Demystifying Master Data Management

Years ago, a global manufacturing company lost a key distribution plant to a fire. The CEO, eager to maintain profitable relationships with customers, decided to send a letter to key distributors letting them know why their shipments were delayed--and when service would return to normal. He wrote the letter and asked his executive team to "make it happen." So, they went to their CRM, ERP, billing and logistics systems to find a list of customers. The result? Each application returned a different list, and no single system held a true view of the customer. The CEO learned of this confusion and was understandably irate. What kind of company doesn't understand who its customers are? Unfortunately, most companies don't have a precise view about their customers, products, suppliers, inventory--or even employees. Whenever companies add new enterprise applications to "manage" data, they unwittingly contribute to an enterprise-level confusion about a corporation's overall view of itself. As a result, the concept of master data management (MDM)--creating a single, unified view of an organization--is growing in importance. Naturally, where there's a big technology problem, there's a host of vendors lining up with sophisticated technology to provide an "out-of-the-box solution." The rationale is that by plugging an MDM technology into existing applications (CRM, ERP, logistics, billing, etc.) you can build that one "true view" that will then feed consistent, accurate, and reliable data back into these systems. But, as early adopters are finding, MDM isn't just a technology effort. Although MDM technologies can have a dramatic effect on a company's performance, no software is going to magically solve your enterprise's data problems overnight. And the root of this problem is poor-quality data--and the manner in which it's created. Consider a recent report from The Data Warehousing Institute that found 83 percent of organizations suffer from bad data for reasons that have nothing to do with technology. Among the causes of poor-quality data were inaccurate reporting, internal disagreements over which data is appropriate, and incorrect definitions rendering the data unusable. Organizations must understand that improving their data--and building the foundation for MDM--requires them to address internal disagreements and broken processes. Staff must agree on exactly what constitutes a "customer" or a "partner," and how to resolve any disagreements across business units. Departments and divisions need to agree on hierarchies of customers and products and how to resolve duplicate records across sources. Rather than a technology-focused effort, the project becomes one of political-strategizing and consensus-building. For any company contemplating an MDM strategy, it's important to start small and refine an approach over time. The typical enterprise has massive amounts of data spread across multiple databases and departments. A pilot project on a subset of applications can show you what works--and what doesn't--when integrating data from multiple sources. However, merging this much information cannot be done in a vacuum. The best way to gain control of your data is to work directly with I.T. analysts, data stewards, and data owners who understand where data is held--and where problems exist. It's important to set up a task force--often called a "Center of Excellence"--to guide the MDM project, with both I.T. and business users providing guidance and support. This gives everyone a sense of ownership and responsibility while also creating an approach focused on best practices across divisions and departments. The Center of Excellence can define the parameters of master data and institute control policies to regulate who is allowed to input and change system data. Well-defined parameters will ensure you know what to look for when you review your data, and will help determine the standards for inputting it, going forward. Only after you have these critical definitions can you properly deploy MDM technologies. By providing a single version of the truth, MDM can have a decidedly positive impact on your business. You will know exactly what products your customers have, what items you buy from selected vendors, and how much inventory you need to meet customer demand. And if your CEO needs to send a letter to customers, don't sweat: The truth will (finally) be in the system. About the Author Tony Fisher joined DataFlux as president and general manager in 2000, and is now its CEO. He has guided DataFlux through tremendous growth as the company has become a market-leading provider of data-quality and data-integration solutions. Tony is also a prominent speaker and author about emerging trends in data quality, data integration, and corporate compliance efforts. Prior to his post at DataFlux, he was the director of data warehouse technology at SAS.