Update: Klout seems to be listening. On December 2, CEO Joe Fernandez posted “What Does Klout Measure?” on Klout’s blog. You can find the link below.
At the end of October, Klout changed its standard for measuring influence. The Klout score is a 100-point scoring system that gives marketers, employers and users of social media the opportunity to surmise from a single number someone’s ability to drive action on the social Web. Recently, Klout changed the way it arrives at this score, resulting in dramatic changes for many of its users. Some scores went up while others went down.
Originally established as a scoring system for influence on Twitter, Klout now measures influence on a dozen social networks, including Facebook, Foursquare, LinkedIn, and Google+. More networks are being added, and the company’s CEO Joe Fernandez predicts the total will reach 20 by the end of 2011.
A person’s Klout score is based on three subscores: True Reach, Amplification and Network. How these scores are calculated – Klout’s PeopleRank algorithm – is Klout’s secret sauce, and on October 26 the ingredients changed. With the additional number of networks and the increase in data – over 2.7 billion pieces each day – the company realized it had to change its algorithm so that scores would more accurately and transparently reflect a person’s entire social influence.
Earlier in the year Klout had stated that the addition of networks to a person’s profile would create an “additive affect” and would not detract from one’s total score. But overnight, Klout scores changed dramatically. Many influencers saw large drops, leaving them wondering how Klout could have gotten it so wrong before. Mine dropped 30%, from 70 to 49. My true reach went from 10K to 2K, my Amplification went from 53 to 5, and my Network went from 71 to 18.
Since its earliest, Twitter-only days, I’ve been intrigued by the Klout system of scoring and have extolled its utility to my clients, and so the abrupt change in Klout’s “standard” came as quite a surprise. There was an outcry throughout the Klout ecosystem. I joined in as journalists, social media pundits and practitioners reached out to Klout for an explanation. We were referred to the Klout blog, which offered only a palliative post, at the end of which I expected to see a smiley face. Why would a company striving for more transparency not be more transparent with its users?
I began researching influence online, and very quickly one thing became resoundingly clear: influence is what’s wrong with social media.
The Larger Context of Influence
“Our friendships and professional connections have moved online, making influence measurable for the first time in history,” reads the first line of Klout’s About page. Klout would have us believe that our online connections represent the sum total of the measureable population of people who influence us, and that for the first time in history what constitutes influence in our lives is, indeed, measurable. If only hubris were measurable.
Quite simply, only a fraction of influence in social media is currently measured. Influence can only be approximated based on the very individualized uses – by millions of people – of hundreds of social media channels. Klout measures about a dozen channels and then distills that measurement into a single score.
But the real issue isn’t how or whether we’re able to measure influence. What concerns me is that social media influence is following the pattern of other forms of influence that have had pivotal effects recently on our society.
I don’t occupy anything these days, except perhaps my office. But I share the growing concern that America has done little to curtail the influence of our financial industry over Washington. I bring it up to illustrate a point about influence. It was influence, for instance, that facilitated the 1999 enactment of the Gramm-Leach-Bliley Act, a law that repealed the restrictions of the Glass-Steagall Act on commercial bank participation in investment banking activities. Influence again kept the Commodity Futures Trading Commission from regulating derivatives. The result was the creation of the financial industry’s securitization conveyor belt that led to the mortgage lending crisis, which has pushed about 50 million Americans to the poverty level while creating a top 1% that in 2007 took in over 23% of the nation’s income, conditions we haven’t seen since 1928, just before the Great Depression.
Could an online class system form from social media influence? When we look at the evolution of social media influence, we begin to see distinct divisions.
During Social Media 1.0, some of our existing offline relationships moved onto the Web. (It would be more accurate to say bits and pieces of these relationships were copied over to the Web.) We connected with some of these friends on networks like MySpace, Facebook, LinkedIn and Twitter. And then during Social Media 1.5, we expanded our social graphs by adding people we hadn’t seen or heard from in years – old friends from high school, grade school, our childhood neighbors – people we hadn’t thought about much over the years. We even began adding people to our social networks based on relationships we had in our existing graphs. Friends of friends became first connections.
During Social Media 2.0 we began seeking even more friends, fans and followers, taking advantage of weak connections to build out our social graphs and engage larger and more influential numbers of people. While social media theorists and practitioners maintain that quality of engagement is more important than quantity of connections, none will dispute that there is a positive correlation between influence and number of friends, fans and followers. Size matters.
Klout currently scores over 100 million people, most of whom have not signed up for a Klout account. The company uses information from individuals’ public channels like Twitter and private channels like Facebook to build its data. More than 3,500 companies use this data to reach influencers with targeted marketing. In May, 2011, Fernandez told Forbes reporter Kym McNicholas that certain hotels in Las Vegas look up your Klout score when you check in to see how influential you are so that they may determine whether you should receive preferential treatment. The better they treat you, the hotels reason (if you have influence), the more likely you are to rave about your experience to your online community.
In social media, anyone with a small number of friends – at any level of engagement – does not have absolute influence. They can’t check in to a hotel in Las Vegas and be treated with privilege. Their numbers don’t cut it in Vegas.
And so at the end of Social Media 2.0, we face a new greed, one with the same unpleasant characteristics and ill effects of financial greed.
Not that influence is a bad thing, particularly when it’s formed and used in the aggregate. Even instances of individual ability to drive positive change are favorable when that change sincerely reflects the desires of many. But influence in social media is being co-opted by the second-comers to social media – marketers – whose interests aren’t necessarily those of the community.
An individual becomes an influencer in social media because her interests are well aligned with those of her online community. An accretion of friends, fans and followers occurs around this influencer, in some cases exponentially, as word spreads of the value offered by this individual. Marketers identify such influencers and engage them on various social media in order to reach their communities with a message engineered to slip seamlessly into the conversation. No harm, no foul, particularly if the marketer’s interests are genuinely aligned with those of the community.
But as gamification and award-based practices (like those of the hotels in Vegas) become rife in social media, the behavior of the influencers changes through incentive. Badges, mayorships, leaderboards and free orders of fries can cause someone with influence to act less in the interests of their friends, fans and followers and more in the interests of the marketers offering these rewards.
As badges build and connections accumulate and Klout scores climb, and as more and more people buy into the idea of engaging influencers to build their own influence, the ecosystem skews toward the interests of the few. And so the interests we find at the top of the influence pyramid once again are out of alignment with the interests of the many. Influence at one time may have been organic on the social Web, but no longer. As with a lot of things these days, social media influence is for sale.
Where’s the social value in engaging only with influencers? Where’s the responsibility? A representative from Klout offered insight into Klout’s point of view when she answered my email questioning the drop in my score:
The reason you may have seen a drop is that we’ve improved our ability to take into account how *much* you influence someone. For instance, if I rarely like or comment on anyone’s posts, but choose to do so to yours, that is more meaningful than if I like 60 posts a day. For users whose engagement primarily comes from others with high activity, you may see a drop in your Score. [sic]
What part of the secret sauce is it that tastes really bad here? If I’m in the middle of an Arab revolution, and I retweet updates at a rate of 60 a day or I like 60 related posts a day, will that reduce influence? Do I care? I just want regime change.
Hotel rooms and French fries are one thing, or shopping for shoes online, or asking where to rent a bicycle for the day. But what if I want to learn about an emerging presidential candidate in a party other than my own? If I begin asking around in various social media, I’m led to destinations online that have been Liked, commented on, retweeted or otherwise shared by people in my social graph. My results are skewed because of our aligned interests – we’re members of the same political party, for instance. The results are further skewed if the activity of my social graph includes Likes, retweets, comments or shared objects that have been strategically inserted into their online conversations by influencers wishing to spread political messages that unfavorably portray this opposing candidate.
I wonder if Klout will ever see the need for a new badge: Trusted Objective Observer. Or if they’ll ever come up with a score that indicates how un-influenced a user is. Perhaps that score already exists, and that at some point soon instead of striving for a higher Klout score, we should strive for a score of 1.
Update: Klout’s reaction to these and similar concerns: “What Does Klout Measure?”