Visitor Engagement: Is it Worth Measuring?
There is so much talk by marketers about visitor engagement and what it means, but there appears to be no universal definition for it. There certainly seems to be little to no agreement in the marketplace on if and how engagement should be measured. Sounds like something we should avoid altogether, right? Well, if you are a traditional direct marketer who is used to analyzing and optimizing your marketing efforts based reams of hard, undeniable response and conversion data, then the idea of measuring engagement might seem unwarranted or maybe a bit scary. I mean, what's the point in measuring something that requires bias, judgment and creativity?
An interesting argument I just read by Matt Belkin of Omniture suggests that there is no place for subjectivity – let alone, creativity – in the arena of measurement. His post on Omniture's blog in July called the whole notion of measuring engagement a "fad." Having played in the web analytics space for several years now, I can appreciate Matt's position because tools like Omniture, WebTrends, and Google Analytics provide robust reporting on the traffic sources coming into your site(s), as well as the activities taking place on your site(s).
However, the problem with having so much visitor data available at your fingertips is that it can be difficult to determine the value of all of the variables. Many, if evaluated in isolation, do not appear to represent any tangible business value.
So, should you just ignore them? Should you avoid analyzing the behavior of a visitor that spends several minutes on your site watching video, reading content, or completing a series of seemingly trivial tasks. Absolutely not! I agree with a large population of marketers and analysts who believe that you just need to get creative in how you measure this activity. This is the point at which you create your engagement scoring model.
Essentially, an engagement scoring model requires that you prioritize multiple variables within a group and assign a weight to each one. Bill Bruno, a principal at Stratigent, a web analytics consulting firm, wrote a very nice primer on how to develop an engagement score.
The subjectivity is that you may have to determine your weighting or value assignment of variables without too much understanding of their individual importance. Hardcore direct marketers and web analytics experts may not approve of this approach, but if you understand your business well and are confident in how your campaigns are planned and how your site is designed, your engagement score may begin to tell you a story that will enable you to optimize your programs and site in new and meaningful ways. If you feel your engagement scoring model is ineffective or is not clarifying anything for you, you may need to consider reassigning values or combining different variables altogether.
The creation of these models is not an exact science, hence the reluctance by some to construct them. But, if you are delivering compelling content and feel it provides value to the user despite not being a part of your conversion funnel, why not assign an engagement score to that activity? For example, if a visitor plays a video on your site, watches a full minute of it, replays it and then forwards it to a friend, why not assess the value of those activities and consider using them as part of an engagement score. In this case, each successive activity may represent more value than the previous action (e.g., viral/forwarding is more valuable than a replay which is more valuable than watching the video in entirety).
Measuring visitor engagement can be an important step in improving your web analytics capability. If you feel you are lacking some key insights, try building one. If it fails, try reworking it. If you are still having trouble gleaning any meaningful insights out of the model, please tell me about it below. Or, be sure to comment on Matt Belkin's post and tell him how you got caught up in the fad.
Comments
Mr. Ellinger,
I agree whole heartedly...Nicley put!
~Shey
Posted by: Shey O'Grady | October 7, 2008 09:45 AM