PeerIndex plans to build a DNA profile of your expertise

Viewsflow appeared in the last few months as a sort of high-brow filtering mechanism for Twitter users. It’s majored on finding the best people who tweet about Peak Oil, or China’s economy, for instance. It’s been producing quite a nice little daily newsletter full of the kinds of links you’d expect more from an editor on The Economist or The Atlantic Monthly. And that’s precisely the point. If you can work out who really knows their stuff on Twitter, you can get some great knowledge and insight, right?

The realtime nature of the stream lends itself well to the coming world (which may already have arrived?) where we start to make trading decisions based on the data we can pull from social networks. This is well known as a discussion.

Now, like many startups, now Viewsflow has stumbled on its real business. In around June/July it’ll be launching PeerIndex, a data play based on the algorithms it developed building Viewsflow.

The trouble with the content on social networks is that the real underlying data right now is not really that great. Yes, Tweetmeme, for instance, gives you an indication of retweets and therefore how popular a link is, but what it doesn’t give you is how “interesting” or influential the people are who retweeted that link. That’s actually far more useful.

A story on China recommended by five people is just another story. A story on China recommended by five people who are known to be discriminating on China is likely to be a highly qualified story. There’s a big difference there.

In addition, although startups like Unvarnished are trying to produce a “Yelp for people”, this is actually going to be pretty unhelpful (aside from the massive class actions it will probably attract) and no doubt the data will be bad and gameable. What you need are Morningstar rankings for people, in the way Morningstar ranks Mutual funds.

So PeerIndex is not a pure Twitter play. The idea is to build a topic based database. Right now PeerIndex is currently indexing 724,000 people, based on the real acitivity of what they share on anything from Twitter to blogs to LinkedIn, you name it. They are adding 17,000 people a day. It’s like building up a DNA profile for expertise.

Interestingly, there is also a cost for over-sharing – noisy people who spam their feeds get punished in the index. PeerIndex will rank people in their ability to share things which other people find interesting.

PeerIndex will therefore be an analytics layer cast over third party platforms. The idea is to identify a community of specialists across any platform.

So what’s a use case? Well, a third party could basically build custom publications fast, e.g. around business news about the car industry in India. Or solar in Africa.

The database could also be used to allow a business to find people who know about niche subjects or can influence their communtiy.

Some obvious competitors to PeerIndex reside currently in the Twittersphere like Klout, twittergrader and twitalyzer. But Twitter influence monitors like Klout look mainly at followers. PeerIndex will look at how a particular person is part of a diffusion network for interesting content. So whether you have 1 follower or 1000 it doesn’t matter if everything you post is crap. Thus, the people who game their Twitter follow count and post rubbish will become even less influential. Hooray for that.

And in fact the results for a search on “Venture Capital” influencers on Klout and twittergrader are clearly showing the problems with their follower approach.



And here is the answer on Viewsflow which is using the PeerIndex engine.

There is a six person team behind London-based PeerIndex, headed up by CEO Azeem Azhar, a serial entrepreneur and angel investor.

The investors include Bill Emmott, ex editor of The Economist; Sherry Couto, angel; Stefan Glanzer, angel; Ab Banerjee of Raw Communications, Ken Olisa, Ditlev Schwanenflugel, Anish Chandaria and Shamil Chandaria, Chair of Omniphone.