If your BI application only reports the past, you’re missing out on a major goal of BI systems. BI systems should be forward-looking, so they can inform decisions. In order to do this, you need some sort of analytics, whether they be simple questions like “where are we against plan” or “how does this compare to last year”. Or they can be more complex, as in “which customer segments provide the greatest possibility for our promotions” or “what would happen if we realign our territories in a different way?”

If you are looking at these sort of analytics, you have a choice to make when you design Oracle BI systems that run against the Oracle database. BI is great at presenting data, but where should your calculation and aggregation logic be? How can you capitalize on Oracle database capabilities such as Oracle Data Mining, Oracle R Enterprise, or Oracle OLAP cubes? For that matter, how does Oracle In Memory change the game? Now we can compute aggregates directly in the database very quickly in memory. These all raise the question: “What should be done where?”