If there's one thing most master data management (MDM) and product information management (PIM) experts agree on, it's that MDM projects will succeed or fail based on the willingness of business executives to elevate data governance to a more important role than any other technology.
To create a viable MDM or PIM software implementation, manufacturers will need a full-fledged MDM or PIM application -- after all, data management tools are still required to manipulate data. Before technology buyers start evaluating data models, product architectures and software performance, however, they must understand the needs and goals of the business.
Those business needs and goals are "the most valuable and powerful criteria for evaluating MDM and PIM," said Michael Dortch, principal analyst of DortchOnIT.com. "Before an enterprise even begins to evaluate candidate solutions and vendors, it must have a clear picture of what it's trying to achieve, what it's done so far to try to meet those goals, and how those efforts have fared to date."
Because business goals and needs change with business and economic conditions, every enterprise must establish, document, enforce and manage consistent processes for assessing and prioritizing its own business needs and goals. "This information should then be used as a foundation for assessing and comparing candidate solutions and vendors, just as knowledge about who's using what information how often and for what should be used as a foundation for any MDM or PIM efforts," Dortch added.
Of course, the starting point for evaluating MDM or PIM technology will be affected by the scope of the project, including what kinds of data will be included in the MDM or PIM project, who the users of the data are, and how the data is currently stored. A good way to locate these and other evaluation priorities is to find existing problems and pain points in the data management process.
"The first step for any kind of master data management program is to clearly identify those business processes that are impeded by the absence of a unified view of the product landscape or that could be dramatically improved by providing that unified view," said David Loshin, president of Knowledge Integrity, a Silver Springs, Md.-based information management consultancy.
"Next, the organization must be aware of how the business applications will need to be modified to accommodate the master environment," Loshin said. "In some situations, this in its own right requires a significant level of effort."
Dividing up the MDM evaluation process
When a company is ready to start evaluating MDM tools, Andrew White, vice president of supply chain management research at Stamford, Conn.-based Gartner Inc., recommends breaking the process down into manageable subcategories. According to Gartner, an MDM or PIM solution should address eight categories: data modeling; data/information quality/semantics; metrics/analytics (of the process, and data); workflow; business services; architecture (of the tool itself); performance and scalability; and security. If the solution doesn't address these areas out of the box, then it should be able to do so through customization or modification.
Most manufacturers run an ERP system in one form or other, and the ERP software will have a big impact on the MDM and PIM product selected.
"An ERP system coupled with some of the more sophisticated identity resolution techniques might actually be as effective as, if not more effective than, a segregated product master," Loshin said. "Because the ERP system maintains its own single copy of 'the truth,' one might consider the ERP system's data set as the master environment."
The ERP option as a data store, in conjunction with MDM and PIM tools that are either built into the ERP system or added on, will depend on the size of the manufacturing company.
"For some users, mostly smaller manufacturers that run most of their business -- more than 80% -- on ERP, then yes, ERP is presumed the master," White said. "However, ERP is not extensible, so when a firm calls [itself] heterogeneous, this becomes costly."
According to White, "this is what manufacturers have been trying to do for years -- extend the item and vendor and customer master into the ERP system." He noted, however, that "ERP was not designed as an information infrastructure tool for a 'single view' across the application landscape" and that this is why MDM has become a standalone market.
"Most existing ERP systems do not have the workflow, data quality and role-specific user interfaces needed to manage the master data lifecycle," said Bill Swanton, vice president of industrial products research for Boston-based AMR Research. "Companies will need to array a variety of tools to solve their problems, including data integration, workflow and data quality technologies."
Elevating the use case of MDM
In fact, once manufacturing-based organizations create a foundation for MDM (although PIM may be more tightly focused), there's a good chance they'll find other areas of the business that need help in the area of MDM. And while finding these other areas may complicate the MDM initiative, they may also help elevate the cause of MDM throughout the enterprise.
"Understanding your customer is just as important as getting your act together regarding your product information," said Aaron Zornes, founder and chief research officer for the MDM Institute. "So your customers could be your distributors, your end consumer, whether it's a business or consumers themselves. For example, Caterpillar is keeping track of its distributors. Michelin is keeping track of its tire distributors, and Pepsi and Coca-Cola are keeping track of their bottlers as well as their suppliers. So customers are still a vital piece -- it's not just PIM."
So while many manufacturers have an underlying assumeption that their master product data and supplier data are under control -- because it's in their ERP system -- other parts of the business, including the back offices, accounting, engineering, or production departments, could be hurting, Zornes said.
In an effort to find the best catch-all application for MDM across an enterprise, manufacturers run the risk of believing that a single solution might be the answer. "Customer data integration (CDI) is also important, though it's not as important as PIM for manufacturing, and organizations need to understand that both are fundamentally different products," Zornes said. Although CDI and PIM solutions are both MDM products, he said, "you probably will not find one vendor that can do both to your satisfaction."
Because most MDM, PIM and CDI vendors focus on one type of data, Zornes warns that companies shopping for MDM software should watch out for PIM providers that claim that, "Oh, by the way, we also can do CDI" -- or vice versa.
"For a while, some vendors were [claiming] that you could do both CDI and PIM within the same software set," Zornes said. "But now they are backing off those claims. A bank has very different CDI needs than a hard disk manufacturer."
Chris Maxcer is a freelance writer who frequently contributes articles to SearchManufacturingERP.com.