ORLANDO, Fla. -- When it comes to master data management (MDM) best practices, companies must behave better than a bunch of children fighting over a toy.
Not that John Radcliffe, a research vice president for Gartner Inc., was suggesting MDM is a child’s game with his analogy.
It’s difficult managing the many forms of customer information, he said last month at the Stamford, Conn.-based firm’s
For more on the use of MDM
Learn how Gartner puts Oracle and IBM in the lead in its Magic Quadrant for product MDM software
But companies, more often than not, fail with MDM software because of internal company politics, Radcliffe said. Departments aren’t sharing information, and companies sometimes allow in-house rivalries to prevent everyone from seeing the bigger picture, he said.
“We live in a world of silos, with organizations between organizations,” Radcliffe said. “Marketing hates sales; sales hates service. They all hate finance.”
Instead of bickering over a toy, Radcliffe urged attendees to instead look at “the reality of life,” which is “the single view of the customer.”
Gaining that view won’t happen overnight, even with the best processes in place and little inner friction, Radcliffe said. But if a company wants to thrive, it needs to adopt an effective MDM strategy.
Radcliffe stressed the importance of governance in such a strategy; in other words, the assignment of people to manage the data. It is their only job.
Radcliffe was preaching to the choir with B. J. Flowers, the director of enterprise information delivery for TXU Energy, a retail electricity provider in Dallas. Her company hasn’t found anyone who wants to manage data.
“It comes down to governance, and no one wants to do governance,” Flowers said. “There’s no role for that. We tried to do it. It’s not sexy.”
MDM should connect CRM, other systems
Radcliffe also spoke of the sprawling infrastructure that companies have in place to process customer information.
Companies often have multiple CRM and transactional operational systems, a centralized marketing platform that may be connected to a contact center, surveys of customers, external data bought from sources and many social networks to track, he said.
“You’ve got this channel and that channel,” Radcliffe said. “Life is getting complicated. You’re not alone if you think managing customer data has become a difficult task.”
Creating a new MDM hub would work wonders, Radcliffe said. It can solve data quality issues and the semantic differences that occur when mingling programs, he added.
But Radcliffe reminded the audience of the true hurdle to installing an MDM hub: cost. “It can be expensive,” he said.
Companies could otherwise consider using two different content aggregation technologies, a mashup or a portal, Radcliffe offered. But these don’t solve data quality issues and semantic differences, he said.
Data governance key part of MDM best practices
Ultimately, though, MDM best practices come down to governance, Radcliffe said. Companies need to decide who will manage the data, he said.
Often, data quality issues arise when nobody has an interest or when people have too many interests, he said.
Employees will typically think, “It’s in a database somewhere; it must be IT,” Radcliffe said. “But IT says, ‘I don’t create this. I don’t understand the impact of getting this wrong, the value of it. This is really a business issue.’ ”
On the flip side are the turf wars that occur from the politics of silos, he said.
A company needs to create something in between, and that is a data steward, Radcliffe said.
What is master data management software?
Master data management (MDM) software combines data governance with technological tools in an effort to ensure higher levels of data quality and data consistency throughout an organization. MDM programs typically make use of middleware to create a master data hub that becomes the single source for all of an organization’s approved data definitions. The data governance aspect of MDM focuses on aligning business and IT in an effort to identify and fix the flawed business processes that lead to inconsistent data.
This manager of data, he said, jumps into the lifecycle of information, establishing the root of building an MDM framework: Who’s creating data? Where is it flowing? Where is it being retired?
As soon as a company understands the lifecycle, it can then say, “We need this guy, that guy,” Radcliffe said.
Radcliffe said he’s starting to see a hierarchy of MDM hubs -- multiple levels of systems that are “hard merged” locally but “soft linked” centrally.
For instance, a bank has an MDM hub for its retail side, another for its commercial side and another for real estate, not to mention another hub when the bank acquires another bank.
MDM is necessary if a company wants to link to social networking and the mobile world, Radcliffe said. Mobile and social have “led to an explosion of data generated outside an organization’s walls,” he said.
MDM can’t govern the external data, but it can identify references to customers, products and brands, he added.