Twitter reached its hockey-stick moment several months ago, and is now growing like wildfire. No one has a clue how to monetize it yet, but being an extreme case of “build it and they will come”, the company keeps raising money and attracts more and more users. Commercial interests are quickly taking over though, albeit bottom up. People like Guy Kawasaki use automated tools to feed Twitter with a rapid stream of tweets designed to promote their business and private brand. That’s perfectly fine, as Twitter’s lack of rules indirectly encourages such behavior. It is augmented by the 140 character limit, which promotes bite-size clutter.
The question I ask myself is what new tools the Twitter community – soon to be known as “the Internet” at this growth pace – needs. There are plenty of desktop and mobile clients, innovative front ends, automated tools. But what’s missing?
Relationships are unilateral on Twitter – Anyone can follow (“become friends with”) whoever they wish. Users are identified by name and avatar, but except for a bunch of celebrities and real friends, the majority of users one follows are complete strangers. This implies the unimportance of the identity of individuals generating tweet streams. The power of Twitter comes from the combined content generated by of millions of users. Problem is – this “Über-stream” it’s too messy. This leads me to thinking that Twitter could use some sort of categorization mechanism. Hashtags has built one, but it requires voluntary tagging which only a fraction of users bother with and suffers from chaotic folksonomy wherein thousands of tags are mostly meaningless.
So how do we go about categorizing the tweet stream? That’s what I’m working on these days…