Analyzing Your Relationships

Posted by Shaun Abrahamson on Wednesday, 4 June 2008

How we interact with others, is increasingly under the microscope. In 1996 when I set out to write a thesis, I spent some time looking at IBM research which was analyzing e-mail communications. Who was talking with whom, to try understand various dynamics of communications in the enterprise. It looked interesting then and a few companies seemed to appear and disappear around this theme.

At the core of Google’s success, was the understanding of how to interpret the excellent work which people do when creating links between content. If you havent seen it, best to check out “Anatomy of a … Search Engine”.

Today, in addition to the links, we are generating enormous amounts of data as a byproduct of our communications. And companies are emerging to attempt to make sense of it. Where existing analytics focuses and WHAT people are doing on a site, the new tools enable us to understand WHO people are interacting with. And this has some interesting implications. Following are a few examples:

Daylife – keeps making news analysis more interesting. And recently unveiled some demonstrations of just how useful their analysis can be in the form of “social graphs” and other interesting analysis tools. My favorite is quite simple – for a given phrase or topic, figure out which news organizations or journalists are providing coverage. PR 2.0? We like. Some examples for: social graph, obama, mccain, wii fit.

Pluribo – is being very secretive of late. However, they have already demonstrated some of their potential to understand and interpret user opinions. Part of the process involves identifying more influential and reliable contributors. And so we expect, that part of this will show up in the form of affinities between contributors and categories, between contributors and one another, etc. Sorry they are back in stealth mode, so no examples.

SocialAmp – is providing a new view into interactions on social networks. They have created a rating systems to indicate how “influential” users are, for example. While most of the talk seems to be around advertising, some recognize the inherent value of conversations. It looks like Social Amp may be arriving at a Page Rank equivalent for social networks. Much like in the real world, people need cues as they work the room – Social Amp may have one social networks.

Flickr - Interestingness seems like something which might come out of a Monty Python piece, but it more than just clever naming. Flickr doesnt go into much detail about how this works, but there are plenty of speculators who build on their hints. The end result is that without knowing anything about the photos, Flickr is able to identify the most interesting ones by studying interaction.

While some focus on the loss of privacy, I think it is fascinating to think about how much we are able to learn from simply observing the byproducts of communications. Hopefully its good for business, but more importantly, perhaps we’ll understand one another a little better.

Disclosure: we are investors in both Daylife and Pluribo.

Posted in: Analytics | Tagged: daylife, flickr, google, graph theory, interestingness, link analysis, pluribo, social amp, social graph | No Comments