The BayCHI Web2.0 panel last night (see technorati tag baychi for full coverage) was as good as ever. Everyone on the panel emphasized that openness was very important, that remixing was key to the web. But what came through was that companies like technorati and flickr are very happy to let you leverage their APIs … AS LONG AS YOU DON’T MAKE ANY MONEY.
That’s the line in the sand. The housing maps guy was comfortable remixing google and craigslist because he has no advertising content on his site. Technorati and Flickr are happy for developers to build tools using the Technorati and Flickr APIs. But don’t try to sell the tool without talking to them first!
What this says to me is that providing open APIs is a brilliant marketing move (for data-rich companies like craigslist or technorati). But it is not as interesting from a business perspective as I had originally thought. Let’s say I build a great tool using the google API. In order to make any money from it, I have to get into a business discussion with google. At that stage I’m at a serious disadvantage in the negotiation: I’ve invested a lot, they can make that investment worthless by simply refusing to come to terms with me.
Some other insites, in random order:
You already built an API: why not document it?
AJAX / RIA applications essentially already have an API (since your code is calling the server to get data on a continual basis). Hackers can reverse engineer the API simply by sniffing their own http traffic. So you might as well publicise and document the API so you get all the free publicity.
There’s a continuum of ways to make your site easier to remix.
XHTML : easy to scrape, not a bad start
RSS : easy access to freshest content (from a category, particular author, topic, etc)
API : full access to the database (as much as the site owner wants to provide)
Remixed applications compete at the UI level:
Paul Rademacher (from housingmaps) brought this home to the audience. If you’re writing a remixed app, you don’t own the data. So it’s easy for others to copy you, you don’t have any kind of barrier protecting you from competition. This means that remixed applications succeed or fail at the UI design level (since others can remix from the same sources as you, finding a useful combination of data sources is not enough).
Hyper-growth of Blogging : like the internet circa 1996
David Sifry showed charts that show the number of blogs and blog postings doubling every 5 months. That is exactly the kind of growth that drew Jeff Bezos to quit his job on Wall Street and start Amazon. I get the impression most of these web 2.0 entrepreneurs had a moment like that, when they realized that under hyper-growth, it’s always possible to find SOME way to make money, and the best thing to do is jump in feet first: you can figure out how to make money later.
Architecture of Participation
Web 2.0 apps that succeed do so because selfish behavior makes the app better for everyone. Thinking about how to make your application have these properties is very important.
Post – Panel Writeups
PeterMe says(and I agree) that the discussion was way too technical for the BayCHI crowd. I enjoyed it, but people who don’t know what an API is would have been pretty lost.
Bill Scott has some nice notes. For people who couldn’t make it, this is a really good summary.