Students

Working on master data management? Smart strategists plan for convergence

Chances are, service-oriented architecture and an active data warehouse are coming to your IT shop, if they haven't already done so. According to Michael Goul, professor of information systems at the W. P. Carey School of Business, those technologies will dovetail with master data management (MDM), and smart companies should plan for the convergence that's on the way.

2007-05-23

Not too many years ago, you might have had a cell phone, a digital camera, a pager, an electronic day-planner, a boom box or portable CD player so you could carry tunes on the road And, all of these items would have been separate devices. Today, you can get those functions covered in one slick little cell phone. Michael Goul, professor of information systems at the W. P. Carey School of Business, invites you to consider this convergence and come clean about how many gadgets sit in your personal electronics cemetery — aka a junk drawer — because you purchased various mobile devices to consolidate functionality over the years. "Did you anticipate the convergence?" Goul writes in a recent paper. "Did you pinpoint what type of converged device you wanted— and guess what bells and whistles it would have — before the convergence took place? "If you're like most of us, you selected vendors along the way, only to switch devices and providers a few times as you figured out what you really wanted and needed" in mobile-electronic gizmos, he continues. If you'd followed a prudent path and had a strategy, you probably did a better job than most in transitioning through technologies, he adds. "And, you probably saved a lot of money." Goul uses the mobile-devices example to drive home his belief that companies should consider an upcoming convergence as they plan their master data management strategies. Two important technology developments probably will impact your corporate computing, he says. Chances are, service-oriented architecture and an active data warehouse are coming to your IT shop, if they haven't already done so. According to Goul, those technologies will dovetail with master data management, and smart companies will plan for the convergence that's on the way. Aligning forces   Wikipedia defines MDM as an IT discipline that focuses on the management of reference or master data "shared by several disparate IT systems and groups." According to the online encyclopedia, companies need MDM "to enable consistent computing between diverse system architectures and business functions." That may involve computing between different companies that are sharing information across corporate boundaries. Service-oriented architectures often come into play with such collaborations. In these systems, IT professionals combine small, computerized units of work into larger applications or automated business processes. Airline websites, for instance, may be "mash-ups" of services provided online by the hotels and car rental companies so that you can book your room and wheels at the same time you pick your flight. Likewise, a retailer like Lowe's might harness web services that provide specifications on appliances, so that employees such as the store's kitchen designers can help consumers plug this or that dishwasher into a virtual kitchen and see how it might fit in. Coincident with the proliferation of SOA, Goul notes that next-generation data warehouses are likely to change how people access and use data. Standard data warehouses simply collect data and store them, he notes. Sure, the data are valuable to business-intelligence workers, who may use them for sales predictions, pricing models or other corporate calculating, but such activities happen outside the warehouse itself. In an active data warehouse, intelligence will be encoded into business process so that people can start using it at the right time, precisely when they're making decisions. Going back to that Lowe's example, suppose a consumer is working with a kitchen designer and picking new appliances. An active data warehouse might automatically deliver an alert that the dishwasher the consumer is considering has a new, upgraded version with enhanced features. In that way, the ADW helps the retailer and manufacturer up-sell the consumer by delivering the right information at just the right time — when the consumer is making a purchase. Not surprisingly, active data warehouses are converging with service-oriented architectures. Goul feels that "in the context of convergence, it is appropriate to adopt a broader, more strategic definition of MDM." He maintains that MDM should be defined as an enterprise "service," rather than a set of processes. The service scrubs the data and delivers it with enough semantic relevance for subscribing people or systems to use it with near-real-time speed. He also feels the convergence of SOA and ADW should include master data management. What's more, MDM strategy should reflect these emerging IT influences, as well as the organizational structure as a whole. Go configure   When it comes to organizational structures, most business people are familiar with value chains, where several organizations collaborate in standard product-manufacturing or delivery lifecycles. In a value chain, a retail store's inventory-tracking programs might detect an impending shortage of corn flakes and automatically order more. When that happens, a warehouse operation's system might need to order additional corn flakes from the manufacturer. The manufacturer's systems then may have to search for raw supplies, order missing ingredients, schedule corn flake production, track packaging availability and so on. Web services can facilitate such automation by allowing one entity's system to talk to another. For instance, the store might discover the warehouse has a shortage of corn flakes, too. Then, its system could automatically search for another supplier. Or, if the cost is too high, perhaps the system would search for a more affordable supplier. After all, things like cost and product quality deliver value in a value chain. But not all businesses create value through such "chain" reactions. As Goul points out, some businesses are value shops that center around problem solving. Legal-, medical- and accounting-service firms are examples of value shops that make money providing professional services. In such businesses, value relates to reputation, not necessarily service cost, Goul says. Along with value chains and shops, there are value networks, where the business must continually add new members, who gain from their affiliation with that network. Goul cites cellular service providers and insurance companies as examples of value networks. He explains that with insurance, for instance, members gain value through shared risk and through reduced costs negotiated by the insurer. Such business designs contribute to Goul's strategy for MDM planning. According to him, certain value configurations call for greater investment in specific IT functionalities and, if you can't manage that triple convergence at once, you might want to prioritize. Best two out of three Goul admits that, in the best of all possible worlds, a company would consider the convergence of MDM, service-oriented architectures and active data warehousing simultaneously. "Since these three things are converging, you can invest in each one independently," with separate employee teams, he says. "But if you do, you'll end up with a lot of knowledge about each technology and no knowledge of how you're going to integrate them when the convergence really hits." What's more, you'll be grappling with platform proliferation, as well as the time and expense wasted on managing redundant computing systems. Yet, looking at this trio of influences simultaneously will be a challenge because no one really knows how these emerging factors will look in the future, Goul maintains. Consequently, "You might want to look at binary combinations to begin with," he says. And, his paper on master data management strategies outlines binary synergies that might be appropriate for specific industry orientations or businesses. One example he uses is aircraft manufacturer Boeing. Goul explains that the company is undergoing a major business transformation that centers on collaborative business processes. "Instead of having myriad parts, built to complex blueprints and then delivered for assembly, partners are designing parts and examining their fit in what is referred to as 'virtual assembly,'" Goul writes. "Testing occurs in a complex, computer-based model that resides outside Boeing's firewall. Partners are concurrently building different aircraft pieces by creating data that is assembled and checked in real time." SOA helps engineers know when they have designed a part that won't fit, according to Goul. That is, a computer model virtually tests parts compatibilities and manages a standard value-chain operation. Aiding this process would be a context-aware MDM that helps translate data such as parts numbers or specifications so that the information is accessible, understandable and relevant to all the players involved. A different binary combination might be right for the business-intelligence group at American Express, Goul says. There, MDM services handle the necessary semantic transformation that will make information usable to various professional groups. Active data warehousing is key to providing the groups with timely performance information, he adds. At Safeco Insurance, a value network, Goul sees all three IT functions operating harmoniously. Service-oriented architecture already facilitates an electronic quoting system for the 9,000 agents and brokers who sell the insurance. The MDM services can help the company manage the policy and regulatory updates it faces. Once added, active data warehousing can improve up-selling opportunities and support agent decision-making during client interactions, he maintains. He also believes that in addition to considering which IT capabilities to focus on first, smart corporate managers would be wise to remember that MDM middleware "has more flexibility to evolve as convergence takes place." So, if you pick an MDM and then get service-oriented architecture, the MDM probably "can evolve to SOA better than the other way around." At bottom Goul stresses that an MDM solution doesn't operate alone, so it makes sense to consider the alignment of MDM strategy with the two complementary developments of active data warehousing and service-oriented architecture. The convergence of these three functions is coming, just as surely as cell phones will pick up more gee-whiz features in the future. "These elements represent a trifecta of influences that need to be aligned with organizational strategy," Goul concludes. Bottom line

• Corporate computing professionals face a coming convergence between service-oriented architecture, active data warehousing and master data management.
• Focusing on these IT functions independently could lead to wasteful platform proliferation and complicate integration efforts in the long run.
• Tackling all three functions at once may be a challenge. Looking at binary combinations of the three functions might be a smart starting point.
• The value configuration of your business — the way you make money — will determine which duet of functions to focus on first.

KnowIT