It took a worldwide financial meltdown for Twitter to finally cough up the IM hairball. At BearHug Camp, I spent about 10 of the 30 minute executive visitation trying to pin down Jack Dorsey, Biz Stone, and Alex Payne on when exactly Track and IM would be back, and in what order. Turns out the IM part isn’t coming back; it’s been moved from Broken to Build.
Evan Williams delivers the bad news with a refreshing frankness, suggesting the ROI of IM services for a small percentage of Twitter users puts it down the list below other more pressing priorities. And at the bottom of the email, he points at a fledgling third-party service that gives you a way of “tweeting” over the Jabber XMPP gateway. The author is mulling how to provide access to users’ follows. No mention is made of Track, of course.
Williams’ acknowledgment of the realities of surviving the viral success of a platform built on ideas but not necessarily execution comes not a moment too soon. Robert Scoble suggests in a Google Reader note that “Twitter is working on things to improve its already strong lock-in,” namely its dominant user base. But what this really does is focus pressure once again on Twitter’s business model, and begs the question once again why Twitter is taking so long to pass along permission to third parties such as Gnip to provide the scalable services the “small” minority of realtime users crave.
With Microsoft already releasing pre-PDC information about Live Mesh’s PubSub underpinnings, and many of the open source and hybrid third party services using Google App Engine and Amazon Web Services, realtime Tw*tter has a strong chance of surviving the nuclear financial winter. Startups need nothing more than Twitter’s permission and little or no venture money to enter the network of cooperating services.
Now that IM is dead meat, Twitter needs to clarify its deals with middleman services such as Gnip, and open up the kimono on independent developers locking down access. We’ll do our part to shine a light on these permissions or lack of them.