Klout, an app which scores peoples’ influencer on social media, is set to close on 25 May. That’s the also the deadline for companies to comply with European data privacy rules. Coincidence? Probably not. But whether Klout was or wasn’t selling your data, it failed because it was attempting to distill a person’s entire social media footprint into a single two-digit score.
Sci-Fi author John Scalzi verbalised the concerns of many when he described Klout’s mission as ‘socially evil’, by playing on the anxieties of social media users. But social justice was not the reason for Klout’s downfall. The most important reason is simple: it just didn’t make sense for business.
The problem was that the Klout score didn’t – and doesn’t – mean anything useful: even after the inner working of the algorithm were revealed, and despite apocryphal tales that hiring decisions were made on the basis of applicants’ scores, the score itself has little, if any business value.
For anyone buying or selling influence on social media, the Klout score had little chance of displacing above-surface metrics (e.g. followers, likes and shares) as the go-to metric in influence-based transactions. That’s primarily because – on most platforms – everyone can see followers, likes, and shares. Even when these metrics, or the method used to calculate them, are more obscure, it takes fewer leaps of faith for either buyer or seller to come to an agreement on the basis of prospective views.
The alternative, presented by Klout, was an opaque, cross-platform algorithm, which claimed to measure “influence”, but also determined that Justin Bieber was more influential than Barack Obama. If, to any Beliebers reading this article, that isn’t a cause for concern, consider that Klout also rates Justin as an expert in “religion”. That may come as a surprise to Pope Francis, an expert on the topic of “Miss Universe”.
Even if you thought it was laudable to measure social media users’ influence, plenty of people claimed that it was easy to game the system, and in 2011 somebody worked out that a person’s number of followers was sufficient to explain 95% of the variance in their Klout scores. So, Klout or no Klout, it still came down to number of followers.
The negativity surrounding Klout is not a reason to view all social media scores and metrics with suspicion. In a business context, engagement based metrics are far more useful, despite being “below the surface”.
For example, it is common that, the larger an influencers’ audience, the less engagement they get per post. If, for example, you’re able to compare click through rates of a series of influencers, you may discover that using ten smaller, highly-engaged influencers may be both more effective and cheaper than using one influencer whose audience is bigger than all ten combined.
Equally, a score based on engagement would allow influencers who have highly engaged audiences to get paid higher than those with a big following, which may be comprise a higher proportion of bots, and fake accounts.
The mistake that Klout made was to reduce an individual’s entire “impact” on social media into a single score. In influencer marketing, specific scores may be beneficial in specific contexts. But one size certainly doesn’t fit all.
Bret Cameron is head of campaigns at Fanbytes