On Wednesday, Twitter tipped the tuna. By that I mean it started peaking. Adoption amongst the people I know seemed to double immediately, an apparent tipping point. It hasn’t jumped the shark, and probably won’t until Steven Colbert covers this messaging of the mundane. As Twitter turns 1 on March 13th, not only is there a quickening of users, but messages per user.
Twitter, in a nutshell, is mobile social software that lets you broadcast and receive short messages with your social network. You can use it with SMS (sending a message to 40404), on the web or IM. A darn easy API has enabled other clients such as Twitterific for the Mac. Twitter is Continuous Partial Presence, mostly made up of mundane messages in answer to the question, “what are you doing?” A never-ending steam of presence messages prompts you to update your own. Messages are more ephemeral than IM presence — and posting is of a lower threshold, both because of ease and accessibility, and the informality of the medium.
Anil Dash was spot-on to highlight “The sign of success in social software is when your community does something you didn’t expect.” A couple of weeks ago it became a convention to start messages with @username as a way of saying something to someone visible to everyone. Within the limited affordances of the tool, people started to use it not only for presence, but a kind of shouting at the party conversation. Further, when you see an to someone who isn’t in your social network, you find yourself inclined to go see who it is or add them if they are a friend who just joined. This kind of social discovery goes beyond seeing friend lists on profiles, aids network structure and quickens adoption.
While the app is viral (you have to get others to adopt to be able to use it), mobile social software has great word-of-mouth properties. At Wikimania this summer, a buzz went off in my pocket when I was having dinner, which prompted me to get Jason Calacanis, Dave Winer and the brothers Gillmor to adopt. Wednesday was the first day of TED, so a bunch of A-listers spread it. At SXSW it seems to be the smart mob tool of choice, and there is even a group for it with a feature I’ve never seen before, JOIN.
Most recently there has been a rise in fake identities and even celebrities. Partially because people want to form more than one group, sometimes as integration points with other communities. Some of the groups I’ve spotted include AdaptivePath, Barcamp, Technorati (a hack that begs people for blurbs in WTF), Techmeme (a hack that posts new top stories) and Wordpress (release updates). Andy Carvin hypothesizes Twitter could save lives in a catastrophe, but group forming is already ahead of his theory with the USGS Earthquake Center on Twittter.
This week most of my company joined Twitter and I set up http://twitter.com/socialtext for no reason in particular. I posted the login in a private wiki page to let anyone contribute. But when Moconner saw how simple the API was, he wrote a bot to let us post from our IRC channel. Now we have a low threshold way to express group identity that fits with the way we work.
Liz Lawley well addressed the differences of this form of presence and criticisms of mundane content and interruption costs. She highlights “exploring clusters of loosely related people by looking at the updates from their friends. There are stories told in between updates.”
However, I do think the the interruption tax is significant — especially with the quickening of adoption. You use your social network as a filter, which helps both in scoping participation within a pull model of attention management, but also to Liz’s point that my friends are digesting the web for me and perhaps reducing my discovery costs. But the affordance within Twitter of both mobile and web, that not only lets Anil use it (he is Web-only) is what helps me manage attention overload. I can throttle back to web-only and curb interruptions, simply by texting off.
Good thing too, because back when it was called twittr people held back believing what they posted would be interrupting on mostly mobile devices. Lately I think people just go for it, and most consumption is on the web or other clients. I’d love to see some research on posts/user, client use, tracking @username, group identities, geographic dispersion and revealing other undesigned conventions.