Making correlations between events can be a slippery slope when it comes to marketing measurement. The untrained eye may look at two events and assume a connection that might actually have a very weak association, or not be related at all. It's a common trap to fall into for those marketers who generally don't have deep statistical or math skillsets to properly assess measurement situations.
Because of the meteoric rise of social media as a significant marketing channel, there is a lot of pressure to crack the puzzle of social media measurement. There have been many debates on the subject, and there are certainly plenty of people saying they have the answer. But sometimes it feels that folks are trying fit their solution into a model the way someone might try to cram themselves into tight jeans -- in both cases, it's a lot of work and there's not much room for error.
Whether they're in-house social media stars or external agencies and consultants, no one really wants to admit to whoever's signing their paycheck that there's still a lot to be figured out. Brands want definitive answers if they're going to continue spending the kind of dollars we've seen flow to this channel. No one is arguing that there isn't some major value here, but I think we all know the arms race for "likes" is over. Having "like" for "likes"-sake is not going to satisfy a CMO who's on the hook for these expenditures and business results.
Full disclosure here -- I work for a marketing attribution software vendor, and we are able to show advertisers the ROI on their paid social advertising just as we do for other channels such as online display or paid search. This definitely solves a big piece of the puzzle, but when I talk with social media practitioners, they articulate that there's a lot more value in social media than just the amount of direct sales it brings in from paid channels. There's brand affinity, earned impression value, and of course engagement (the single most difficult metric to define in all of digital marketing).
Ultimately, unlike most paid media that reaches a user with an impression, has a direct measurable [ROI] impact, and then fades from memory with time, social media practitioners approach this medium as something that can literally buddy-up with consumers. Yes, of course the goal is to impact sales, downloads, leads, etc., but how to measure this lingering, ongoing impact of owned and earned social media in the marketplace compared to the direct effect of paid social media is still the Holy Grail of social measurement. What metrics do we currently have that could be correlated to truly address this issue?
To dive in deeper, I spoke with Eric Swayne, director of social analytics and insights at M/A/R/C Research, a published author, award-winning frontend developer, and all around good guy to talk to about the state of social media measurement.
Josh Dreller: Eric, what's your take on correlations in social media measurement?
Eric Swayne: The caution here is to avoid thinking every correlation is a causation -- not all things that are related actually push or pull each other. Before you start correlating everything together, try some simple gut-check questions: "Does this activity directly move these customer actions?" and "If my marketing is intended to change perceptions, is there a way I can measure perceptual changes rather than correlating with something further down the chain?" Correlating metrics that have a natural relationship with each other gives you a better chance of success when trying to move business outcomes.
Dreller: What are some ways social media practitioners are correlating metrics to events?
Swayne: A number of companies have correlated social conversation in the days leading up to the release of a movie with how that movie performs at the box office. Those metrics typically do have a correlation, but generally because great movies get people excited before they come out, which creates conversations. It doesn't necessarily work the other way -- getting more social before a movie launches doesn't make the movie better; a hotly discussed bad movie will still tank at the box office.
Another example would be correlating social media with stock prices -- again, something many companies are doing, and a relationship that typically reveals some solid correlations. But positive buzz isn't an indication of a strong company; it's a digitized measure of the word of mouth surrounding a brand. A rising stock price can have a strong relationship with an increase in positive conversation, but typically those are both driven by an external factor, not each other. Apple could drive tons of positive buzz by announcing it was giving away the iPhone 5 for free, but that definitely wouldn't drive its stock price higher.