the matter with metrics

Twitter is my personal canary in the coal mine of world events.

A coup? An outrage? A celebrity death? I miss nothing. Why, I have mourned the loss of leading figures before they themselves even heard they were dead (sorry ’bout that, Gordon Lightfoot.)

Yesterday, I heard the wailing and gnashing of teeth as soon as I opened my laptop after lunch.

Sometime around noon, Klout’s algorithm shifted. And revealed a great deal about itself – and us – in the process.

Klout defines influence as “the ability to drive action.” Klout claims to measure influence across social media platforms. It collects data on users’ engagement on Twitter, FB, G+, Flickr, etc., and collates those multiple analytics into a single, shifting number. You go up if you’re doing well, down if you’re losing influence. Or, say, if you spent a whole day offline. Merciful heavens.

Klout has been embraced as an objective third-party tool for business to tell which self-promoting social media gurus actually have real capacity and reach. It has also been embraced a pet hobby for bloggers intent on giving each other mischevious +K points on topics like “belching,” “Kansas City airports,” and “hairy backs.” It promotes that use less loudly in its press releases.

Klout claims to measure both reach – how many people you influence – and scale – how much you influence them. It also takes into account the influence of those you influence. Meaning, on the surface, if you engage with leaders in your community or corner of teh internets, you yourself are more likely to exert leadership influence.

If you’ve been in the habit of checking your Klout, you may have seen a change in your score yesterday. And if you had Klout anywhere above, oh, 55 or so, you may have seen a drop. Klout posted a graphic (scroll down here) to support their claim that the majority of users would see their score stay the same or go down, but a straw poll of the canaries tweeting out sturm und drang on my Twitter feed yesterday afternoon suggests that the people clipped hardest by the new algorithm were the ones best positioned to actually give a shit about Klout.

(Disclosure: I went from an all-time high of 64 to a 57. Pass the hankies.)

Last week I ended an academic presentation of social media with a screen capture of my Klout score at the time, tongue-in-cheek. Thank god. I’ll never see that number again.

But, as I noted on Twitter, showing it off to a non-social-media-using audience isn’t a whole lot different than bragging to them about that high score I got in Super Mario Brothers back in 1993. It, too, was still a lot lower than some friends’ scores. It was higher than others. What it gave me was a sense I was improving at a game I was trying to learn…which is pretty much what I think Klout is good for.

(Admittedly, the old algorithm could be gamed, and was skewed by random RTs by celebrities, for instance. It rewarded cliqueishness, and highly sociable people with access to established networks. However, while the new Klout claims to be more transparent, I don’t actually see the explanations of how my acts translate into data anywhere in my new Klout interface. I’d like to: for my thesis research, it’d be fascinating.)

But. The lack of transparency, however touted, is not the problem with Klout’s new algorithm.

Maybe Klout needs to become my new canary in the coal mine of social media. Because the problem is bigger than Klout, and it is threefold.

1. We are beginning to buy into what we think our Klout tells us about ourselves.

Social media practices are identity practices, particularly on networking platforms like Twitter and Facebook (the prime sources of Klout data). Many of us put a lot of time into social media, and are aware that our expertise has growing cultural capital. People have learned to care about their Klout. For some, it’s a very real calling card for very real money. For others, it’s one of the few reflections available of whether we’re succeeding in a varied game in which there are no maps. Even if we watch it tongue-in-cheek, clearly a lot of us watch it.

This accords power and veracity to the metric.

Now, social media has always involved metrics: comments, Technorati, numbers of Twitter followers. But, for the most part, if one desired to increase those numbers, the path was relatively straightforward: one engaged more. Klout, with its complex algorithm drawn from big data that judges our most mundane interactions, is different. It’s not only measuring us, it’s assessing us. It’s designed on behaviourist principles, with rewards and virtual pats on the head when we – ratlike, often not entirely sure what we did to warrant the praise – succeed on the terms its algorithm values, and framing losses in score with banners that proclaim “Oh no! Your life is over Klout has fallen -1 in the past 2 days!”

We are highly conditionable beings. Klout is conditioning us to care about Klout, and to value ourselves – in this identity economy of social media – in terms of it. Which one could argue we’ve been doing ever since 2001, when Joe next door got more blog comments than we did and we cried in our beer and felt small and alone, but. It’s not the same. Because not all engagement is created equal, in Klout.

2. We’re being influenced by our own “influence.”

Used to be, if you happened to be someone who valued the metric of comments, a comment was more or less a comment. Yes, a comment from a blogger of known scale could feel somewhat a visitation from the archangels, but at the end of the day, the comments added up and like votes, each counted (if you were counting).

Relationally, the comment from a famous blogger might be an avenue to connection and networks that might ultimately serve some strategic purpose, but use value can’t really drive relationships in a one person-one vote economy.

Klout, though, works to devalue the nature of many social media communities, particularly those whose networks and relationships aren’t based entirely in use value. Some animals are more equal than others. In new Klout, I now get notices along the bottom of my screen about which contacts have gone down in score recently: in case I want to dump them, I assume, like dead weight. Bye, Mom! It’s all business.

Social media wasn’t supposed to be all business, especially business as usual. Social media is, uh, social. And relational: it’s a form of augmented reality, a network for all sorts of purposes, well beyond use-value networking.

But because Klout rewards use-value networking over other forms of engagement, it fosters an increasingly use-value environment. In Klout, it matters a lot more if you get a famous person to click your link or RT your content, especially if that person doesn’t regularly engage in clicking or RTing or sharing or whatnot. This makes some sense, in terms of assessing influence. But IT ALSO AFFECTS BEHAVIOURS.

The peer-to-peer relationality of social media – already grappling with a relatively new breed of user whose sole goal is building platform as a path to old guard institutional or corporate success – is undermined by the kind of behaviour that cultivates status over relationships. Status is part of the game. But when it becomes the whole game, the broad, rhizomatic networks get boxed in and wither, and then we’re back to something a lot less interesting than social media. And like the new Google Reader, a lot less social.

Yes, there is a pattern here. We are gradually being directed away from sociality and towards business-like behaviours by the business interests that design and profit from the platforms we use.

Social media, which was once a bit of a rogue blowing smoke at the establishment, is being taken in hand and given a tie and a briefcase. We’re like the rebel who’s been told s/he got the highest mark on a class test: we suddenly don’t know what to do with ourselves.

The problem: the test was rigged. And will always be rigged.

3. We’re allowing a metric to do a human’s job.

I’m not saying Klout isn’t trying, in terms of assessing influence and engagement fairly. The problem is, it can’t.

Klout today claims that I am as influential as Her Bad Mother. HA. Klout also puts me two points ahead of Finslippy and three ahead of George Siemens. If I buy that, I’m ON SOMETHING.

My influence and reach and social media fame and probably my throw to third base are all somewhat more modest than those three. My Klout score ultimately reflects that I’m frittering away more time on Twitter than they are, as they’re too busy with jobs or book tours or speaking engagements.

Because their actual influence – their name recognition within their respective fields, their public profile, their contacts, their capacity to leverage social media influence into dollars – is, in each case, greater than mine. That doesn’t negate mine, or anything. But just because Klout says I have equal influence doesn’t make it so.

Klout attempts to create an objective representation of something that is complex and subjective beyond the capacity of any algorithm to capture.

It appears that a lot of business interests have bought into the idea of Klout as a marvellous, miraculous objective third party observer, collating all the variables and doing the dirty work of sorting out for them who matters.  But just because scoring is helpful in a competitive neoliberal economy – “crucial,” even, according to the author linked above – doesn’t mean it’s actually valid. Or even possible.

All algorithms and metrics are products of their design. They are rigid, no matter how flexible and complex, and they cannot make exceptions or comprehend the subtleties of human relational interaction based solely on numbers, no matter how many numbers they use.

Influence is a relational measurement. It is a human measurement. Like intelligence and learning all the other things we stupidly insist we can measure, simple because we NEED effective comparisons, influence exceeds our grasp.

We may need to understand how to compare apples and oranges. It doesn’t mean we can, especially with mere numbers. This is true in education, and this is true in human relations and influence.

And while the game of seeing how we measure up may be entertaining, it’s only valuable if one is embedded enough in the relational networks it claims to assess to know when to take it with a grain of salt. Liz Gumbinner at Mom 101 wrote an exceptional post about this last month, giving thanks for savvy PR people and corporations who recognize good writing when they see it, who understand that this game is more than numbers.

I’d like to see more of them. I don’t wish my Klout canary in the coal mine of social media dead, but I’d like it seen for what it is: a decorative little bird, useful for entertaining and reflecting back the notes one is, uh, tweeting. NOT the measure of value in social media.

We need to stop handing over so much power to metrics. They have a place. But it’s THEIR use-value we need to assess, not the other way around.