September 19, 2011
The subject of Klout has come up a few times on Twitter, so I’m posting this so I can point people toward a few articles I’ve found useful. Something I can’t do in 140 characters. Which proves my eventual point.
[Update (12/11/11): I wrote this post before Klout renovated their algorithm in a very opaque way, sending many to their keyboards in anger as their klout scores plummeted for no apparent reason. While I’m not that surprised Klout don’t tell people more (it would make it even easier to game), they do appear to have crashed their cred (see posts added, below). But my problem was never with their algorithm or lack of transparency – it was with the very idea that you could judge someone’s influence by measuring their activity.]
[Update (1/4/2012): It’s April Fool’s day, so i couldn’t resist adding the picture below, courtesy of PRDaily on pinterest]
More updates at the end of the post.
For those who don’t know it, Klout purports to measure your social media influence by looking at your profile on Twitter, Facebook, etc., and those of your followers, how often they retweet you, etc. They’ve got some fancy mathematics and sexy terms like True Reach, Amplification, and so on, which you can read all about on their site.
The idea, of course, is seductive. How much easier it would be if you could boil everything down to just one number! No actual reading required – even an MBA can compare two numbers! 😉
My gut instinct was to ignore Klout, simply because nobody can be boiled down to a single number, and certainly not in a field as complex as influence and ideas. But when I saw job ads appearing with minimum Klout score requirements (I’m helping a client recruit), it bothered me so much I did some research. Here’s what I found:
As Mack Collier points out in Online Influence Is More Than Just Social Media Activity, Klout and its competitors measure social media activity, not influence. Moreover, as discussed in the comments, it’s pretty easy to game the system – i.e., perform activities which raise your score, but do nothing for your influence.
In Does a high Klout score make you a Twitter influencer? Not really, Mark Schaefer goes further, debunking some cherished ideas about the influence of your Twitter account. He has evidence that it’s probably zero, which is a lot less than Klout thinks.
Echoing my initial reaction, someone pointed out that it’s just not possible to boil influence down to a single number.
“What are you influential at? Financial markets? Electrical systems? Human resources? Hot rods? Surely you are specialized in something and not others … Having this “average out” is doing you a big disservice.”
Tom Foremski concurs, pointing out that:
“If you … need to consult these services, this probably shows that you are clueless about who is important in the very markets that you are selling into… [Your Klout score] tells you nothing about the context … about the hot topics, the major issues in those sectors; and nothing about the culture ….Klout, PeerIndex, Empire Avenue, and the others, provide shortcuts without insights”.
And in case this is all too theoretical and you like numbers, read the deconstruction by Alex Braunstein, statistician and search quality engineer. After analysing the scores of four comparative groups of users, he concludes:
“there are some serious inconsistencies with Klout that render it nearly meaningless in some circumstances. It often does not correctly order individuals in terms of how influential they are, is easy to game higher simply by adding a Facebook account, and does not respect some very basic monotonicity rules.”
Slacktivism does not measure influence
But how can this be? How can actress Alyssa Milano (Klout: 84, almost 1.2 million followers on Twitter) tweet a link to a book on Amazon and not drive a single additional book sale? Particularly given the book was, ironically, Connected: The Surprising Power of Our Social Networks & How They Shape Our Lives?
Because Twitter doesn’t work like that. Neither does Facebook, or indeed much else. That’d be too easy.
You can reach people through all sorts of channels, but real influence, in my book, is when you change behaviour. Perhaps you made them think something new. Maybe you made them act. Maybe, one day, it appears in a policy or company decision
Prompting someone to retweet or Like something might be nice, but it’s not influence. Reading a tweet or clicking Like is called slacktivism for a reason – it doesn’t demand much effort. And that’s why it’s relatively easy to get people to do. By the same token, however, it won’t make a lasting impression on them, because they have not invested much. They’ve moved on already.
Influence is about ideas
The best way you can influence someone via Twitter or Facebook is probably to send them somewhere else where they can read something useful and substantial. Like your blog, if you have one. Or someone else’s.
Because influence is, above all, about ideas. If you can get the right idea in front of the right people, then they’ll carry it forward. So it’s not quantity of followers and social media activity, which is what Klout and others measure. It’s the quality of ideas that counts. And they’re usually on blogs. Which Klout ignores. Completely.
As Mark Schaefer points out:
“Marketers should be looking for influence in blog communities. That is where the real magic is happening. The connections built through a blog community are extremely strong, especially when compared with the weak ties on Twitter.”
So the good news is that the extra effort that goes into a blog is worth it.
The bad news is that tracking the propagation of ideas is hard – you need to read and understand them, rather than just order numbers on a scale.
Which is why hiring someone on the basis of their Klout score is more than just lazy, it’s stupid.
Go on, prove me wrong
Want to prove me wrong? Go ahead, try and influence me! I’m open to counterarguments.
But you’ll need to use the comments. Because there’s not enough space in a tweet.
Which means that, by trying, you’ll be proving my point.
(author smugness evaporates as he realises he’ll now get no comments)
PS And since you asked …
Go on, check out my pitiful score if you must (my score is 41, probably because I’ve always found Facebook irritating).
Who Klout says I influence is laughable – they’d probably have done better if they just chose 5 random commenters to my blog. It really shows how completely they ‘miss’ the influence of ideas longer than 140 characters – but how can they possibly track those?
What they say I’m apparently influential about is even better:
- Journalism, although I’m no journalist
- Public relations, although I don’t do that either
- Politics, although I’m apolitical
- Food ????? well, yes, I do eat, but I’ve never mentioned it to anyone online, so frankly I’ve no idea where that comes from
and … wait for it …
Yep, apparently I’m influential about Klout itself. Does that make this post more influential? Does the tail wag the dog?
[April 2012-present]: since the Great Klout Algorithm Shakeup, some other posts worth a look:
- My Klout Experiment And The Disturbing Results
- Klout, Blondes and Other Distractions
- Why Klout is dangerous
- ROI is a good book…but Klout still sucks!
- Klout Doesn’t Really Measure Influence [STUDY]
- What Your Klout Score Really Means (Wired Magazine)]
- The Problem With Measuring Digital Influence (Dr Michael Wu on Techcrunch)
[20/11/2012] The Techcrunch article (via @thefashioncloud), in particular, is worth a close look, as it spells out the absolute impossibility of modelling digital influence – without independent sources of data to validate the model, you end up with circular reasoning.
Moreover, such models are usually based on massive overgeneralisations of the data; their algorithms are more susceptible to gaming and are easier to game. He concludes that “An influence score is really just a measure of how well people game the influence scoring algorithm“, which leads to a rather perverse result:
If you tweet a lot yesterday your influence score jumps … would you continue to tweet more? Most people would… This has created a lot of loud mouths who are not actually influential in any meaningful way.
Because behaviors that game the system are typically a lot simpler than behaviors that are truly influential, IEO [Influence Engine Optimisation] will tend to changes people’s behavior in a way that pushes them further away from being truly influential.
The short version: caring about Klout makes you less influential. Hence the title of this post.Author : Mathew Lowry