"Our ambition is to up-level the model of social networking. We want to support asymmetric relationships with brands and celebrities - but we also wanted to support intimate communication with real life friends and family. Could we do both of those things? Being all things to all people is the path to hell - but could we do it?"
Bradley Horowitz, VP of Product Management and a leader of the Google Plus social network, said a lot of very interesting things today in an interview with Tim O'Reilly at the Strata Summit on Big Data. One of the most interesting, though, was to quietly introduce the data scientist behind Google Plus and its future development. Dr. Andrew Tomkins (above) used to be the Chief Scientist of Search at Yahoo but two years ago left and joined Google. No one noticed. At least no one indexed by Google News. Today we learned what Tomkins has been doing since joining Google. He's nerding-out on social network user activity data and gathering observations to help Google Plus aim to be all things to all people. Andrew Tomkins is charting the path to hell.
Plus as Data-Driven
"We have a complex service that contains a lot of components and we want to do it in a data-driven way," Horowitz told O'Reilly on stage this afternoon. "You'll begin to see data-driven insights come out into the light of day soon, based on a bleeding edge understanding of how people are using social networks. We have a research team lead by Andrew Tomkins, who is one of the great data scientists and who is coming up with insights that are informing the development of our service."
What kinds of data is the Tomkin's Plus research team surfacing so far? Horowitz noted several in his interview.
- Users of Plus are two to three times as likely to post content to private circles than they are to post it publicly. In other words, the numbers say the Circles metaphor is working so far.
- The new Search functionality will in the future identify topic experts, based on algorithmic analysis of the things they discuss and other signals.
- Automatic prioritization of social contacts: "We have great data to determine who you really care about. The phone contacts list is key. The data belongs to the users though and we need to find the best way to serve it up to them."
Horowitz and O'Reilly talked dreamily about a sensor-rich future where almost unimaginable technologies were built on tidal waves of data. "Imagine we all opted-in and donated our microphone sensors in this room to capture an aggregate of data," he imagined. "There will be sensors like dust everywhere and it will be [technologists' job] to harvest that data and return it as killer apps."
Who's Andrew Tomkins?
Andrew Tomkins, whose time at Yahoo overlapped with Horowitz and who's worked at IBM as well, is quoted in Steven Levy's deep dive coverage of Plus at launch regarding Sparks, the persistent search feature that sounded today like it was going to get a refresh as part of the launch of search.
Horowitz said today that Tomkins leads the Plus research team, is one of the best in the business and that research is going to drive future development and publicly shared insights in the near term future.
Tomkins got a Computer Science undergrad degree from MIT in the late 80's, then a CS PhD from Carnegie Mellon in the 90's. He's co-authored 12 research papers about data mining published across 9 different publications in the past 10 years.
But he has only shared one post publicly, outside of Circles, on Google Plus - ever. He's followed by two hundred people on that service and not a single journalist (until this afternoon, at least). He seems quite open to talking to press, though, so hopefully I'll follow up with him after today and offer a more detailed discussion of his thinking.
None the less, if Paul Adams, former Googler turned Facebook employee and author of the widely-read presentation that inspired the Circles metaphor, is the philosophical granddaddy of the hot new social network - it sounds like Andrew Tomkins could be the man who is leading the young service into a data-driven future.
Good luck, Dr. Tomkins, may you build something that is all things to all people. And if "don't be evil" seems an antiquated Google slogan, then perhaps we should say...may it not in fact be a path to hell.