What Can My Lack of Expertise in Waffle House Tell Us About Lazy Metrics?
As many of you know, I’m writing a book about metrics for John Wiley and Sons Publishing called Social Media Metrics for Dummies. What you may not know is the struggle I’ve had trying to figure out how to address the issue of services offering incomplete metrics that become some kind of standard for folks too lazy (or, to be fair, too pressed for time) to do much deep-diving on their own. Things like Klout, for example.
Now, don’t get me wrong. I know what Klout is trying to do, I think. And they have adjusted their algorithm a number of times. My score was a 79 in the beginning (until I spent less time online, when it dropped to 72). It hovered at 69 through several more algorithm changes until it now is resting in the 56 or 57 range after the latest one. What bugs me about Klout is that it isn’t a complete metric. It doesn’t tell you anything about the person except, conceivably, how noisy they are on various services.
I’ll use myself as an example. This morning I finally “hid” a topic that Klout has dubbed me as “influential” in for two+ months now: Waffle House. I like Waffle House and all, but I don’t live near one, and I don’t go visit them when I travel. How did I get “influential” about it? Doing a little digging, it seems it began when I shared a tweet from a WSJ article several weeks ago about the Waffle House disaster plan and Waffle House trailer offices – how they handle natural disasters and get their stores open quickly with limited menus where possible to be a local source of food and comfort, and also how they use their online presence to spread the word.
By a sheer coincidence of timing, that tweet got re-tweeted a gazillion times and picked up on several Tumblr blogs*, entering the Tumblr share network. Shortly after that, it emerged as a topic I’m influential on in Klout, never having discussed it before or since. Now you tell me: is that an accurate measurement? I should be influential in a variety of things, like hockey, music, food, wine, football, mma, film, politics and other things I discuss and have deep conversations about frequently (none of which show up) or in social media, content marketing and emerging media (which do show up).
*To me, my “influence” in Waffle House was a better measurement of TUMBLR’s influence than mine. I don’t have a Tumblr blog, but those who did caused most of the impact. To not take into account deep conversation and conversions over noise indicates a failing on the part of Klout and narrow metrics like it. I’m sure they are working to address it – it’s plain they want to be the go-to metric source for measuring influence – but they have a long way to go (not to mention other problems with privacy and trust issues and some nefarious practices to solve first).
In the book, Klout gets a relatively positive mention, but with the double caveat of “use with caution” and “not intended as your sole metrics solution”. If the folks at Klout would like to have a conversation with me about this and discuss case studies or what they are working toward, I’d love to. I like to keep an open mind.
Meanwhile, if you want to put understanding your Klout on steroids, have a look at this nifty data set from always-insightful Chris Penn. It will rock your socks. Also, I’d love to hear what goofy thing Klout thinks you are influential in.
Aside: Single focus metrics options that show a more complete picture could include Smarterer, by the way. I really like where they are headed and hope they soon integrate with LinkedIn and other services. Disclosure: I wrote the bulk of their Twitter test and edited it during their private beta phase, though it’s now open to public edit.