must-see for anyone building a web site or application, or even an intranet application since within an intranet, your "customers" are your employees.

Some technical notes from R. Tyler's presentation:
  • If you're building a facebook application, try to use fql instead of the apis. If you don't know what this means, email me or leave a comment and I'll reply - or come to the IRC channel for Facebook and ask anyone there: it's on freenode and is called #Facebook.

  • If you're working in python and using the excellent pyfacebook module, I'd suggest adding cjson to the code. The module is written to use simplejson, but cjson is much faster. What I did was to try import cjson and fall back to simplejson on an exception. If you're using django, pyfacebook is the way to go. Also, I prefer to use pyCurl to urllib2 for heavy http work.

  • Learn to love the Facebook developers wiki.

No comments:

Post a Comment