Thursday, September 28, 2006

Tech: Don't waste your time on a Data warehouse!

Dratz did an interesting post on data warehousing.

In effect: why bother?

In concept, a data warehouse is laudible: one place where you can store all the organisation's data, integrated across business units...

But why? Do you have a business case for it? A good reason you can't extract the intelligence you want from the existing databases?

Well maybe you can't. But maybe you can. Dratz pointed out that the "if you build it they will come" philosophy is the ruin of many a technologist. It just doesn't necessarily happen.

At my workplace, there's a data warehouse. But does it contain feeds from all the company's repositories? Noooo. In fact, much as you'd like to think the organisation is all one big happy family, there remain concepts of ownership. Competition, even.

Break down the information silos? Look at the business silos first. Work on the business issues, the internal politics. And tell me integration is axiomatic.

All I'm saying is, think more carefully about what you're doing. If there's a green light, start small. Don't try to bring on board business units that aren't yet interested - or even jealously guard their IP. Datawarehouse by all means, but only as and when there are specific business returns.


Bazza said...

That looks like the kind of advice that firms pay big bucks for, Stephen. Is this your profession?

S Simmonds said...


It wasn't me that said it, it was Dratz. You might want to give him lots of money.

I largely agree with him, although sometimes it can pay to have a visionary aim.