Organizations that set out to achieve the lofty goal of enterprise master data management (MDM) often begin with...
an analytical MDM program focused on the information in their enterprise data warehouses.
But no matter whether you invest in analytical MDM as a first step or a self-contained initiative, it can help give your end users better insight into business operations by ensuring that warehoused data is accurate and consistent -- provided, of course, that you properly manage the MDM process.
SearchDataManagement.com talked to analysts and technology professionals with MDM experience to get their advice on how to ensure the success of an analytical MDM project. In no particular order, here are the top five analytical MDM tips they had to offer:
Do your homework when evaluating analytical MDM software offerings. There has been a longstanding debate between software vendors and IT industry analysts as to whether analytical MDM should be considered a standalone product category. Some vendors claim that analytical MDM software is exactly what they’re marketing. Some analysts, meanwhile, think that operational and analytical MDM tools go hand-in-hand with one another.
For example, Forrester Research Inc. analyst Rob Karel said he views analytical MDM tools as “essentially governance-driven operational MDM products” with embedded workflow and data stewardship capabilities geared to supporting information such as financial data. "I really use 'analytical MDM' to be less the product category and more the architecture and business objective," Karel added.
But the vendors appear to have won the battle, said Dan Power, founder and president of Hub Solution Designs Inc., a Hingham, Mass.-based IT consulting firm.
"They fought tooth and nail over that kind of classification schema and said there is definitely operational MDM and there is definitely analytical MDM," Power said. "And the market bore them out because [analytical MDM] is one of the fastest growing areas of MDM overall."
Wherever an organization falls in the debate, it's important to conduct extensive due diligence when vetting MDM vendors, according to Power. "Everyone now will tell you that they can do analytical MDM, but don’t try to put a square peg in a round hole," he said.
Put the business in the driver's seat. Analytical MDM focuses on improving the quality of data used to gain business insights. That's why it's necessary to have a cross-section of representatives from the business side of an organization involved in the analytical MDM decision-making process from the beginning.
"It’s all about the business," said Holly Anderson, a former manager of enterprise data management at the U.S. Department of Education who now works for another government agency. "If your data management program cannot demonstrate how it enhances or improves or services the business, I think you’re sunk in the water."
Power agreed and said that many analytical MDM projects are too IT-driven. As a result, they often miss the mark when it comes to meeting business requirements.
"Put the business in the driver’s seat and don’t make it all about the technology," he advised. "Dig into what the business needs."
Make sure the organization is clear on terminology. Words, phrases and numbers can have different meanings depending on which member of an organization is talking, and that can cause big problems on MDM projects, according to Al Moreno, IT solutions architect with TopDown Consulting Inc. in San Francisco.
"If I’m the president of the company and I’m talking to a sales executive and I’m talking to the head of my accounting department, we’re all talking about gross profit but we all might mean different terminologies," Moreno said. A key part of a successful analytical MDM program, he added, is enabling users to look at a particular data point and be able “to say with certainty that we’re all talking about that number in the same way.”
An MDM system can help an organization enforce rules around terminology, Moreno said, but he also thinks it’s important to work with people face-to-face to make sure everyone is on the same page. "I tend to see that as a governance issue," he said.
Don’t forget to define business rules in the MDM system. Organizations often define at least some data-related business rules inside of extract, transform and load (ETL) tools used to pull information into a data warehouse. Others define those rules within the data warehouse itself. But the rules tend to be difficult to access, maintain and update in those systems, analysts warn.
That's why Moreno and consulting firms such as The Information Difference Ltd., which is based in the U.K., believe it's important to maintain business rules within the MDM system itself.
"It’s preferable to have [them] inside of one system where everyone can discover the rules, everyone can see the rules and everyone knows what’s under the covers," Moreno said.
Make sure long-term goals are architected into the system. For some organizations, analytical MDM is all they need and all they want -- but for others, analytical MDM is a stepping stone to larger MDM plans, according to Karel. If you’re in the latter group, he said, your analytical MDM plans must be undertaken with an eye for the future.
"If you are truly looking for this to be a scalable system that is going to support both transactional and batch MDM requirements, you have to make sure the technology decisions, the architectural decisions and the governance decisions [are made with that in mind]," Karel said. "You can’t build a business rule that is just for the data warehouse if you eventually expect the call center and finance and order management to also leverage that master data."