Klout Score’s Anti Rapture: Scores Suddenly Drop to Hell

By Dave Wieneke | in Web 2.0: The Social Web, Web Analytics & Measurement | 23 May 2011 | 2 Comments

The Tweeting world woke up today to find that many Klout scores have dropped, in some cases to nearly nothing.

Regardless of what caused this, it represents a major screw-up for the firm that wants to be the “credit score” of the social media world. This is a serious problem for Klout because their measure of influence relies on so little to start with, and competitors are at the ready with equally as little behind their scores.

If this were an intentional change, the firm would have used announcements to managed this change. But the Klout blog has nothing to say. It’s a malfunction, and for Klout, that’s quite a problem. The “Tweeting class” has actually started to act like this score is a fair judge of their influence, so any arbitrary change, even in error, will set them off.

This looks like a glitch. Anyone who has followed Feedburner subscriber numbers is used to seeing Sunday adjustments where numbers drop precipitously for a day then recover. Even Google Analytics has occasional divot days, during which metrics appear way off but then get fixed. But for a new measure, glitches can turn believers in to skeptics.

Klout’s weakness is that it is so open to trickery. Stowe Boyd blogged about Adriaan Pelzer of Raak, who created Twitter bots that spewed interesting things at one and five minute intervals. Guess what? Klout regarded them as celebrities.

What’s wrong with this picture? To start off with, it should not really be possible for a bot to reach a Klout Score of 50 within 80 days merely by Tweeting random (yet entertaining) rubbish every minute, should it?

What Does Klout Score Really Measure?
“Influence” is a terrible choice of word for what’s being measured. Klout score and the like measure the qualities of ENGAGEMENT.

You see, prime ministers and presidents have INFLUENCE; armies listen to them. Warren Buffet has influence over investors. You can see that through engagement.

One wouldn’t try to measure their influence through tweet analysis. That’s momentary engagement. You can measure how connections are reciprocated. Communist leaders on Twitter tend to have lots of followers, but they follow nobody back. Does that mean they have more influence? Or just that they are selfish and self-absorbed? It’s open to interpretation.

As I said to the Boston Globe back in February:
Of course, influence can’t be reliably reduced to a single measure any more than relevancy can. But such scores can take on a reality of its own. FICO credit scores are more valid than tarot card readings, but they are both dependent on who believes in them.

Let’s see if we can deflate some of the language in our ninja-rich industry. The dot.com smug-fest lived off its own inflated words; maybe we can spare social media, and ourselves, that same fate.

And, if seeing that lower Klout score is getting to you, here are some free or cheap options that provide a social capital/sociability measurement of some kind. Until one service builds credibility and proves validity, people will focus on their score that “feels best.” And today, that sure isn’t Klout.

• PeerIndex – http://www.peerindex.net
• TwitterGrader – http://twittergrader.com
• Tweetlevel – http://tweetlevel.edelman.com
• Twitter Grader – http://twittergrader.com/
• Twitalyzer – http://twitalyzer.com
• How Sociable – http://www.howsociable.com
• Postrank – http://www.postrank.com