A little while ago, there was a lot of chatter in the blogosphere when the CEO of Klout, a social media analytical tool which attempts to measure ‘influence’, tweeted a Salesforce job post that required a Klout score of 35 or above. This kicked off a debate about whether a Klout score is a useful requirement when recruiting. Social media has revolutionised the way that recruiters work, so is Klout the next step in gaining an edge when sourcing the best talent?

For those who have never heard of Klout, it is one of a new breed of companies that uses the big data available on social media and attempts to analyse it to deliver insights that may not be immediately obvious. In this case, Klout attempts to work out how much ‘influence’ you have online. To do this, you register your social media accounts to the application (e.g. Twitter, LinkedIn and Facebook) and it then applies a complicated algorithm to give you an influence score of between 1 and 100 – the higher the score, the more influential you are. Klout isn’t the only company doing this, Kred is another company with a similar product.

At first this sounds great from a recruiter’s perspective, as it is another tool to give you insight into the quality of your candidates. After all, who wouldn’t want a candidate with a high level of influence? Everyone knows how important influence is in your professional life; one of the best selling books of the 20th century was Dale Carnegie’s How to Win Friends and Influence People. An influential candidate must be a good candidate.

Unfortunately, there are several things wrong with this picture. Frankly, the idea of being able to quantify something as qualitative as influence is laughable – influence as a concept is about as nebulous as a sparrow’s fart. To create a score of influence makes several false assumptions:

  1. Influence is a linear scale (i.e., a score of 40 is a consistent amount better than a score of 20). This dramatically oversimplifies the nature of the world and human interactions. This assumption has to be made because otherwise the score would be meaningless.
  2. Influence is objective. This assumes that we will all be influenced to the same extent by someone. Influence is about as subjective as you can get – Justin Bieber has an incredibly high Klout score (the first person to briefly achieve a score of 100), but I couldn’t give two hoots about what the annoying Canadian teenager is up to. To one of his fans however, his every word is gospel.
  3. Influence can be accurately measured online. This assumes that as social media becomes more and more ubiquitous, our lives will increasingly be lived online. However, the average time we spend on social media each month is measured in hours – the rest of the time, we are out in ‘meatspace’, interacting in the real world. What happens in the real world is far more important than what happens online, and that is where real influence is to be found. What we do on social media may have no bearing on our professional and real-world selves: one software developer I know only uses his Twitter account to post bad puns.

Even ignoring those critical errors in assumption that Klout makes, there is one big reason why you should avoid using Klout: we have no idea how its algorithm works. We do have some details about how Klout works – for example on Twitter it measures how many people are talking about you, what you talk about and what gets retweeted. But, for the most part, we have no idea what calculations go on to arrive at that score. My most retweeted tweet ever, for example, was a story about my wife playing a dirty word in a game of Scrabble with my gran and mother to get rid of her ‘Q’ tile. It was retweeted by Neil Gaiman (a very influential figure on Twitter) and received dozens of retweets and dozens of replies. Unfortunately, that tweet has nothing to do with my professional experience (recruitment or management studies). Klout, however, may well assume that I am therefore most influential in the fields of Scrabble and [dirty word].

Without being able to see Klout’s algorithm and critically appraise it, any number is completely arbitrary and, as a result, meaningless. There are stories of Twitterbots being given high Klout scores despite only being an automated feed. And, despite the fact that we don’t know exactly how the algorithm has worked, a number of people have come up with various ways to ‘game’ the system and artificially inflate their Klout scores. If it can be manipulated that easily, then it loses whatever shred of credibility it had left.

Network theory, the academic subject that studies how humans interact with one another, is incredibly complex. In networks, all the vagary, irrationality and chaos of human nature is multiplied and refracted in our interactions with one another. People have been studying it for years and we are still yet to come up with a consistent model of human networking behaviour. Yes, there are patterns that we can pick out, but there is no mathematically consistent model to predict human behaviour in networks to a reasonable degree of precision and accuracy. The idea that a company in California has managed to do in a few brief years what decades of some of the world’s best academics have failed to do is something I find very hard to swallow.

So, for the recruiter wondering whether to start using Klout as a metric as a recruitment tool, please think again. It isn’t going to make your job easier or improve your candidate quality.

Andrew Fairley

Andrew Fairley has recently completed an MA in Management with The York Management School, focusing on strategy, innovation, HR, and organisational behaviour, and has just begun a PhD investigating the UK internet startup industry. Prior to this, he spent 2 years as a Recruitment Consultant, working with clients from SMEs to blue-chips, sourcing IT staff.  You can find him on Twitter or LinkedIn.