31 Jan Quantifying the Impact of Social Media: Where the Edelman White Paper Got it Right, Got it Wrong and What We Should Do Next
Earlier this month, Jonny Bentwood of Edelman published “Distributed Influence: Quantifying the Impact of Social Media.” Edelman gets picked on a lot (often deservedly so) for the highly visible mistakes they’ve made in the social media space, but there’s no question that they are both deeply involved and pretty darn thoughtful on the topic. This white paper fits the latter; it was a thoughtful analysis by some major players in the game on how you measure online influence.
What they got right:
1. This white paper builds on the concept of a Social Media Index, which is an effort to get all of us in this space to think about a standard way of measuring of influence. The index is not perfect (as they openly admit), but it is thought provoking. It’s a good starting point and a must read.
2. They gave some additional insight into how meme’s get started, concluding that there are:
- meme starters,
- meme spreaders,
- meme adapters,
- meme commentators, and
- meme readers.
This is analogous to Forrester’s Social Technographics Ladder, which measures web participation from low to high as:
- Inactives (52%)—>
- Spectators (33%)—>
- Joiners (19%)—>
- Collectors (15%)—>
- Critics (19%)—>
- Creators (13%).
3. They begin an interesting discussion that influence as measured by the power of an individual isn’t the way to go, but by the power of the idea and how much it actively moves within the space. Credit to Jeff Jarvis for that thinking, which I think is 100% right. This is similar to what Radian6 does in their monitoring tool when they allow you to search content by the activity surrounding it (i.e., total comments, unique commenters, etc.). I firmly believe that this is likely to be more useful than chasing an amorphous, ever-changing group of influencers for reasons I explain below.
What they got wrong:
Early on Bentwood acknowledges that this white paper doesn’t even try to be the definitive answer we’ve all been looking for, but another piece of the conversation. It’s a valuable one at that. These “got it wrong” statements are meant in the same way: to contribute another series of thoughts to this conversation, not to criticize even in a small way. So here goes:
1. For the bulk of this white paper, the roundtable seems to conclude that influencers are influencers by their very nature. I don’t believe this is true.
Relationships (both online and off) are contextual. A powerful influencer in one situation is a meek influencee (or someone who simply doesn’t care) in another situation. Steve Rubel has influence in social media conversations, but likely has no influence in conversations regarding women’s shoes, vitamins, Bollywood, etc. (Steve, if you are, in fact, a player in the women’s shoe world, more power to you, but hopefully the point is well made. 😉
2. Social networks and trend starters are too complicated to ever be identified in advance.
That “influencers” exist and “do their thing” is the premise of the Tipping Point, a favorite book of mine. But we all know (or hopefully should know) that the world is vastly more complicated than that. Duncan Watts of Yahoo has done some groundbreaking research on social network behavior that suggests that the power of the idea is in the context of the moment, and the influencer on a given idea varies widely and unpredictably. As a result, we should not be trying to pin down the “powerful people” as some fixed set (n=what?), but rather build better systems to help us quickly identify powerful people by topic.
3. Personality-type is not the horse we should be betting on.
According to Bentwood, the roundtable concluded that “a system equivalent to Myers Briggs was needed for micro-communications. This would enable people to be able to map target media, meme creation, consumption and sharing habits.” In my view, that sort of effort would miss the boat entirely and be of dubious value. As someone who works in a multi-client social media agency, I have to work within conversations by topic. I believe that people’s passions and influence are not innate qualities of the individual that make them either an influencer or an influencible, but rather vary widely based on their comfort in and knowledge of a particular topic relative to others in the discussion. (Think of the bossy know-it-all around the office who is meek and mild when his big brother is around.)
Whether they ever choose to participate may relate to their personality type, but among those willing to participate, influence-level is not fixed. Max Kalehoff’s statement in the white paper that, “You don’t go for the most influential but the most easily influenced” is interesting, but misdirected and sounds suspiciously like targeting 1.0. (I acknowledge that I don’t know the context of Max’s statement, so I may have missed some nuance he injected.)
What we should do about it:
Based on all this, what do we do about it? What’s the ideal situation for moving forward? Here are my thoughts to add to the conversation:
1) To move the social media index forward, we’ll need a system that serves as the white pages of social media with RSS.
In other words, one of us needs to build an open directory into which people can voluntarily put their various online personas (blog, Facebook ID, Flickr account, LinkedIn, Twitter name, Utterz account, etc. etc. etc.). There’s no question that everyone would not do this, but social media types are openly creating online personalities and want to be found, so active “creators” users would. While Naymz and a few others do something similar to this, the utility comes not from just a “look-up” feature, but the ability to use this data to analyze contribution levels. (The social media index is interesting, but to make it work in the long-tail we have to be able to both automate it and use it as a tool within a monitoring platform.) So we’ve got to be able to allow folks to query this data in an automated fashion as part of our monitoring.
2) With the dataset we’ve created in Step 1, we will be able to improve our online monitoring so can identify both the most virulent ideas that are percolating within each conversation (measured by the activity around the idea including comments, photo tags, unique commentors, length of the comment, etc.) and cross reference it with the most active contributors in each conversation (by volume of contribution, across multiple platforms and media type).
3. If we then build an algorithm that weights the activity around contributions (audience receptivity to a contributor as measured by their engaging with that contribution) with volume of contributions (across multiple media like blogs, Tweets, Flickr tags, etc.), we’ll have a useable, ever-changing snapshot of which individuals are controlling and guiding a particular topic at a particular time. Ideally, this system would allow you to adjust weights like a graphic equalizer to get different outcomes depending on exactly what you’re trying to learn with your search.
Summary
When you work for a social media agency, you end up participating with the “true believers” in the world (like the folks on this roundtable), but you also participate with the openly skeptical and the cautiously optimistic. They neither know nor care who Robert Scoble is and their universe (online or off) is unlikely to intersect with his. But I firmly believe that there are billions of interesting conversations taking place out there, being steered (intentionally and unintentionally) by hundreds of millions of micro-influencers. To chase them all as somehow wired differently than us is not the way to go. To have better tools by which to find them within the conversations just as you need to is the way to go.
With that, I’d like to tag Jeremiah Owyang for his thoughts on the concept of context-based influence versus personality-based influence. I’d also like to tag Duncan Watts for his thoughts (although I can’t seem to find his actual blog, so doubt he’ll see this). I’ve likely bastardized and twisted his brilliant thinking into something he’ll barely recognize. I’m always curious as to what Geoff Livingston thinks. And, finally, of course, since I’m contributing to a roundtable that I did not attend, I’d love to hear from the actual participants, including Jonny Bentwood, Max Kalehoff, Sarah Petersen (who is either too shy to blog or needs some SEO help), Henry Copeland, Jeff Jarvis, Steve Rubel, Keith O’Brien, Richard Edelman, Rick Murray (another hidden/nonexistent blog-irony noted), David Dunne, Peter Kim, Charlene Li, and Dr. Walter Carl (congrats on the baby by the way—she’s very cute).
Now, who wants to build this tool I described? I’d certainly pay to subscribe to it… Heck maybe we’ll build it, who knows…
