Master Data Management: Changing the Information Landscape

And necessary – albeit complex – business strategy for effective data management. By Leslie Knudson

The issue at the heart of the data management problem is nothing new: as growing volumes of data overextend information management capabilities, the quality of that data keeps plummeting as organizations try to piece together remote and disjointed information that is often conflicting and, eventually, of no value.

The disconnect that results from the silo practice of information storing has become too prevalent to ignore and the business consequences – including poor customer service, business moves based on false assumptions, incongruent decision-making and so on – too severe to sustain. While businesses have participated in data governance for many years, the need for a broader initiative to govern entities that go across the enterprise (such as customer, product and supplier) has become decisively apparent.

What has thrust master data management into the limelight is its far-reaching and overriding ability to synchronize operations and dissolve the level of inconsistency that might exist between internal systems and applications. This means that the human resources department can possess the same, cleansed data as the sales department and the marketing department, ensuring optimal customer service, synchronized business moves and internal consistency.

Getting technical

The number of information catchphrases floating around the industry has created a tendency to lump related terminology (such as data governance, reference data and applied master data management) into a single, fuzzy, gray area – which is why some incorrectly assume master data management is just a fancy term slapped onto an old practice. Consequently, understanding the explicit definitions of, and distinctions between, these terms is key to executing them properly.

Simple data governance pertains to the formal delegation of information ownership involving a grouping of people, processes and technology working to utilize data as an enterprise asset – and may include a hodgepodge of a governing body, assigned data stewards and outlined communication processes. But the growing volume and complexity of data has created the need for organizations to take a more strategic stance to break down information barriers existing between internal divisions and obtain a single view of master data.

While reference data persists across transactions without being dependent upon other identities for identity or meaning, master data is reference data for which there is an agreement on a single view, meaning that reference data can be shared consistently across applications.

“Someone once said that the difference between reference data and master data is that master data is reference data with consensus,” says Henry Morris, Group Vice President and General Manager for IDC’s Integration, Development and Applications Strategies solution research group. “There’s an organizational component that requires people to come to an agreement on a particular view of customers, products and those things that go across business units and applications.”

Master data management, then, is the set of processes involved in creating and maintaining a single view of master data through one central hub. True master data management extracts rules and information out into the horizontal layers and then replicates the information back to individual units. The broad umbrella of master data management encapsulates a number of applied master data management areas, including customer, product, location and financial account.

The overall market for master data management software includes both the applied master data management software and master data management infrastructure software. Applied master data management software includes categories of applications specifically geared towards managing distinct classes of master data, such as product or customer, while the master data management infrastructure software supports the overruling processes required to establish and maintain a policy hub for master data.

Key business drivers

Compliance is arguably the biggest business driver behind the proliferation of master data management software (data quality, integrity and security rank high on executives’ priority lists to help stay in line with strict regulatory guidelines), with efficiency running it a close second.

Besides reducing regulatory risk and improving compliance agility, organizations are also looking for consistent data across the enterprise to ultimately improve relational tasks such as marketing effectiveness and customer service. Organizations seeking to satisfy reporting needs and create improved operational efficiencies – especially in the context of retail or supply chain industries – are turning to master data management as well.

Another influence driving the increased focus on master data management is that ownership of master data is shifting to high-level business executives rather than resting solely on the shoulders of IT. Because master data is now being viewed as such a valuable asset and critical resource for efficient business operations, data quality and integration issues have been escalated to upper management concerns.

All these drivers aim to reach the essence of master data management, which is obtaining a ‘single version of the truth’ – a concept that is nothing more than the ability to harmonize data across all internal units and applications so information remains uniform and consistent. While this idea of searching for a ‘single truth’ sounds more like a fantasy sci-fi expedition, the concept of achieving a unified view is a very tangible aspiration for businesses. Reaching this level of truth entails consistent, real-time information about customers, products and locations across all systems and divisions.

Mapping a master data management strategy

Determining the best master data management strategy begins with each organization first defining the objectives they wish to achieve and, ultimately, what level of inconsistency can be tolerated across internal processes.

According to Morris, IDC has categorized master data management projects into three main scenarios: management reporting, data synchronization and single point of origination. These scenarios vary in ascending degrees of complexity in relation to the core objective of each. The management reporting scenario is typically implemented by organizations looking to solve management reporting issues and it involves reconciling master data at the hub to drive reporting for compliance and business performance management issues. This scenario poses the least difficulty and will suffice for organizations simply wishing to ramp up their reporting.

The data synchronization scenario seeks to build consistency in operations by synchronizing master data from the hub back to local systems. Because this scenario permits master updates to occur, it allows local systems and system owners to retain a level of autonomy – which is ideal for a customer master data project where locally relevant attributes exist and wouldn’t be updated via a central hub.

The single point of origination scenario is by far the most challenging strategy because it involves the most fundamental changes to business processes. With this scenario, all master data changes originate at a single hub rather than having master data changes occurring across an organization. However, organizations employing this master hub approach may see performance problems surface as a result of record creation flowing through a single point, which will inevitably create a bottleneck.

“I would advise just looking at the business purpose you are trying to achieve,” says Morris. “If you’re trying to solve a compliance problem then the first strategy might be fine. If you’re trying to establish a strategy so you can deal with coordinating marketing activities, synchronization is probably fine because you can live with some level of inconsistency. If you get closer to real-time operations and want to work around the call center and real-time dialogue with customers, then you’re getting closer to thinking around single point of origination.”

Implementation challenges

Granted, implementing any of these master data management strategies is no easy task. Based on which data management strategy is right for a particular organization, there are varying degrees of difficulty dependent upon each approach. Businesses with a history of mergers and acquisitions or global organizations with independent units under unique cultural or regional structures will encounter an added array of complexities.

Embarking upon a master data management initiative requires much more than momentum and the right software. Synchronizing views across multiple applications and systems requires gutting out existing processes, and may possibly include a severe transformation in terms of individual and departmental ownership, access and control of data. Due to the organizational component inherent in delegating data governance roles and relinquishing control of data from one department to another, master data management services are much more likely to flourish than master data management software.

“Usually the biggest problem that people have is around the organizational issues,” Morris says. “You’re making one system the only one where you can introduce new products, and if there are several in the organization, someone is losing control. So the ability to resolve those differences in the organization will obviously require some attention at a fairly senior level.”

While it’s common to hear people warning against a ‘big bang’ approach, converting an organization’s gamut of internal information management processes at once is virtually impossible anyway. To successfully approach a master data management project, organizations should proceed one step at a time – taking into consideration key business objectives and key performance indicators, gaining the support and participation of relevant stakeholders and determining the scope of systems that will be included.

It’s highly recommended to limit the scope of the initial deployment by beginning with a pilot project and then progressing in carefully planned steps. Likewise, it’s also essential to incorporate training and consulting services in the early stages of a master data management project.

Forecast for the market

It is predicted that market demand should see an explosion of master data management services and software over the next five years. According to IDC, the market for master data management reached US$5.4 billion in 2004 and should grow to US$10.4 billion in 2009, a compound annual growth rate of 13.8 percent.

The vendor market is already becoming lightly saturated with master data management software as produced by the likes of key vendor players such as Hyperion, Informatica and IBM who has built upon its master data management offerings through several recent data-management acquisitions.

In the software arena alone, IDC forecasts master data management software revenue to grow from US$2.6 billion to US$4.6 billion over the forecast period (2004-2009), a compound annual growth rate of 11.8 percent, and master data management services revenue to grow from US$2.8 billion to US$5.7 billion over the forecast period (2004-2009), a compound annual growth rate of 15.6 percent.

The bottom line is master data management is certainly not heralded as any magic pill or quick fix, but should be viewed as a growing body of best practices that when carefully leveraged can drastically transform an organization’s information management capabilities and, in turn, significantly enhance operations and revenue.