On platforms and apps

I was reading Fred Wilson's blog yesterday and saw that 4 more startups recently launched services on top of Twitter. It caught my interest, so I shot a tweet across to Fred asking why Twitter themselves weren't building more apps. (Fred's VC firm are investors in Twitter)

Fred's response was simple and I wholeheatedly agree with it: "should they be? Or should they be focused on making the platform better?" Twitter is a platform, first and foremost, and they need to keep improving its reliability, speed and APIs. However, Fred's reply also got me thinking. If I agree with his response but I still have my nagging question from the start of this post, what's up?

A bit of pondering and an always helpful discussion with Pippa later, I had my finger on it. For me, the question is not whether Twitter should focus on platform or apps, but where you should draw the boundary between the two. What does this mean?

Well, let me take the example of Google Maps. That's a platform - you only have to count the number of times you encouter in your daily forays across the web (or visit this page at Programmable Web) to see that. As a platform Google Maps could just provide draggable maps and a geo-coder (the tech that let's you turn a postal code into a map coordinate). Google could have left the local business search, driving directions, UGC mapping etc to 3rd parties. These services might be called apps. Sat on top of Google's platform.

But google didn't keep it's platform to a minimum. It built functionality into the core offering. That meant that some of these core services can work together a bit better and faster (driving duevyions to home via the supermarket). And that created value for users, for Google and for developers who could focus on writing really groundbreaking services on top of an already highly functional platform.

So where does this lead us with Twitter? Well I think just maybe they need to expand what they think of as their platform. Not too much but maybe a bit. One way to do that would be to see what little bits of code are being constantly reinventing for the service? Those are candidates for platformhood. 

A good example here is the functionality required to run Fred's lyric of the day club. That's a pretty simple idea and requires some pretty simple code. Why can't you just do that on Twitter? I'm sure there are a bunch of other little bits of code that fit this bill. And the right implementation has two benefits: first your developer community can get on with writing great apps (instead of simple little hacks) and second you can expand your 'developer' base to include highly non-technical people (imagine how many lotd-style clubs there would be if you didn't need to write a Twitter bot and own a database to build it.)

I'm sure there are a bunch of other things that Twitter should do 'out of the box'. What do you think?

PS - I'm really enjoying Twitter, so I hope this is seen as constructive feedback rather than a rant :-)