What can I say about data quality that hasn’t been said a hundred times already? Plenty, as it turns out. Honestly, it was the most difficult chapter to finish in Healthcare Business Intelligence-the book. It is among the first chapters I wrote, but I think the reason I had such a difficult time with it was because I feel that the topic is so important and critical to the success of BI in healthcare that I needed it to be perfect. I needed it to not only convey the importance of data quality, and frame it in a way that was achievable, but to also throw out the idea of 100% data quality.
A few years ago I was reviewing a contract for a new client (we provided reports to our external customers of our services) and I noticed a line in the contract that guaranteed 100% data quality. It’s hard to fathom this. If you stop to think about it, we were had very little control over the data that the client sent us (one large historic batch load) or the data that would be sent to us regularly as part of the service. I didn’t have to sweat it out too long, as it turned out, you can’t contract for that, because you can’t control it. If lawyers understand that how come most business users don’t? In truth, most people do understand it, they just don’t care. I am certain that my eighth grade math teacher Mrs. Connelly would argue with me, but I think 100% is relative, at least when it comes to data quality.
What business users want is reliable data, trusted and consistent data. They do not want all data for data sake and they certainly don’t want bits and bytes of data without context.
The point? Data quality is hard, but important. There’s no such thing as 100%. Data governance will help define data quality procedurally for your organization. You should start here.
Leave a Reply
You must be logged in to post a comment.