Influence vs Obeisance at the Independent

There was an educational article in the Undiependent yersterday, in a sort of "OmiGod" way. It was about the UK "Top 100 Twitterers", and was ostensibly based on the PeerIndex algorithm. Now those of you who know about measuring influence on Twitter will know we are in Iteration 4 of Influence Monitoring.

Iteration 1 was "how many followers do they have" - the logic being that if you have lots of followers, you are influential. Turns out that lost of peopel gamed the system by following as many peopel as that can, and as so many had "autofollow" on, they got followed back. Ergo Joe Social climber is now instantly influential.

Iteration 2 was to try and gauge influence by the "Follow/followed" ratio - ie if you were truly influential, then lots if people follow you and you follow no one else. Cue above spammers to start unfollowing many of the people thay had followed to get their ratios up. Given no auto-unfollow, all their "audience" remained. The only problem was, it was also becoming clear that ratios were not the be all and end all, people actually had to say something (ie volume of twts) and make it interesting enough to actually influence other people. Not only that, it was becoming clear that on social media you were having to be social - the "pimp and dump" mode was in fact counter-pruductive.

Iteration 3 was to be a bit more sophisticated, and to see how often other people repeated (re-tweeted) what you said. This of course led to a Retweet Explosion by the sort of people who want to be Influential on Twitter (The Social Media Climbers) , but the only measurable influence was lots of p*ssed off people saying "stop retweeting everything or I'll unfollow you, you eejit" or similar, and changing of Retweet rules on Twitter. Also, as many grumpy people pointed out when Ashton Kutcher and friends had a "race for a million followers", people may let him influence their decision on a movie to watch, but not really on things they felt he knew nothing about. But this was the Explosion of Corporate interest in Social Advertising, so the newly minted Social PR agencies had to have something to feed the beast, so needs must and many attempts - some very good (eg Mat Morrison) and some much less good - were made to understand more deeply how influence works on a social network. And guess what - it's complicated

Iteration 4, where we now are, is trying to be more sophisticated - trying to blend in all the above approaches, plus add some form of "where is this person an authority", some understanding of "when is a retweet an influence and when is it just some prat trying to curry favour with the person they are channeling" etc, and even looking at outside calibration in other social media systems.

Anyway, the Independent yesterday invented Iteration 5 with its Top 100 UK Twitter methodology. It is so breathtakingly simple we wonder why people didn't think of it before:

Twitter is not yet evenly distributed: some fields, like technology and science, have very large communities of fans; others, like literature or art, have more incipient Twitter communities – but are none the less plainly influential. So we also searched specifically (further down the rankings) for people who had particular resonance in certain fields; and further refined our list by focusing on those who were especially trusted by other experts. This was where the advice of our expert panel came in.

Our five-person panel comprised Azeem Azhar, founder of PeerIndex; Ian Burrell, media editor of The Independent; Julia Hobsbawm, founder of Julia Hobsbawm Consulting and the media networking business Editorial Intelligence; Steve Moore, director of the Big Society Network; and Stefan Stern, cultural commentator and director of strategy at the Edelman PR agency. They analysed the initial results of PeerIndex's calculations, made recommendations and disqualifications, and argued for extra weighting for certain key tweeters whose significance to the medium is, in their expert judgement, greater than the numbers imply. Such interventions were kept to a minimum but give our final list the crucial virtue of having been assembled through the prism of human intelligence.

So, what you do is you take the output of an Iteration 4 algorithm system (Peerindex in this case). you take a long, hard, considered look at the algorithmic output and say "f*ck that for a game of soldiers, who are all these odd people - we need more people like us and some slebs with big t*ts". In other words, get a small closed circle of meedja types, look at the output, and chuck away anyone who doesn't look like a sleb or a current meedja darling, as those are the Right Stuff - the people who are (to quote) "plainly influential" and "especially trusted by other experts".

Well no, because if they were they would have higher scores on the algorithm, surely? - the population on Twitter is now quite large, its goes way beyond early techies. The reason - I would argue - that their scores are not higher is that people on Twitter are not being influenced by them as much, and there are (I propose) two main reasons for that:

(i) The Mainstream media is sleb-obsessed, but its mainly because slebs are walking advertising billboards, whereas said media have no tight feedback mechanism for measuring (a) what their readers really think of said slebs and (b) how influential they are (For example some slebs are influential insofar as their lives are total train wrecks - but no one would actually buy anything on their say so)

(ii) Twitter is (largely) still ABC1 demographic - people of wit, discernment and taste - or at any rate, they expect to see interesting interaction with their favourite famous ones. Going from Mass to Social Media is a bit like the jump from Silent to Talkie movies - suddenly one finds some of the stars of the silent screens sound ridiculous, and the 'Net effect is people run away from them.

The fundamental problem is that the Mainstream Media is unable to recognise anyone that the mainstream media doesn't recognise, if you get my drift - which is of course why it is becoming daily less relevant, while a new lansdcape of media performers have come up via blogging, Web TV et al.

Now I believe the algorithm is probably as good as any of this Iteration (It has to be, my Peer Index score means I should have been on that Indy Top 100, in about position 75-80* ) but you have to ask, why bother with an algorithm method if you can just grab a quick coffee clatch and over a circle-jerk decide, based on (your perception of) mainstream media fame, who is influential on an entirely new medium. Straight up Hello or Cosmo's alley I'd of thought, but the Erudite, Intellectual, Independent?

Or is it now the Obeisant, Populist, Undiependent?

Update - a few days later, Jeremiah Owyang makes the same observation


* Hence this post - and the grapes were sour as well