Oct 19, 2009

I Want To Be A Platform

We all want to create or invest in platform businesses - those that enable other businesses or participants to add features, content, build other things into/onto the platform, extract associated but not directly related content therefrom, thereby creating infinite scale and value. And valuation.

But it doesn't work so easily. I submit that the best platform businesses evolve, they are not created sui generis. Further, they evolve from applications - applications as representative of specific pain points and use cases.

Applications can and do become platforms. It's extremely hard for platforms to begin as platforms and at the same time find the specific problem they are solving, in a way that encourages usage and enhances the users' experiences and leads to growth.

So another core betaworks philosophy: it is infinitely harder for a platform to spawn value added applications than it is for an application to evolve into being a platform.

We've seen this many many times before, for example:

AOL - began as chat/email application, then evolved into "portal" platform.

Yahoo - initial pain being solved was navigation - application was a directory, also later evolved into portal platform.

Twitter - application= messaging, now becoming a communications and media (news, links, more) platform.

Google - search application into media platform.

More specifics: we built switchabit as a content routing platform, a way to move any piece of content from its base format (mp3. jpg, blog post) and location to another service. Nice, decent usage, but too many and varied use cases (flickr to twitter, tumblr to blogger etc etc) which I believe got in the way of mass adoption. It did, however utilize a URL shortening service (bitly), which solved a very focused pain. Thus that application was born separate from the platform. Similarly, Twitterfeed does one thing - as a content routing application - moving one piece of content to one or two other places. Result: 350k publishers and about 600k feeds running through the service. Maybe it will now move to platform like attributes. Maybe not but it has proven it can solve a specific pain and scale.

In practice, then, for us when we build a service, or invest, we look closely at the use case to determine what kind of application it might become. We don't expect any evolution, though as dreamers we hope that there will be one. Show us an application, not a platform.

blog comments powered by Disqus