Tuesday, June 14, 2011

Social Media Measurement: Winning or Winging it?


People are "winging it" with measurement in social today. Marketers embarrassedly tell me this daily.

In my social psychological opinion, those of you who aren't abiding by any standards, measuring a little sentiment here, a little influence there, and a lot of buzz everywhere are falling prey to a cognitive bias: The Imposter Effect. You're denying yourselves the credit of being bona fide experimenters.

With no standards yet, and no evidence that a global social media metric standard spans all business goals and outcomes, the best thing you can do in measurement today is effectively operationalize your variables. Operationalize in the scientific sense: define the ambiguous concept you're trying to get at by coming up with a relevant way to measure it. We're an industry without gold standards. You have no choice but to wing it, but you can still do so empirically.

When social scientists study things like "health," "happiness," and "satisfaction" with marriage, jobs, or life, we rely on proxies. You often hear things like "health, as defined by number of doctor visits in a month," or "happiness, as defined by size of smile." In lieu of (or sometimes in addition to) getting self-reports of "how healthy/happy you are," these objective measures act as a best bet or starting point, eventually with some validation as to why that operationalization was selected.

Take a typical desire in social business measurement today: "we want to tie engagement to business results." So, how do we operationalize engagement? Importantly, this doesn't mean you should settle for a kitchen sink approach and add up everything to yield engagement, instead chose wisely - a realistic manifestation of what it means to be engaged. As I've said before, think about:

  • Objectivity-- Really think about what engagement means; don't arbitrarily involve variables simply because they’re available (e.g. # friends).

  • Reliability - Look for something that gets at engagement over time. Don't incorporate variables that measure the same thing multiple times (e.g. friends on Facebook + followers on Twitter + connections on LinkedIn).

  • Validity - show that your variables predict a meaningful behavior (e.g. statements of connection with the brand and likelihood to purchase, etc.). A great trick in defining variables is to think of the inverse. Get positive results of what you want (i.e. engagement) and negative results of what you don't want (i.e. disengagement).

As always, when you operationalize, highlight the right aspects of your business-- things that measure important movement and things that matter to those who consume the numbers.

This is also posted on the Dachis Group collaboratory. Join the conversation there to access a wider network of social business professionals.

Photo Credit: flickr.com/mlrogers

No comments: