The road to enterprise-wide master data management (MDM) is paved with small, departmental or tactical wins --...
but be careful -- because it's easy to lose sight of long-term MDM goals, according to attendees at last week's Gartner MDM Summit in Los Angeles.
Technology professionals say it's important to get executive support and clearly define business drivers prior to implementing enterprise MDM. Then it's time to figure out how the program will be initiated and gradually proliferated.
"These things have to start with the business theme -- otherwise, nothing is going to get funded, and there's not the social buy-in, which is really what drives all of this work," said Eric Adler, development manager for information delivery at the Bill & Melinda Gates Foundation. "Finding small wins in [specific] spaces is really a great way to [move forward]."
But be aware that at anytime problems can arise and cause the enterprise MDM program to stagnate, Adler warned. The budget could dry up for unforeseen reasons, for example, or short-term success in one tactical area could lead to complacency before enterprise-wide goals are achieved. That's why it's important to always be mindful of the big MDM picture.
"If you isolate one part of the stack, we're going to lose. That's all there is to it," Adler said. "You have to look at it from end to end."
MDM is an increasingly popular methodology that combines technology with data governance policies and procedures in an effort to synchronize and improve the overall quality of business information. The phrase enterprise MDM refers to the sometimes lofty goal of mastering data across all of an organization's business units and systems. Implemented properly, enterprise MDM promises to improve everything from the customer experience and supplier relationships to business intelligence (BI) reporting and analysis.
"MDM is a discipline and as such you will not be able to manage it with technology alone. It actually contains many elements which have to be managed in conjunction with each other," said Dimitris Geragas, senior director of consulting with Stamford, Conn.-based Gartner. "Start by simply looking at [where you want to focus] and figuring out what are the pieces that you will put in place to make this happen."
The specific business drivers behind enterprise MDM vary depending on the user organization's makeup and culture, said Steve Strout, CEO of Black Watch Data, an Atlanta-based data management software firm. He added that it's important to understand those drivers when making the case for enterprise MDM to higher-ups or formulating an execution strategy.
"You have analytical data that is very important to [some] companies. You have operational data that is very important to operational companies," Strout said. "You have different things that are important to different people."
Strout said that as with any large and complex project, small wins are essential to enterprise MDM success.
"The days of having projects [last] two years before seeing any results are pretty much over," Strout said. "Thus, a best practice is to break projects into small, 60-to-90-day chunks with specific deliverables and work through each one of them. This puts delivery of benefits to the business on a much faster track."
An organization might begin an enterprise MDM implementation by focusing on a specific data domain -- such as products, customers, suppliers or employees. And Strout recommends taking a multiphase approach.
Phase 1 is a good time to establish the basic data governance requirements, Strout said. Some key questions to answer during this phase include: Who owns the data? Who will steward the data on behalf of the owners? What are the specific data governance policies around managing the data? What are the specific procedures and metrics that tie back to those policies?
Once the data governance policies are determined, it's time for Phase 2. Strout said this is the right time to get the data harmonized, standardized, deduplicated and enriched. During Phase 3, organizations should implement any technology that will be used to maintain the MDM program over the long term.
"Add the appropriate MDM technology for ensuring data accuracy at time of data entry," he said. "Build the workflows for approvals [and] tie everything back to [data governance] policies, measures [or] metrics."