Klout’s no good and here’s why

I was recently shocked to see social media thinker David Graham say that the sentiment around Klout is going to change. If anything, the sentiment around Klout is going to die off as more people realise the algorithm and methodologies behind it are seriously flawed.

Klout adds impetus to the saying that “there is no truth, only perception” because it acts as a mouthpiece for it — showing users what it perceives to be the truth about their online personas.

This filter of perception isn’t infallible either, it’s incredibly easy to game it. Yesterday I logged in and gave John Beale (not a bully at all) +K for “Bullying”. The new updated version also takes Wikipedia articles into account, especially ones with a large amount of links embedded in them. Most sycophants have written their own Wikipedia entries, and it’s not hard to rustle up a couple of hundred links either. Telling Twitter the funny things that Klout considers one to be influential on, is also becoming somewhat of an internet meme. Cue the Klout applause.

Graham argues that the fact that Klout is increasing its use of signals from 100 to 400 is a good thing, but that’s only good if the underlying algorithm and methodology is correct in the first place. In this case, when I tweet an old schoolmaster that I hope he has a few beers for the rugby, the Klout algorithm assumes I have Klout when it comes to “Beer”. That couldn’t be further from the case as I’m a “Rum and Coke” man in real life. This happens because Klout incorrectly assumes that this social meta-data is somehow linked to influence and expertise.

The quantity of data points is not as important as the quality of the data point and the relationship between data points. In the above example, Klout assumes that because I am telling someone to have a beer, that I have influence over them. It doesn’t take into account whether that person takes me seriously or not.

What’s worse is that this inaccurate means of judgement is being used in real life scenarios (job interviews) to (ironically) influence people with their decision-making, says CEO Joe Fernandez:
“Klout is basically your social credit score. Consumers should care because it affects the way employers, companies and everyone looks at your ability to spread information as a critical part of the attention economy today.”

Other than the serious methodology flaws, Klout is flawed because it doesn’t accurately portray our interests and influences in real life, merely what we broadcast on the internet. While that’s got more to do with human nature than the makeup of Klout, it is a damaging result of it. If Klout is to be of any use, it is to show us how separated our online persona is from our real life one.