Relevance-Aware Activity Streams



Liz Gannes of GigaOm does a nice job summarizing a recent interview of Twitter CEO Ev Williams. Williams provided a few significant insight into how he thinks Twitter (and by extension real-time activity stream apps)  is different from email and search.


"Blogging (Williams was previously the founder of Blogger) and Tweeting can be different (and better) than email, he said, because people who have something to say can find their audience. That’s a much better situation for both the publisher of the information and the consumer of it. So recipient-based media can scale better “in a world of infinite information,” he said.

That’s also a contrast to Google, said Williams, which serves more purpose-driven needs versus Twitter’s focus on “an interest-based world.”

“Google is very good at ‘I need to solve a problem, I need to buy something, I need an answer,” he said. “Twitter is more ‘I’m interested in many things, I don’t know what I need to know.’” Where Google is more likely to be gamed by a company like Demand Media, Twitter is a different beast.

However, there’s still the problem of filtering information on Twitter. “What we need to get much better at is scaling that system so you don’t have to pay attention to everything, but you don’t miss the stuff you care about,” Williams said. He said more such products were on the way.

Williams cited Google CEO Eric Schmidt, who said recently that more information is now produced in two days than was in all of time before 2003. Williams said automated streams of information from services like Fitbit and Blippy —  in addition to the proliferation of media — will only add to that problem.

Williams also said to expect forthcoming products that would help filter relevant tweets around events, similar to what it’s doing with location. This would go beyond the user-developed convention of hashtags, he said, though he didn’t elaborate.

Developing relevance-aware activity streams will require a flexible filtering experience that enables users to filter across a number of different topics or variables.  It seems that Ev hints at developing certain types of filters – for example filtering based on events.  

Does that mean we'll see a slow roll out of category based filtering tools versus a broader, user-driven filtering tool capable of coarse and/or fine filtering?  Will it be an open-friendly approach or closed?

Developing relevance-aware activity streams is an important step in the evolution of the real-time web.  

If, as Ev points out, recipient-based media can scale better in a world of infinite information, then the relevance aware filtering model must be able to scale as well.

(image from


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s