Following my review of two important books in Social Media analytics in the first issue, I wanted to take more time to discuss the concept of influencer, one idea that certainly is at the center of Social Media Marketing. Why? Well, simply because marketers believe they can reap benefits through influencing influencers.
I also believe that the focus on influencers, and thus all the efforts put in understanding how influence works in Social Media, comes from the acknowledgement (or absence of acknowledgement, strangely enough) that, for most companies, Social Media audiences are by far much smaller than their client base, or potential base. In short, since we are basically not reaching a significant number of people, those we do must have a higher value. Using the 800 million Facebook member number as a potential market is as moronic as saying there are 2 billion Internet users; there may be, but they’re not on my page/site!
Since this is not a dissertation, I cannot review all the definitions of influence out there. Suffice to say here that influence is the capacity, or propensity, of having others change their mind, and/or do something. It can’t really get simpler than this. Since measuring how what someone said convinced someone else is rather very hard to do on a large scale, influence in Social Media is basically measured as behavior. For data collection, and automated analysis purposes, the behavioral dimension of influence is the more practical one to adopt; at the end of the day, don’t we rather want people to do something (buy), than have nice thoughts about us (although the latter can be connected to the former)?
In most Social Media Analytics products so far, influence has been established by the size of someone’s audience, and the number of times their contents (Tweets, posts, videos, etc.) get relayed through their audience member networks. Inevitably, audience size seems to be an obvious criterion, since scope of reach, mainly how many people can be exposed to a message, must be an indicator of one’s influence. Which also means that celebrity would play an important role here, since, to my knowledge, very, very few people have organically become famous, and gathered enormous following, thanks only to what they did on social platforms.
However, audience size, in itself, does not seem to guarantee influence defined as making people act. N. Christakis and J. Fowler (authors of the excellent Connected) gave a great example of audience size efficiency in a very good article; they sold more books from a tweet by a person with 4,000 followers than from a tweet by a celebrity with over a million. True, the data set is extremely small, but it offers a good illustration of how differences in audience composition can have an impact. Famous blogger Anil Dash also commented on how suddenly going from 18,000 followers on Twitter to 300,000 didn’t actually change the number of retweets and replies he was getting. And Kevin Hillstrom tweeted some time ago: “Just as many people click through my links today at 2,188 followers as 18 months ago at 300 followers”. If we are to believe these three examples are representative of anything, then audience size is probably more a measure of someone’s celebrity than actual influence, again here seen as the propensity to make people do something.
Obviously, measuring influence via people’s actions outside the social platform can be quite a complex and daunting task. Not surprisingly, what products will look at is how many retweets, mentions, likes, comments, grades, scores, etc., one gets. The more of those one gets from one’s audience, the more influential one is. We could discuss endlessly about the significance of each type of action. How “influenced” am I when I retweet something, or is it me who want to influence? Could be either. If I publish a funny video on my Facebook page, and it gets “liked” a lot, is that manifest of my influence in any way?
We could also debate about the intrinsic costs of “liking”, “retweeting”, “sharing”, etc. Social platforms have made those actions so frictionless these days that one wonders how much influential one must have to be to make you do them. Of course, there is an immense gap between making someone retweet one of my tweets, and make that same person buy something. Still, this is via those sharing functionalities that products will establish if someone is influential.
So far, I think we are still very much in the broadcast paradigm when we think about influence. This is very much about how a message source broadcasts to the masses, while being relayed through “antennas”. A one-to-many paradigm.
But what is really social about it? Do we really understand social dynamics by parsing social platforms users through hierarchies of retweets, likes, share numbers? And if one looks at one’s audience, how much of a social network is it? Is it a strong, or a weak one? Are my followers/fans connected beyond just following me, or are they themselves connected to one another? Who are the most important nodes in that network? Me who send out messages, or those who relay them? From a social network point of view, who is the most important: the Retweeter or the Retweetee? A post by Steve Boyd (found in Marshall Sponder’s book who very lightly touched the concept, but unfortunately didn’t develop it further) about “betweeness” is a good example of how we could look at the notion of influencer.
As for me, I don’t have any answer for now. Coming from the social sciences (Sociology 1985), I tend to lean toward the corpus of social theory knowledge for better, more insightful explanations. Unfortunately, it hasn’t permeated yet into analytics products the marketers could use.
So, hold your horses; I am not convinced anyone really can deeply analyze social media yet. We are still at the page views, hits, unique visitor stage. I believe in two, three years’ time, a lot of progress will have been made in social graph and analytics, allowing us to discover the unsung social heroes currently eclipsed by the shining stars.