Last month, I wrote a blog post on Tips for Succeeding with MDM or Data Quality Initiatives. Cliff Longman - I assume it's the same Cliff Longman who is the CTO for master data management (MDM) solution provider Kalido - wrote a response noting that it might help readers to understand there are actually two types of MDM, operational and analytical MDM. He thought much of the information in the post applied to operational MDM.
Actually, I have no idea if the information I referenced applies to analytical or operational MDM. That's because this type of distinction is seldom made when vendors, analyst or trade journalist write about MDM.
Assuming I'm not the only one confused by this, I decided to do a little digging into these two types of MDM. What I found is that it's pretty hard to distinguish the marketing terms from useful descriptions. Shocking, I know.
Longman explained the difference thusly:
"Operational MDM seeks to straighten out the data in the operational systems for operational reasons. That's why it is big, risky, invasive, and expensive. Analytic MDM seeks to leave the operational world alone (phew!) and simply concentrate on compiling a view of data that can be used for analytic purposes. Much quicker, cheaper and less risky!"
Therefore, Longman suggested companies opt for analytical MDM.
As with all things IT, not everyone agrees. And there are a few other issues to consider.
In a 2006 article for Intelligent Enterprise, Ventana defined the difference between operational and analytical MDM implementations and argued most global organizations will need both:
"Operational MDM (O-MDM) is focused on the distribution, synchronization or exchange of master data to ensure consistency in transactional operations. Analytic MDM (A-MDM) is concerned with the management of the master data items and associated hierarchies required for aggregation and analysis."
Or, put another way, operational MDM synchronizes master data, including product or customer information, across the company's transactional applications, while analytical MDM reconciles data "drawn from a variety of sources to deliver integrated, consistent business intelligence across the entire business."
As I read it, this means operational MDM is making sure all the sources stay on the same page - pushing the information to the individual sources - while analytical MDM is sorting it all out in a hub to create one view - pulling it together from the various sources and then analyzing it in the hub. Typically, analytical MDM is seen with business intelligence tools.
To complicate things, there are also two different flavors of MDM products - one evolving from customer data integration (CDI) and one evolving from product information management (PIM). As the name suggests, one is designed for for customer data, the other for product data. (I should also note that IBM introduced a third flavor in 2007 - its multi-form MDM solution, which promised to let you use the solution for either customer or product data.)
The Intelligent Enterprise piece explains that both fall under the category of operational MDM. It also explains which vendors specialize in which approach. It's good for a general overview, but the information is from 2006 and no doubt dated by now. (As I said, people tend not to separate the two, so I haven't found a more recent survey - stay tuned.)
You might also want to check out this more recent piece, a sample chapter from "Enterprise Master Data Management: An SOA Approach to Managing Core Information," published by IBM Press. It explains why companies need MDM and the types of MDM solutions, including a discussion on industry-specific MDM offerings. It also explains a third usage for MDM -collaborative authoring, sometimes called collaborative MDM.
Finally, there's this explanation of operational versus analytical MDM from Roger Wolter, who originally worked on Microsoft's MDM product team but now (well, as of an 2008 post, which was pre-layoffs) worked as an architect on Microsoft IT's internal MDM project.
Wolter explained that the data for the two styles actually looks the same, though there are a few exceptions:
"Transactional MDM might have a few more attributes associated with a given entity because there are things the operational system cares about that that aren't required for analysis. An Analytical MDM hub will probably store more hierarchies than a transactional hub because there are generally hierarchies that are interesting in analysis and reporting that the operational system may not care about."
He noted that the two styles also load and publish data differently. In either case, he argued these differences are incompatible, so you could probably just use the same MDM hub for both, assuming you ensure the solution will support both:
"My take on it would be to look for an MDM solution that can support both transactional and analytical styles I think in most cases the logical progression would be to start with analytical MDM to master the data models, rules, technology and stewardship required to manage your master data in a less mission-critical environment. Once you have achieved some successes in analytical MDM, you can use the same data, models and processes to manage the master data for your transactional systems by just adding the publishing logic to push the master data into the operational systems."
I'm sure that some vendor will ping me to tell me what I miss - as they should. This isn't meant to be a definitive list of MDM categories. Rather, I'm just trying to make sense of it and wanted to share what I've learned thus far.
At this point, here's the one thing I can tell you: The MDM market is confusing and CIOs are going to have to be on their toes. Not all products will address all MDM needs, so you'll need to think long-term to ensure the solution will do what you need when you expand your master data program down the road. Otherwise, you may have to invest in a new MDM solution and, if you're not careful, you could just wind up with MDM silos.