As late as yesterday I was involved in yet a data quality issue that wasn’t caused by that the truth wasn’t known, but caused by that that truth wasn’t known in all the different databases within an enterprise and of course exactly not (thanks Murphy) by that application that needed that information due to a new requirement. Yep, the column was there alright, but it wasn’t updated, because until yesterday it didn’t need to be.
The data architecture in most enterprises isn’t perfect at all. Through the information technology history of that enterprise many different systems has been deployed ranging from core operational applications, data warehouses and lately also web frontends.
It’s not that we don’t know about how master data management can help, how service oriented architecture (principles) is a must and how important it is to document the data flows within the enterprise. But gee, even for a modest sized organization this is huge and even if we strived to do it right, when we succeeded, the real world has moved.
Well, back to business. What do we do? I think we will:
- Make a quick fix that solves the business problem to the delight of the business users
- Perhaps prioritize up that sustainable technical solution we planned some while ago
Have a nice day everyone. I think it is going to be just fine.
Liliendahl on Data Quality 