Where is analysis headed?
The current trend among reporting and analysis tools is to build in more robust processes that handle the heavy lifting of analysis. Adobe introduced a few ideas at the Adobe Summit earlier this year and a recent video on Google's Content Experimentation hints at something called "Market Insights".
As with the democratization of any highly skilled function, this has implications on how we work. In the past heavy-duty analysis was done in the backroom by the propeller heads, and I've done my fair share so that isn't an insult. In that scenario a fair amount of time was spent on explaining the data and the process itself. When it is baked in that all disappears and we're left naked asking the question: What do we do now?
Well first, we have to pay attention since we're now looking at what is vs. what should or could be. It is that gap that generates action.
report fatigue that Gary Angel wrote about recently. The reality: Everybody wants reports, but nobody actually uses them. The reason is simple; if things are running according to plan then there is no reason to look at them since they answer old questions with new data. This is why KPIs represent a paradox: they are critical, but they don't change much.
One of the advantages of baking in insights, or at least deviations from expected patterns, is the chance to build alerts that notify us only when we need to pay attention. And intermittent reinforcement is a much more powerful motivator than the weekly batch of sameness.
While the purists may argue that 'modeling is an art', the benefits of baking the science in out weigh that one.