Why we shouldn’t take social media metrics too seriously

In the last year, I’ve become increasingly preoccupied by why we shouldn’t take social media metrics too seriously. In part, this preoccupation is analytical because following this thread has proven to be a useful way to move from my past focus on individual users of social media to a more expansive sociological account of platforms. The lifecycle of metrics from being a project of platform engineers, through to being a feature of platforms onto something which are meaningful and matter to users elucidates structure and agency as it pertains to platforms. As does the subsequent utilisation of these metrics, laden with meaning by users, in order to model these people and modulate the environment within which they act. By saying we shouldn’t take metrics too seriously, I’m drawing attention to the way they are used as a mechanism to mould the behaviour of users and the risk that uncritical embrace of them leaves us being enticed by platforms in a damaging way.

However beyond this concern, we shouldn’t lose sight of how easily they can be fudged and how unreliable they are. This is a concern which Jaron Lanier powerfully puts forward on pg 67 of his new book:

First, why believe the numbers? As discussed in the previous argument, much of the online world is fake. Fake readers, fake commenters, fake referrals. I note that news sites that are trying to woo advertisers directly often seem to show spectacularly greater numbers of readers for articles about products that might be advertised—like choosing your next gaming machine—than for articles about other topics. This doesn’t mean the site is fudging its numbers. Instead, a manager probably hired a consulting firm that used an algorithm to optimize the choice of metrics services to relate the kind of usage statistics the site could use to attract advertisers. In other words, the site’s owners didn’t consciously fudge, but they kinda-sorta know that their stats are part of a giant fudge cake.

It’s not so much that they are meaningless as that their meaning is often unstable. There are occasions in which it might be necessary to engage with them but we have to do this carefully. One of my projects in the next year will be to try and produce guidelines about this interpretation which reflect what we know about the sociology of platforms while nonetheless recognising that metricising our activity on social media can sometimes serve as strategic purpose.