The Dirty Data Conundrum

Loraine Lawson

In a recent IT Business Edge interview, Ron Schumacher expressed a shockingly contrarian view: Maybe having a single version of the truth isn't necessary.


Schumacher, the CTO of Initiate Systems and a proponent of federated CDI, suggests that sometimes it's just fine if John Smith is known as John Smith to sales, and J. Smith to the ERP system. (To be clear, he certainly doesn't advocate this situation in every case.)


Thinking about this idea, it occurs to us that he might have a point. It's perfectly possible that Mr. Smith may not only be happy with different versions of his name, but different versions of his address, as well.


Aaron Zornes, a monthly columnist at DMReview, probably wouldn't agree. He sees having clean data as one of the ten best practices for CDI-MDM.


And SOA data problems are also guaranteed when there's no single version of the truth, although in that case it's a higher-level problem, having to do with definitions, such as: What is a customer?


These various points of view all show that dealing with the inconsistent data issue is problematic. Cleaning up everything would hardly be cost-effective for most companies, but sticking with the (dirty) status quo can lead to real problems, as well.

Add Comment      Leave a comment on this blog post
Oct 30, 2006 7:21 AM Matt Stitz Matt Stitz  says:
What a useless column/article!  Tell me why it's okay to have different views of the same customer.  Essentially your breaching the topic of which data should a central hub master, and which data it should be a pure reference file for.  Where do you draw the line between the two and why? Reply

Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.