What is Klout, and how well does it work?

A few days ago I wrote on my other blog about some online software tools, among which was Klout, which I've been trying out Some online software tools | Khanya.

Klout is a website that purports to tell you how much influence you have in social networks, and who you are most in contact with. First impressions were difficult to gauge, because I discovered that Klout takes a bit of time to get up to speed. It starts off by adding your Twitter followers (or is it followees) as "friends" or "influencers", and then after a day or two goes on to Facebook, Google+ and other networks. So you probably need to use it for a week or so to see how it actually works.

Even after a week, however, it becomes apparent that it is heavily weighted towards Twitter. If you want to list your "friends" it loads the people you follow on Twitter, rather than your Facebook friends, for example, and it appears to display them in random order. Taking one of my blogging friends, Miss Eagle, and doing a comparison, I can learn what Klout thinks are our spheres of influence:

Sorry if that's a bit hard to read: you can blame the new and downgraded Blogger interface, which does not appear to let you adjust the size of such pictures before posting them. What it says is that I do 93% of my stuff on Facebook, while Miss Eagle does 100% of her stuff on Twitter. And it also says "You use Facebook as the primary way to spread your influence. Twitter is Miss Eagle's primary network of influence."

How accurate it is, I have no idea.

One of the things I remarked on in my original post was the topics in which Klout appeared to think I was influential: Some online software tools | Khanya:

Among the rather strange things Klout tells me is that I am influential in Singapore (first and last time I was there was back in 1985) and that I’m more influential in “Celebrities” than in “Christianity” — if you look at the tag cloud in the right side bar you’ll see that “Christianity” is quite big, and “celebrities”, if it appears at all, is very small.

I've since discovered that this can be altered in various ways. You can remove topics that don't really interest you. You are also given points, and you can use these to add new topics to either your own page, or to those of your friends and influencers. For example, I used five points to add "Socialism" to the topics on which another blogging friend, Chris Hall, had influenced me. Klout had apparently not detected it, and once I added it, it moved to the top of his influential topics.

But the "topics" of influence also seem arbitrary, and quite bizarre, and this, I think, is one of the biggest weaknesses of Klout. I wanted to add "Missiology", which is my academic speciality, as well as that of a lot of other people in my network, but Klout would not let me. It was not in the "topic dictionary" The nearest I could find was "Theology", which is fine as a generic topic, but I also saw "The California Pacific School of Theology (Japan)", and a couple of other similar entries. That is a really silly topic -- if you are going to add one theology school, you should add them all, but surely a "topic" is for a discipline, not a single institution. If a single school can have a topic all to itself, then surely every single theology school in the whole world should have its own topic? But it makes more sense to have a discipline as a topic, and not to include only a couple of the institutions where that discipline is taught. Yet a whole discipline, Missiology (aka Mission Studies), which is found in hundreds of schools around the world, has no topic at all.

I wrote to Klout about this, and their reply awas not reassuring. I also looked to add "Church History" as a topic. Nothing doing. They offered "Church", "Baptist Church" and "Winston Churchill".

"Baptist Church"? What about other denominations? Try "Anglican Church"? No, nothing doing. You can have "The Riverina Anglican College (University)" as a topic, but not the "Anglican Church" or "Anglican Communion".

And it gets worse. Not only can't you have Church History, you can't have History. You can have Black History Month, but you can't have History.

Now Missiology may be a fairly obscure academic discipline, but History is big. But sorry, historians have no Klout, not even if they are black historians. They only haveKlout if they are Black History Monthians.

I would thus say that Klout's topic dictionary is well and truly screwed up. It is badly thought-out, badly designed and badly implemented. It has lots of very narrow sub-sub-topics, but in many cases the main topics that they should be under are missing.

In the light of that, news stories like this are very scary indeed: What Your Klout Score Really Means | Epicenter | Wired.com:

Last spring Sam Fiorella was recruited for a VP position at a large Toronto marketing agency. With 15 years of experience consulting for major brands like AOL, Ford, and Kraft, Fiorella felt confident in his qualifications. But midway through the interview, he was caught off guard when his interviewer asked him for his Klout score. Fiorella hesitated awkwardly before confessing that he had no idea what a Klout score was.
The interviewer pulled up the web page for Klout.com—a service that purports to measure users’ online influence on a scale from 1 to 100—and angled the monitor so that Fiorella could see the humbling result for himself: His score was 34. “He cut the interview short pretty soon after that,” Fiorella says. Later he learned that he’d been eliminated as a candidate specifically because his Klout score was too low. “They hired a guy whose score was 67.”

Consider the case that I mentioned above -- Chris Hall. Klout had not picked up "Socialism" as one of his topics of interest, yet when I added it, it turned out that he was more influential in that than in any of the topics that Klout did pick up. And perhaps Klout had done something similar with Sam Fiorella in the story above, but because he did not know about Klout, there was nothing he could do about it.

If these things are flawed, there's no way of telling how Klout calculates influence even on the flawed and inadequate data it uses -- what if its algorithms are as flawed as its data?

Klout is an interesting concept, and it is quite fun to compare yourself with your friends and to see which topics you are interested or influential in. But I'm not sure how seriously it can be taken when its data are so obviously flawed.

You can do something to improve it, though. You can check your friends, and see if their topics reflect those that you discuss with them most frequently. You can help to make their scores more accurate -- provided, of course, that their areas of expertise have made it into Klout's topic database in the first place.

Oh yes, and if you're feeling kindly disposed towards me, please retweet a few of my tweets. It's not that I'm looking for a job in marketing or anything, but you never know when you're going to need it.