Yesterday I wrote about the different styles of master data management. It can be pretty confusing, as I mentioned. So, today I wanted to share an item that might help simplify your choices.
It's important to know the terms vendors will use, but for your own knowledge, do yourself a favor and read this recent post by Forrester Analyst R. "Ray" Wang. Wang doesn't get bogged down in all the varieties and incarnations of MDM products. Instead, he looks at the three architectural styles of MDM:
- The cross-referenced registry style
- The hybrid harmonized reference style
- The transactional operational data store style
Please note: He's not adding to your list of MDM choices. Instead, he's describing the three ways to build MDM. And, he has made it much easier to decide which option is right for your business by focusing on the pros and cons of each architecture style. He even outlines, in the accompanying picture, the typical business use for each MDM style, as well as how it affects data usage, data model and deployment times.
In short, it's the Cliff Notes version of how to pick an MDM style that will fit your business needs.
For instance, one of the MDM styles he discusses is cross-referenced registry. He explains that this approach provides a read-only, but real-time reference for the data. It can be deployed rapidly - a mere one to three months versus the 12 to 24 months required for an operational system that's going to synchronize the data across your systems. But, it doesn't address data quality or de-duplication of source systems.
I asked Wang by e-mail how these architectural styles aligned with the types of MDM. He replied that analytical MDM is best supported by options one and two (the cross-referenced registry and the hybrid harmonized reference style), while operational MDM is best supported by options two and three (the hybrid harmonized reference style and the transactional operational data store style).
One caveat about the inserted table: It's impossible to read within the post. I was frustrated that I couldn't double-click to enlarge it and complained that he should include a bigger version. Wang pointed out that you have to use the old-school method of right-clicking and choosing "View Image" to make it readable. I had a big "duh" moment, but I'm glad I bothered to mention it, because the information was well worth the trouble.