It’s Social Media Week in New York City, and the digital scene is buzzing with conferences and brimming with ideas and insights about a culture that’s constantly in flux. Yesterday, I attended the Advertising Research Foundation's event dedicated to 'The Science of Social Media' (#ARFSMC), and found myself particularly engaged by a presentation given by Jeff Doak, Chief Technology Officer of Converseon, called Social Media Mentoring Metrics. Jeff’s presentation revealed some staggering insights into the analysis of social media metrics that confront the challenges of analyzing metrics via the confines of automated tools that are still in the infant stages of their technical evolution.
The first monitoring metrics that came under the lens were metric returns for “volume”—which is basically to say how many times a brand is mentioned, period. The context for which a brand was mentioned in is, of course, up for grabs, and that is one of the biggest problems. Volume without context is really pretty useless, and it’s really up to humans to decide how that volume aligns with overarching business objectives. One trip to Twitter shows you that your friend Bill did, in fact, wake up this morning with his usual glass of Uncle Matt’s and drove to work in his Prius—and not much else. There’s your volume.
Well, what about influence? Maybe Bill’s got the hottest blog right now on a low-emission lifestyle and over a hundred-thousand devoted readers who are hanging on his every tweet. That’s not something your influence metrics are going tell you at face value, and those metrics are never going to provide perspective for the way your brand fits into Bill’s sphere of influence. That’s the kind of pithy insight that only comes with human analysis.
Lastly, we really can’t forget about sentiment metrics, which is to say how people are talking about your brand. Current social media monitoring tools measure something called “automated sentiment” which is basically an algorithm that searches for “sentiment words” and their proximity to your brand or product name. Huh? This analysis seems logical in a mathematical way, but we’re talking about words and that arbitrary phonetic symbolism that begets meaning. Human beings can also be pretty witty, and as Jeff so bluntly puts it, “machines don’t get sarcasm or slang.”
Enter the hammer and the scalpel. Jeff prescribes the use of one these tools, metaphorically of course, to analyze your social media monitoring metrics. The hammer is most appropriate for those analysts looking to smash into that data and create some cool charts that’ll impress the boss. For the rest of us, it’s really the scalpel that going to come in handy. The scalpel is best suited for that kind of detailed analysis that requires a surgeon’s precision and a human being’s understanding. Scalpels let you bisect tiny segments of consumer data to bring under the lens and expertly carve around other irrelevant brand chatter. It’s the scalpel that’s going to help you understand if it’s Bill’s preference for your organic orange juice that’s driving sales, and how you might leverage that affinity to connect with your consumers.
Right now, we’re working with social media metrics that have been around for as little as eighteen months. The measurements are bound to get better over time, and it’s nice to imagine having tools that create easily accessible and understandable data. But that’s just not where we’re at right now. I have faith in the human ingenuity that will eventually lead to greener pastures, but until then, I’m keeping my scalpel on hand. Kudos to Jeff from Converseon for an enlightening presentation.
