Data management has another master
New solutions market aims to create structure for information-based initiatives
By Cole Ollinger, contributing editor -- Manufacturing Business Technology, 1/1/2006 12:00:00 AM MST
Master data management (MDM) is a bold name for what actually happens when IT vendors blend existing principles with a few subcategories and specialties—this time, data integration and standardization among them—and thereby announce a new solutions category. But MDM technology, as a concept, seems to resonate with the prospective audience. A recent survey by Boston-based AberdeenGroup reveals 72 percent of manufacturers currently practice some form of it, or are actively planning to do so within the next two years.
Generally accepted definitions of MDM include merging various enterprise data into a single "system of record" or "one version of the truth," as regards customers, orders, products, raw materials, or suppliers. MDM automates data transfer and reconciliations between repositories where such information is stored—and among the applications that capture and use it—while ensuring the highest levels of data quality and standardization.
The truth is MDM means different things to different people.
According to Boston-based AMR Research, "The real goal of MDM is to create a structure for information management that lets the business effectively and efficiently use this information for business value."
Meeting that goal involves a huge range of initiatives, systems, and IT assets, including data warehouses and data marts; data cleansing and standardization efforts; customer data integration (CDI); EDI; product information management (PIM); and metadata management. MDM also is a discipline relevant to ERP, CRM, supply chain management, and BI systems, as well as application development environments.
Take CSI Wireless, a Calgary-based maker of equipment for wireless networks and global positioning systems. Many of its managers rely on three types of data. Day-to-day information is pulled from test stations on the shop floor. Historical information on production operations is stored in data marts. And data that gives executives an overall view of how efficiently all plants are running at any given time is available via reports or dashboards. CSI wanted MDM to ensure contract manufacturers were meeting the very tight specifications of its products.
"We want to test the shop-floor information to ensure we're making and shipping quality products," says Mike Pratt, director of information management services. "We also want to catch issues proactively to ensure that specifications are not wandering. And to avoid product returns, we want to identify any potential problems within the production cycle."
Pratt looked at solutions ranging from those offered by large ERP companies to product documentation vendors before choosing the SigmaSure solution from SigmaQuest, whose products are closely aligned to CSI's core needs on the shop floor. SigmaQuest uses MDM principles to deliver what it calls product performance intelligence—or PPI.
According to SigmaQuest CEO Nader Fathi, SigmaSure extends manufacturing execution systems, allowing companies to collect feedback about how products are designed, tested, and built. Production data is captured on the plant floor, then matched and synchronized with engineering data. The system then "calls home" so executives and engineers can monitor ongoing production and compliance with specifications, performance targets, and regulations—for plants around the world.
"PPI gives OEMs and brand owners global visibility and control during the entire life cycle, from design and testing, to manufacturing and post-sales service," says Fathi.
Not the norm
Like many IT veterans, CSI's Pratt has a healthy skepticism about the latest IT flavors of the month. But the company's efforts with SigmaQuest were successful enough that Pratt expects similar MDM tools will be deployed across other functional areas, business units, and product lines. "SigmaQuest generates the data and feeds it up into data marts, but it also has the analytic capabilities we need," he says. "It helps us build trust with our manufacturing partners."
Bill Swanton, an AMR Research VP, believes companies like CSI Wireless that take on independent MDM initiatives are somewhat atypical.
"Usually, master data management initiatives will be part of other large-scale initiatives, like strategic sourcing, shared services, or a CRM program designed to present one face to the customer," says Swanton. "This is especially true for companies with complex technology infrastructures. Unifying material numbers and 'de-duplicating' data have been standard features of ERP consolidations. Figuring out ERP standardization is largely a function of figuring out MDM. If you don't have pure, clean data in your ERP systems, you won't get the full return on your investments, especially if you're consolidating systems."
Swanton estimates 30 percent of calls to help desks about enterprise applications are traced to incomplete or incorrect master data—not problems with the software itself. And it's more like 80 percent for closely related enterprise applications that must use the same master data, like SAP's R/3, and Advanced Planner and Optimizer solutions.
Those figures shed light on why MDM "has been one of those neglected undertakings in IT, according to William Mougayar, a VP and service director with AberdeenGroup.
"The larger the organization, the more difficult MDM is on an enterprisewide level due to sheer complexity of coordination required, and the need for multiple business units to agree on a common terminology," says Mougayar. "Another challenge is that several monolithic software applications include their own mini-MDM features, so it becomes appealing for users to take a shortcut to the exercise within the confines of stand-alone projects."
Beyond ERP
In addition to ERP consolidations, a mix of external business drivers—e.g., the globalization of manufacturing, the rise of contract manufacturing, and more complex reporting requirements—are fueling the MDM market and the specialization of solutions. Technical factors such as the drive toward service-oriented architecture (SOA) also play a role in MDM market growth.
Data in legacy systems must be made ready for use before an SOA can truly take hold, says Dave Hollander, CTO of Contivo, which offers metadata management and semantic integration solutions. "Data must be decoupled from processes and platforms and then broken down into elements before it can be translated and shared for use by JAVA or XML scripts," he says.
Establishing standards for even relatively simple data elements can be tricky, however. Hollander cites country-of-origin data as incredibly complex due to different regulations and definitions across countries or among business partners. "Web services can automate the lookup, but XML schemas are needed to drive the interfaces to ensure all relevant relationships are factored in and the right data is used," he says.
AMR's Swanton cites materials data management as another deceptively complex task. Adding a new record may require data from up to 20 nodes, meaning automated workflow capabilities are an important feature of MDM tools. "If that process isn't automated based on clear dependencies, the risk of bad data goes up exponentially," says Swanton. "And you can't solve these problems manually, because there simply aren't enough resources."
Who's who in MDM
Before MDM became a widely recognized IT discipline, many companies already were practicing it, using ad-hoc tools built in programs like Lotus Notes from IBM or Microsoft Excel. Companies with large IT budgets and lots of resources for custom development—particularly financial services firms, telecommunications companies, and energy conglomerates—built robust, home-grown MDM solutions long before IT vendors recognized the need. Today, most of the big IT players are selling MDM in some form. Oracle, SAP, Siebel, and IBM all offer some MDM capabilities and tools, and are likely to round out their offerings through acquisitions. Yet while the analyst community continues to assess the size of the MDM market, its growth is being driven by small and midtier solutions providers offering targeted solutions for specific industries or functions—i.e., CRM, supply chain management, or product life-cycle management. In particular, Siperian's CDI tools have been taken up by many pharmaceutical companies, while Informaticamakes a distinction between MDM applications, which it does not sell, and the underlying "data integration foundation."
Other companies have built MDM tools based on client need. Case-in-point is supply chain management vendor i2 Technologies, which got into the MDM market when users asked for it as part of larger implementations. According to Satish Krishnaswamy, senior director with i2, "The drivers for our clients include global data synchronization with retailers and compliance issues, but strategically we believe MDM is critical to moving to an SOA."
Krishnaswamy adds that i2 created a business for enterprise MDM solutions, including some that integrate and synchronize customer data.
"CDI and PIM are important to manufacturers in terms of customer and product hierarchies," says Krishnaswamy. "Suppliers often are customers, too, and manufacturers must recognize their customers' customers. Establishing clear relationships based on clean data, PIM, CDI, and overall MDM strategies can help companies sort through these challenges."
At the moment, there is no single solution likely to cover all of a manufacturer's needs. "If you're doing order management and manufacturing, you'll need ERP and data warehousing," claims CSI's Pratt. "If you're doing design, you'll need product documentation, and if you have reams of parametric data, you'll need some sort of tool to capture and analyze production data. There is hand-off between the data sets and extractions, but you need a unique solution for each level."
Involve the business
To address the "much more insidious issue" of maintaining data cleanliness and quality over time, Swanton recommends involving business owners.
"MDM isn't just an IT project," says Swanton. "Establishing accountability among business owners is one way to ensure the data goes in clean the first time, remains clean, and doesn't degrade over time." Hard metrics around data quality can ensure data management remains a high priority, he adds.
Mougayar recommends creating a senior VP of information and data management. "The position is critical, because information management is a key competency that requires nursing," he says. "You can't just design a metadata schema and let it run its course without constant adjustments. The inherent organizational changes that happen don't allow you to rest on your laurels."
That's the market reality for manufacturers looking at MDM, but the good news is that because the market is likely to remain fragmented, companies can call on providers that will be willing to tailor solutions just for their needs. Of course, they may have to learn a few more acronyms along the way.
MDM services adoption
| All industries | Manufacturing | |
| Source: AberdeenGroup |
||
| Currently use | 23% | 27% |
| Plan to use within 12 months | 27% | 26% |
| Plan to use in 12 -24 months | 27% | 19% |
| No plans to use | 2% | 6% |
| Don't know | 21% | 21% |
Featured Company
Most Recent Resources
- From CAD Systems to Customers - New collaboration capabilities bring...
- Real Time Energy Management - FREE White Paper and Video
- Deployed in 45 days, Zebra Enterprise Solutions Material Flow System...
- How to Eliminate 7 Wastes from your Supply Chain with Real-Time...
- Is Your Website Really Winning You Business
