A good data warehouse and an outstanding BI team can transform the fabric of a company, just as a good marriage can ignite the soul of a human being. With that being said, one must also take a look at the adverse side: a bad data warehouse and a bad marriage. They have more in common than one may initially think.

Bad Marriages and Bad EDWs

In both marriage and data warehousing promises are made and trust is established, but breaking promises and trust destroys both marriage and BI teams. How do you break trust in an EDW? Promising things like single version of the truth and not delivering; promising phenomenal dashboards without understanding the complexity (and dirtiness) of the data; working on solutions without understanding the problem.

I’ve been married now for almost 23 years, and my biggest data victories with customers have come from lessons that I have learned in my marriage. “My cuteness,” as I call her, may bring up an issue, and as a young man I would rush to solve it. What I learned over time can be summed up in two statements:

1. She didn’t mention it for me to fix
2. The “Real” problem was much deeper

If you have been serving customers for a while, you have surely discovered that what they ask for isn’t really what they want, and I’ve written an entire book to help find what they really want and forgot to ask for.

Select Statements

Bad EDWs and bad marriages have similar traits in that there is no thought for their select statements.  I know, this is getting nerdy, but follow me. Long running queries could mean no indexes, bad design etc.  In either case, it means there is an underlying issue of neglect that needs to be resolved. In marriage, you can tend to “select *” from your mate’s bad traits and want to report on them regularly. 

Take a note on what you are selecting. Are you selecting bad traits from bad teams and replicating them? This approach won’t get you far in your EDW and probably won’t work too much in your marriage either.

Expectations

Parts of the issue with marriage and data warehouses are expectations. Some expect both will be easy and automatic. However, if you have been married longer than 15 minutes or have a data warehouse with at least 1 customer, you know it ain’t easy. I know someone told you that the EDW would run itself just like they said it’s happily ever after, after the wedding. Well, someone lied to you. It takes work, and I don’t care how great your ETL routines are, how marvelously you have instituted change data capture, and how great your MDM implementation is. It still ain’t automatic.

We were told that there is only one way of looking at a problem and only one solution. However, there are always other options. And as time progresses, those options will change. Some will grow obsolete and others will make less sense as the business changes. Sometimes finding the right option is as simple as looking through the lens of time and experience. I’ve actually disagreed with myself before, and I’m sure you have too. I think most can look back on decisions and wish they had done something differently. So, there are different ways of looking at the same problem.

But we were told to go looking for the single version of the truth. However, attaining a single version of the truth is a misguided expectation. Different truths come from different people. Some people want ROI based on sales or increase in customer service scores. Some want sales based on expected customers, while others count revenue after the service has been completely rendered. It’s just different. There are ways to unify this, but it’s not the easy explanation we were sold.

There are tons of other ways to relate the two (marriage and data warehousing) together. This first article is part 1 of a 3-part series that may take your BI team and possibly your marriage, to the next level!  Stay tuned!

And hey, don’t worry… BI

Dewayne Washington

BI Pharaoh