As an industry, social digital analytics is still in an early stage of maturity. In contrast, the Web analytics world has moved from a place of measuring “hits” to using advanced statistical techniques to optimize product, marketing and content decisions. Online video analytics was in a similar place to social analytics back in the mid 2000s.
I recall, years ago, sitting in a pitch from a well-known video measurement provider during which the company pitched the possibility of literally measuring every possible action that a user might take, including when the viewer dragged the mouse across the video. How this was supposed to provide insight into content development and optimization was was anyone’s guess. The old adage is true: just because you can do something doesn’t mean you should.
Similarly, there is no shortage of data in social analytics. A colleague of mine shared a dashboard that she received from a third-party vendor that included no less than 25 different metrics for just one Facebook app. Unfortunately for her, their social campaign contained three Facebook apps, a Twitter program and some paid search. Sifting through all of this data to build a meaningful story around her campaign for her CMO was the primary issue.
To deal with this issue, some providers deliver vague metrics like “interactions” with an app. These metrics serve a useful purpose where an analyst is using appropriate statistical methods to group metrics together to uncover specific insights. The goal of lumping vague metrics into a bucket like this is just to show big numbers with the hope that this will win over more funding for social programs. Most CMOs, however, are smart enough to ask questions about what an interaction actually means, and really want to get at the real return on investment.
Presenting vaguely defined metrics and a lack of transparency about what the data really means, how it was instrumented and how it was counted only leads to mistrust about the data and the insights. Every analyst should be prepared to discuss methodology with as much detail as possible. There are weaknesses in every measurement methodology (thank you Dr. Heisenberg); let the client decide to what level to trust the findings based on the level of accuracy of the work.
Speaking of error, one evolution of social analytics that we will see over the next few years is a move toward measurement standards. This is a good and necessary thing. As we continue to observe success in social media marketing, more dollars will begin to flow into these programs. As this happens, advertisers will want to know more and more about what the return is. We have seen this in Web analytics, mobile analytics and video analytics. I am hopeful we will see this soon in social analytics, because it will mark that social is being taken as seriously as it needs to.