Meet The First Miners of the New Social Graph

George Stephanopoulos. Wolf Blitzer. Ana Marie Cox. Three powerful people that you might want to get in touch with, especially if you’re in D.C. One man who has the ear of all three of those powerful people is Tony Fratto, known as @TonyFratto on Twitter. Managing Director at Hamilton Place Strategies, the founder of RooseveltRoom.net, a CNBC Contributor and a former US Treasury & White House official, Mr. Fratto is the least well-known (in Twitter terms) of the mere 41 people that all three of our sought-after stars have opted-in to following on Twitter. He could be the diamond in the rough, the guy you want to know. He’s probably got some important things to say, too.

These days, it’s all about

who you don’t know.

That’s the theory behind a group of very interesting software projects being built on top of the giant graph of friend/follower connection data that Twitter exposes about its users. These tools unearth potential connected influencers like Tony Fratto.


Using People to Find People

Name 3 people whom you admire, despise, work with or otherwise pay attention to and tools like HiveMind, Follower Wonk and Twiangulate will quickly calculate who all those people are following in common on Twitter. Tell those services who you are and they’ll expose the people from that list that you aren’t yet following yourself.

Services like this stand in an interesting place online: they aren’t too hard to build and they delivery huge value to their users, but so far they have had a hard time getting people to try them out.

Using social network data is easiest with Twitter today, because of its more open nature, but this kind of work is being done on top of Facebook as well. (See this week’s article The Man Who Looked Into Facebook’s Soul for example.) When it comes to using people to find people, it’s almost literally a 2.0 version of what some of the biggest companies in Web 1.0 did.

Adam Lindemman, CEO of a synaptic web company called Imindi that burned up trying to enter the earth’s atmosphere last year and has since been mothballed, put it very well:

“This is essentially collaborative filtering used in recommendation, for people as the object rather than Books, CDs or Movies (Amazon) or search results (Page Rank). In other words we are seeing the emergence of something like the kinds of tools that helped us navigate the information and e-commerce glut in Web 1.0 and using them to help us manage the main glut of Web 2.0 – which is the abundance of people and the need to literally mine that noise for signal.”

We spoke with the creators of those three services in particular (Twiangulate, HiveMind and Follower Wonk) to find out what they’re thinking as they build and promote social graph analysis tools on top of Twitter.

The Fabulous Uses of Friend Network Analysis

Henry Copeland, the creator of 8-year old blog ad network BlogAds, is building Twiangulate on the side, late at night.

Copeland is the one who told us about Tony Fratto, based on the Twiangulate map of common connections among those three media stars. He also points to a map his service created of the most inter-connected, yet otherwise obscure, friends of actress Demi Moore. That’s kind of creepy.

A more benevolent use case? Say you’re a reporter covering New York city schools. Who might be a good source to contact? Here’s what Copeland recommends: Look at Mayor Michael Bloomberg’s list of followers on Twitter. See which ones look like education-related organizations. Use Twiangulate to see who both that organization and the Mayor are following. Look for an education-related organization on that list and add it to your comparison. Repeat, until you’ve replaced the Mayor and are just looking at a list of who is being followed by 3 of the most-connected education organizations in the city. There’s a good list of people to be in contact with. That’s pretty hot stuff.

Freaky Geeky Stuff

Why are services like this not taking off like wildfire yet? “Right now we’re each building a fascinating tool, but we haven’t quite shaped the handle right so it fits in the user’s hand,” Copeland says of the three companies we looked at doing similar kinds of work.

“Then, what exactly does the customer look like? In many cases the right user doesn’t yet exist. There is not yet 10k people designated as blahblahblow specialists at companies across America. There are geeks and aficionados and crazy people with visions of something that might be palm trees on the horizon… or might be a mirage.

“People need to develop habits of thinking about this stuff. They need someone to explain it to them. They need to develop an easy-to-use vocabulary set for understanding it and communicating about it. Right now ‘social network analysis’ just doesn’t trip off the average person’s tongue.”

It may not roll of the average person’s tongue, but there’s clearly some very serious interest in it. Kevin Marshall is the creator of a bundle of social graph analysis tools called Wow.ly, including one much like Twiangulate, called HiveMind or Grou.pe. HiveMind will show you who you aren’t following that a group of selected Twitter users are – and it will email you when that list changes.

“I have a couple of friends that are in the VC business around New York city,” Marshall told us.

“They wanted make sure that they knew about any new deals and companies at least as fast as their competitors. So this is all just one tool to help with that sort of thing…without having to constantly check who their competitors are following manually.

“It actually feels a little like an invasion of privacy when you start getting the alerts though, because you get to see when people you are interested in start or stop following other people. But from a ‘business’ point of view, I think it’s really powerful.”

The Low Hanging Fruit

No potential privacy violation would be complete without politics and advertising. Sure enough, our third miner of the new social graph focuses on those two parts of the world.

Peter Bray lives in Portland, Oregon and built FollowerWonk.com. Follower Wonk displays common followers, as well as people that users are following mutually. It also offers some visualizations and the ability to search inside users’ Twitter bio fields. Bray thinks this technology has real potential to change the world, in the short term even.

“As Follower Wonk matures, we imagine an application where a political campaign or a brand can slice & dice followers into highly leveragable segments. For example, a Democratic candidate might be able to find cross-over votes by targeting Twitter users who follow the competing GOP candidate, but also, say @sierra_club or @barackobama. Campaigns, or brands, can find and utilize these natural ‘tension points’, not only in terms of actual individual users, but also what interests (environment?) or messages might be potent. As social media matures, these sorts of strategies, already so common in the world of list management, are naturally going to play a significant role.”

We asked Mr. Bray, if these tools were so darned potent, why aren’t more people excited about them yet?

“In the universe of Google users, the percentage using, say, Google Analytics or Website Optimizer is teensy,” he replied.

“Similarly, the total number of Twitter users that are going to find value in slicing & dicing their followers is small. But I think that the benefits of such user profiling and tracking are potent, and increasingly evident among brands and campaigns.

“We’ve seen this in one of our other applications, Versionista.com, where the McCain site used our Web page tracking and versioning tool to highlight changes to Obama’s policy pages. Who would have thought that tracking changes to a campaign Web site would reveal important talking points? And similarly, you’re right that few people right now realize the power of Twitter, and social media generally, in terms of the ability to find and engage a receptive audience through advanced profiling and analytics. Or the ability to deduce new messages and arguments.”

Can One of These Visions Come to Fruition?

These are three men have built different tools that do roughly the same things but in different ways, and with different go-to-market visions. There are others, too, like Tickery and Klout. Will any of them be able to popularize something where there’s a level of strategic thought required in order to capture substantial value from a social networking world all-too-often dismissed as trivial?

It’s a complicated but thought provoking question. One that will be discussed on Twitter no doubt. Even there, Henry Copeland, Kevin Marshall and Peter Bray will take different approaches, though. There is not a single person that all three of them are following! There also wasn’t anyone else that was following all three of these trailblazers, either. Until I just did.

Is this kind of analysis a privacy violation? How can privacy on sites like Twitter, Facebook and whatever comes next be balanced with interests in innovation and profit? These are the kinds of questions that the web will need to ask if services like these three become popular. They offer so much value, it seems inevitable that they will.

You can find and follow the ReadWriteWeb team on Twitter here and on Facebook here.

Disclosure: Color commentary above was provided by CEO of Imindi, a company the author has had a consulting relationship with in the past. It was also quite articulate.

Facebook Comments