Tuesday 15 June 2010

Why I have doubt about @Klout

As a self-proclaimed figure-fetisjist, I like to get data and stats on pretty much anything that interests me, in order to juggle that into information.
I figured out that I'm 1.5-10 times as "strong" as my current company, that Tweeps are 35 times as active as Facebook users, and other findings - some of which aren't that appropriate to publish (yet?)

On Twitter of course, I have used TweetStats to find about what I tweet, and when (I really should get a sleeping schedule, I know) and Klout to measure my "influence": a well-elaborated set of variables weighted individually, forming one final score

But, since 3-4 weeks, I am doubting Klout: actually, just about since Klout changed their quadrant into a dekahexadrant (16 pieces) aka Klout 2.0

First of all, I highly appreciate the service. Apart from being free, it has always been a very good and reliable way to compare influence across all different kinds of users. It is a small company of half a dozen people, and they do add a whole lot more value to the Twittersphere than that
  • But, since the "redesign", my score dropped from 51 to 37 within one month, where it had been varying between 47 and 54 since February
  • Second, while all individual scores dropped a bit, 5-10% or so, the biggest drop occurred in my socalled Amplification score: it went down from 29 to 10
  • Third, when I want to compose a tweet of my Klout Classification, I don't get @MartijnLinssen, but @adriaanbloem or @rickladd as subject:
Check the red circles: the right one on the top shows my name, the one in the status bar shows Rick's name

Before, it used to show @adriaanbloem: as it turns out, who ever is at the far right of your list of influenced people, will turn up in your Classification score

Try it yourself! Although it doesn't work for everyone, go to Klout, enter any user name, take the Influence Matrix (Klout Classification), move your mouse over the Compose Tweet button and watch the status bar of your browser. People that are having their score recalculated might be showing correct names, but after the recalc it's back to the wrong info
Update 15th of June 9:33 CET:
It's just very erratic behaviour; it comes and goes. In fact, when you refresh the page, it will state the correct name. After a while (?) it will reset back to the wrong one. Somehow the mouse-over will trigger the right value to appear - I know this sounds crazy but that's just what it seems to be. Here's a list of what I just found now, and who are all "back to normal" for the time being. Just pick 5 of their Influenced By people (the idea is that you check out people that don't get checked out that often), and you'll see most of them are off
I reported and explained this in length May 26th - but couldn't reproduce
June 5th I managed to repoduce it, and sent a screenshot - to the same person I reported it to
June 8th, I followed up on that via email - DM's aren't great for bugtracking purposes anyway. I got a confirmation from Klout that they'd received it, would look into it, and get back to me

Call me impatient but after a week I hadn't heard anything, so I checked again and tweeted about it. Word from Klout had it that they still couldn't reproduce (...) so here's my post, trying to explain it once and for all

With regards to the Amplification score: it measures absolute as well as relative ReTweets and Mentions, which is a good idea. But if I count 12 retweets in one week alone, and the total number Klout gives me is 8, it really doesn't help when the answer from Klout sounds: We're a little specific about what retweets we count. That's a 20th century enterprise-answer if you ask me. In fact, that's what they call in England "a polite foff"

With regards to the general Klout score: it's up to 48 now, and although I couldn't get it refreshed yesterday (nor today), 11 extra Klout points in 12 days on the one hand comforts me, and on the other makes me think about the stability and value of it all...

0 reacties:

Post a Comment

Thank you for sharing your thoughts! Copy your comment before signing in...