Sometime last year — I don’t remember when exactly — a box arrived at my apartment. Inside was a handful of bags of Popchips. I ate them, then didn’t think much about them again.
By doing so I became part of an “influencer” program. I got the chips apparently on account of the sway I hold when it comes to snack food as part of the Klout Perks program. Like many regular Twitter users, I’d used Klout to gauge my social media reach on Twitter. I’m apparently a 61. The question that nagged me was why exactly was I an influencer when it came to snacks?
Klout is trying to solve a thorny problem. Marketing has long sought to reach tastemakers and influencers. The process, even as it is done today, is frustratingly handcrafted. Marketing agencies will often tout their own networks of influencers that they can “activate” on behalf of clients. The truth is there’s little science behind this. Now, thanks to all that sharing on social media, there is, in theory, enough data to crunch to determine who is truly influential.
The rub is it’s still quite early in this process. The Klout score is the most prominent example. On its face, there is a problem: how can I have an overall score? I’m sure I have a bit more influence when it comes to, say, digital media and marketing than snack food. Klout couldn’t tell me exactly why I was chosen for the Popchips program. A rep noted that I did joke about ketchup once recently; and I do have a passing reference to cheeseburgers in my bio.
“The idea that someone with a score of 100 is more influential than someone with 60 under all circumstances is obviously suspicious,” said Duncan Watts, a social media researcher at Yahoo and the author of Everything Is Obvious
. “The right question is who is influencer about what. Nobody is influencer all the time with everything.”
That’s a problem. Klout is aiming to be the Google Pagerank of the social influence world, at least according to Megan Berry, a marketing manager there. She noted that Klout can pull in data from Facebook, if people register, and will soon add LinkedIn. Now that we’re sharing more and more online, Klout could amass a large trove of social data to crunch.
“You’re going to see this grow a lot more as more people are aware this exists,” she said. “It’s going to start mattering in ways people don’t realize.”
Already it matters in small ways. The Palms Casino in Las Vegas determines upgrades based on Klout scores. There are even some intrepid souls touting their Klout on resumes. The scores are used, in part, to determine participants in the Klout Perks programs. It has run influencer campaigns for Audi and the like. The campaigns run the gamut from as small as 200 to up to 10,000 participants. Barry notes that marketing agencies typically struggle to scale such programs with their unscientific approaches.
The issue for brands is predicting who will be an influencer for which campaigns. It is always a situation of starting from scratch, said Noah Mallin, group director social media at Digitas.
“What we find is it’s going to be different from brand to brand,” he said. “An action for one brand that might show a high affinity might not be important for another brand.”
The risk is the science just isn’t there yet. To Klout’s credit, it does have real science muscle behind the problem. One engineer is trained in applied learning and pattern recognition. But it is early days for these efforts. The Klout score is based on how people react to messages — whether they reply, retweet and their influence scores.
“The problem is none of them are actually about influence,” said Watts. “What people care about is if they can activate or somehow engage a set of people to some outcome. They never measure that.”
Watts notes this isn’t that new. Eli Roper was working on this problem with his surveys over 50 years ago. Ed Keller has worked on this for a long time, too. The Klout approach is a step forward from people who are self-styled influencer experts without any data, Watts noted.
“At least they have a definition,” he said. “At least they’re taking it seriously enough to come up with a number somewhere.”