Analysis begins with your implementation. I’m not talking about making sure your data is being collected properly. Let’s assume it is. What I am saying is if you aren’t thinking about how your traffic MIGHT behave during your implementation, you are missing out on an entire new (deeper) level of analysis. Think beyond click tags. Think beyond custom variables that show whether they are registered or not. Think beyond a content hierarchy. Think about what you expect users to do on the site and then group them up in your head. Typically we refer to them as advanced segments in a tool like Google Analytics or SiteCatalyst. This post isn’t about segmentation, either.
I’m suggesting we take a step back and think about how most analysts approach analysis today. It’s a reactive approach. We hold stakeholder interviews to set our goals, we design an implementation strategy to measure these goals and other little tidbits, and finally we take the output that’s in our measurement tool and build our our cohorts/segments. This isn’t the wrong way of doing it, but it will quickly become the old way of doing it. With the proliferation of tag management systems, we don’t have to think of the process as reactive. As analysts, we’ve been hamstrung by the bottleneck of IT to the point where we physically couldn’t create and test complex theories and implementation strategies. Everything we’ve done has been a reaction to the event tags, custom variables, props, and eVars that has sustained us over the past few years.
As a consequence, we’ve had to suffer the tedium of spending hours building the perfect segments to characterize what we think are accurate user behavior trends. Let’s change that. Let’s build theories where we classify users based on advanced page sequencing – but let’s do it in Google Analytics. Let’s design complex characterizations of a user and set them as eVars in SiteCatalyst. Let’s do all of this before we collect the data. Let’s act upon that specific segment without the bottleneck of IT (and the nightmare of the QA). Let’s change the way we think about analytics. How do we do all of this? Start here.