MashupCamp (2 of 2)

So the actual MashupCamp (which was Wed and Thurs, July 12 and 13) was a great experience. As the Merc article talks about, the agenda for the unconference was determined Wednesday morning. I'm going to give my thoughts on some of the issues discussed, but a thorough summary is available if you want one. Visit the wikis on the individual topics for indepth info.

The Computer History Museum (where the event was held) was filled with true Silicon Valley nerds, intriguing new companies, giant old ones, VCs, and a few movers and shakers. My impressions of the individual discussions were mixed. Interestingly enough, my favorite one, Jeff Clavier's talk on start up financing, had very little to do with the conference's focus on mashups. Jeff is a great guy and offered some fantastic, straight-shooting advice. He is also one of our (Mashery's) investors, so it was cool to hear him give examples involving our funding process. As a side note, I also got to meet Josh Kopelman of Half.com fame, who I have pointed at on this blog a few times. Again, fantastic guy by all accounts.

Another discussion hit at a fundamental point in the mashup world, and one very pertinent to Mashery: API pricing models. The discussion itself wasn't extremely enlightening and at times was reduced to the affirming and denying of mashup developer's willingness to pay. However, the motivations underlying the discourse were very telling. Already, two seperate camps are forming in the space - the people who want to make developers pay for APIs (and make money off the transactions) and those who want to keep them free. So far, Amazon is the only example of a successful monetized system. It was generally agreed that most models would be based on volume (pay per API call) rather than a licensing system.

The basic problem is actually quite interesting. You have API developers, who do tons of work and then publish it, making their code and data available to the public. Then you have mashup developers who use multiple APIs to throw together cool stuff. This is where it gets tricky - what if this cool stuff makes the mashup developer money? Well ok, you can say, if the mashup makes money, the API developers, who did most of the work, should get a cut, right? But what if the mash up is only making money on ads, i.e. it's just a really popular site that happens to use multiple APIs? A good overall solution would be to charge based on volume while giving the first few API calls for free. But then again, charging in general takes away from the open source feel of mash ups by economically disincentivizing creativity in making new mashups. So it is a sticky problem indeed.

A few more points about the unconference as a whole. It was interesting that one widespread problem was the lack of reliability in the space. API providers reserve the right (in their terms of service) to change their APIs whenever the way. So building a serious mashup (one that aims to make money or perform a hardcore task) is risky because tomorrow Google could just change part of its map API and you'd be left out in the cold.

Lastly, there was an overwhelming cry for better support in mashup creation. Developers seem to really want what Mashery is offering: support via blogs, wikis, forums, documentation, etc. There's a real lack of it out there in the mashup space right now. So from that standpoint, and seeing that almost everyone had heard of Mashery by the time the event wrapped up Thursday afternoon, I'd say it was quite a success. Look forward to the launch of our site (Mashery.com) in a few weeks, and I may blog once more on the unconference over the weekend.