In a recent conversation about social media influence, the topic of influence scores came up as a way of measuring how influential someone is. Various influence measures are available to us for identifying who we should do outreach to, including Klout, Kred, PeerIndex, Moz Influence Score, and many new entrants.
One of the key questions you need to ask is whether you know what goes into any influence score, what the ingredients are. If you don’t know what goes into an algorithm, you can’t know what results it’s delivering for you. Edison Research VP of Strategy Tom Webster pithily defines an algorithm as “data + assumptions”, and it’s the latter half, the assumptions, that cause influence score problems.
What assumptions went into the influence score of your choice? Does it rely on things like follower counts? Engagement of content? Reshares and retweets? Unless you know, you don’t have an understanding of what kind of influence you’re targeting. Influence scores based on audience size (followers, etc.) can mean a marketing program based heavily on reach, but not engagement. Influence scores based on sharing can mean distribution, but not necessarily audience size, so if your program goal is broad reach, that sort of influence score won’t work for you.
If you don’t know what goes into the influence score you’re using, you have two options. First, switch to one that does disclose its algorithm. This isn’t always a realistic choice, however, as many of them promote their “proprietary algorithm”. The second option is to do a blended index of influence scores, to smooth out the differences in individual algorithms. This is famously how Nate Silver called the 2012 elections, by averaging together many different polling data sets.
Whatever approach you take, know what kind of influence you’re measuring before you built a strategy and program on top of it!
Christopher S. Penn
Vice President, Marketing Strategy