I recently attended a Business Intelligence Users Group (BIUG) meeting, and, while half-listening to the (actually really good) speaker (Stacia Misner), one of the things she said sparked an entire train of thought that made me miss ten minutes of her presentation (fortunately, you can find most of her work here).
Where was I? Oh, right, a train of thought…I got to thinking about just What Is This BI Stuff everyone’s so hyped up about (and that, arguably, has paid some of my bills)?
For me, it boils down to something that can, appropriately, be represented as an equation: BI = BK + D. That is, business intelligence is nothing more (or less) than the consummation or melding of data (D) with business knowledge (BK). Data alone is not BI. Business Intelligence is the ability to take raw bits and bytes from some kind of more or less permanent storage and present it to a business individual (who generally hasn’t the faintest idea where the data came from, how it got there, or what’s happened to it) in a format which allows them to make a decision that has business consequences.
Note that the last part of that is crucial, in my mind. I’m not talking about lists of data. That’d be commonly referred to as a report. I’m talking about data presented in such a way that the business user can interact with, parse, congeal, expand, etc. in order to arrive at a meaningful (in the business sense) set of data to support (or deny) a decision.
The issue, I think, is that too often organizations rely on their data folks (or folks who know SQL – which is a whole different ball of wax for another time) to produce this thing called BI. Unfortunately, it requires true business knowledge to successfully meld with data and produce actionable business intelligence. Furthermore, business knowledge takes time to acquire. The Business Intelligencer (I think I just made that up), needs to understand the underlying business model (via something like the Subject Area Grain Matrix). Then, the BI-Guy can start to understand the source data, document the required transformations, and, ultimately begin to create a digital model that accurately represents the physical business. And it’s that digital model, sometimes known as a domain model, which enables decision making via both Structured and Unstructured Analysis (filterable reporting and ad hoc analysis, a topic for another time).
So…ponder that…oh, and Live long and prosper…