Democratizing Business Intelligence?
By John Parkinson
I would never make it in marketing.
I hated the term "AI" when it first surfaced (one of my friends back then used to call it "artificial common sense".) I feel just the same way about "BI."
It's business analytics, not intelligence that we get from mining, filtering and analyzing the flows of transactional and contextual information in which every business swims. Not that it isn't valuable to do the analysis--it is. It makes you potentially better informed about many aspects of business operations. It just doesn't guarantee to make you any smarter.
So there is this big push to give "BI" tools to more people. After all, if it's good for a bunch of back-room analysts, it must be just as good (or better) for lots of front-line decision makers as well.
Let's apply self-service principles to BI--and connect everyone who makes or influences the myriads of operational decisions that occur every day to the tools and data that will help them make those decisions better and faster.
Now there's nothing wrong with this, in theory--just as long as you do it right. And doing it right means avoiding the temptation to make that connection to the underlying data a "live" connection.
Even though front line managers are often best positioned to spot errors in the data and problems with the analytics, you can't afford to let them just wade right in and "fix" the errors. That one source of truth you spent so much time and effort to create just got corrupted.
Sure, you can add controls, audit trails, and checks and balances, but look at what can happen in the "Wikipedia" approach to master data management and cringe at the thought of what this would do to your data.
That's not what democratization should be about.
See also: Kimberly-Clark CIO Ramon Baez on BI
John Parkinson is CIO of TransUnion. To read his columns for CIO Insight, click here.