Here at Redfin, we’ve been thinking a lot lately about Twitter as a development platform. The problem with that platform today is that nearly all Twitter applications are applications for using Twitter, not applications that use Twitter for doing something else.
Redfin, for example, doesn’t want to build another Twitter client. We want to build a Twitter-powered real estate application, that allows our users to follow a neighborhood on Twitter. Any new listing, new sale, price change, agent insight or market update would show up as a tweet in neighborhood followers’ timelines.
This would broaden Redfin’s reach, and give users an easy way to see what’s new in the market in real-time, without our having to ask anyone to register an account, define her own queries, or even open an email. Even better, Redfin users could get text-message alerts via Twitter when a new listing came on the market.
To support this, Twitter would have to build a high-scale system for bulk creation and administration of tens of thousands of Twitter accounts, one in our case for each neighborhood, which I believe Twitter will support one day soon. And the load imposed on Twitter would also be imposed on Redfin, as Redfin.com would have to process queries against every neighborhood, every few minutes. Twitter’s real-time scale challenges would become challenges for the whole Web to solve.
But it would be worth it. As Cloudera’s Jeffrey Hammerbacher recently noted in an engineer-to-engineer talk at Redfin on Hadoop, machines produce far more data than people ever will.
The possibilities for Twitter-ifying the Web are endless, as almost any series of events recorded on the web could become a Twitter stream: every time the Cal football team scores, whenever a friend uploads a photo to Flickr, or a customer lodges a complaint, I want it in my timeline. If Twitter is really going to be the pulse of the planet, it has to let websites — not just people — create and manage millions of accounts.
Twitter, I hope you take the plunge. If you do, the real-time web could increase in size by an order of magnitude.
UPDATE: Robbie Allen at StatSheets responds to our post, explaining that the company tried to deliver exactly what we wanted for sports enthusiasts, which Twitter initially discouraged and then embraced.