Marry your database. Date your dashboard.

May 2, 2011   //   Business Intelligence, , , , , , , ,

I’ve been married since 1995, and dated my wife for five years before that.  Over the last 20+ years, our separate lives have become one.  We own our house and our cars jointly.  We share bank accounts, furniture, children.  We share most of our friends.  Sometimes it seems that we share thoughts and opinions (although that’s mainly because, like any guy, I’ve devoted an enormous amount of effort to answering correctly when Sara asks “What do you think of that color?” and other critical questions.)  We still have our own tooth brushes, but how much longer can I really expect that to last?

When my single friends (there are still a couple out there) date, they keep the commitments to a minimum.  Even when they move in together, they tend to keep it clean in case things don’t work out.  One friend can stand in his condo and can tell you, object by object, what belongs to him, and what belongs to his girlfriend (although there seems to be a bit of ambiguity about the toaster oven).

So what does this have to do with BI?  Marry your database.  Date your dashboard.

Your database and the other critical components of your data architecture like ETL tools and cube development tools require a long term commitment.  You’re going to have to invest in that relationship, make compromises, play to your database’s strengths.  You’re going to be married to that data architecture for a long time (hopefully forever) so take your time to make sure that your choice is a good one.  Find out what her long term plans are.  Agree to fight through the future for better or for worse.

This has real-world implications beyond the selection process.  You should invest in the training for your staff and the long-term vendor relationships that befit your newly-married status.  Invest in the data architectures that will scale with you.  Make sure that your ETL process are rock solid.  When you amortize the cost of those new windows over a fifty-year marriage, the decision gets easier.

Dashboard and analysis tools are a whole different story.  They come, they go.  They’re flighty.  And let’s face it, you might decide in a year or two that you want that newer model.  Your lack of commitment to your dashboard tool comes with its own implications.  The licensing model needs to assume a short-term commitment.  The learning curve on these tools needs to be short and sweet.  In addition, when designing the dashboards and analysis assets, remember to solve complexity in the data architecture rather than the user interface.

No 30-year mortgages with that girl you just met in a bar, although it’s probably OK to buy that toaster oven together.