Master data management (MDM) is an IT discipline that is becoming a world unto itself, despite being a relatively nascent one, compared to stalwarts customer relationship management (CRM) and enterprise resource planning (ERP). Like these other disciplines, MDM isn’t simple. The problem of how to store and use “master” data – the information on which businesses and their business decisions are based – raises thorny questions about strategy and execution, implementation, governance and long-term return on investment. With so many organizations looking at MDM as a critical component of their information management strategies, businesses need a better understanding of what MDM can actually accomplish.


The discipline of MDM was born just a few years ago, when technologies emerged to meet companies’ need to keep up with the rapid pace at which business data changes. Of course, the need for MDM isn’t new. Since the beginning of time, people have been moving to new homes and forging new business and family relationships. The difference today is that those moves and all the changes that go with them – new phone numbers, jobs, credit cards, bank accounts and the like – affect the accuracy of information in countless databases. If these moves and other changes are not noted in a business’s data records and shared with all the systems that depend on that information – a company can pay a high price.


Not surprisingly, many businesses are trying to solve this problem by establishing a single view of their master data. But there’s a flaw in the single view approach to MDM. Given the organic way in which most companies’ information systems have been built – not to mention the reality of information silos, legacy systems and multiple architectures – achieving a single view of master data is like hunting for the Holy Grail. As companies continue to reorganize, merge, make acquisitions and grapple with security and compliance regulations, their need for MDM solutions cannot wait.


At the same time, it is sometimes said that MDM projects will create more silos – precisely what we all want to avoid. The emergence of technology that unlocks data silos gave rise to a bustling market for software that addresses key categories of business information – customer data integration (CDI) and product information management (PIM) solutions are the most popular of these. These software products bring together master data that is distributed across numerous self-contained systems such as customer service, manufactur­ing and procurement. This is important, because even within a single business unit, there are often multiple front- and back-office systems that house critical business data that cannot be shared across the company.


This compartmentalization of data within often incompatible systems is what makes a single view of master data impossible to create and maintain. CDI, PIM and a number of other solutions that consolidate data from across an organization now threaten to create master data silos.


Avoiding the creation of new silos requires a shift in thinking. Instead of focusing on that elusive single view of their master data, businesses need to consider a “right” view MDM strategy. This means focusing on delivering the right view of data for the right audience. It requires technology that can ensure that the data a view relies upon is correct, accurate and relevant. Unlike the simple data aggregation projects of the past, which led to the creation silos of data used by a particular division or group, a right view MDM solution breaks down the barriers between data silos, allowing existing information to be presented in multiple views so that data is delivered based on the needs of the individual viewer.


As such, MDM has to be treated as a strategy, not a software purchase. And solution providers must address multiple views, domains and usage styles of their master data. Before organizations adopt a master data strategy and attempt to consolidate multiple master data types (customers, accounts, products, locations, etc.) in one central location, they must first understand that the key word in MDM is management. There are many different users of master data in an organization – from operational business applications, to applications and internal users focused on the collaborative process of creating data, to existing analytical users like data warehouses and analytics tools. MDM must serve all of those users and offer functionality that addresses their unique requirements – operational business services for real-time data management, collaborative workflows for data creation and definition, and event-based analytical triggers to integrate with existing analytics environments.


The operational style of MDM refers to a situation in which the use and maintenance of master data occurs within operational processes or applications. The master data is used by other systems via real-time, accessible service-oriented architecture (SOA) services. Collaborative MDM focuses on managing the process of creating, defining and synchronizing master data. At the heart of this style of MDM is the definition of master data that can be synchronized with operational and analytical systems and applications. Finally, analytical MDM provides accurate, consistent and up-to-date master information to data warehouses, and enables business intelligence (BI) insight to be fed back into collaborative and operational MDM.


Simply put, master data is defined, created, operationalized and analyzed; the ways in which this is done varies. Solution providers must address all types of data, users of data and data lifecycles.


There is a lot of talk about MDM “approaches.” Gartner, Inc. has effectively defined these and much of that categorization is based on vendor solutions introduced when the market for MDM emerged several years ago. Interestingly, those approaches, whether it was registry style, third-party data providers, application specific solutions or so-called transactional hubs, very often represented one piece of a multiphased approach. But few of the implementations today will start or end with just one style.


A typical MDM implementation begins with a solution aimed at addressing a data issue that is affecting a bigger business process. For example, if a customer information file can no longer house or support some data elements that are required for the business, a new data element, such as a privacy preference or location data for a customer, arises; or, separate businesses are brought together, creating new requirements for views of data and data reporting.


Companies will opt for either a purchased or custom-built solution to tackle the immediate data problem, but they tend to adopt a phased approach, or build onto the solution over time to meet bigger needs in the future. In some cases, companies start with a cross-enterprise vision for MDM. While they have plans and blueprints, they too take a phased approach to execution.


Most of these initial projects evolve into more significant MDM endeavors that touch many parts of a business. Projects that do not expand in scope do so as a result of lack of vision – often because of a change in leadership.


There are a variety of approaches an organization can take with an initial MDM project. A basic clean-up usually involves the initial clean-up and standardization of data to maintain a small subset in a registry for use when needed. An organization may take that a little further by incorporating third-party information to enhance and update their own data for use with other systems. A minimal persisted data hub will consist of a small subset of data. This may be linked to an initial project and develop over time into a more robust master data hub. Transactional hubs rarely are the initial entry point to MDM, despite the fact that many people perceive them to be the best way to derive long-term value and flexibility from an MDM solution.


Existing systems interact with the new MDM solutions in a number of ways. Customer information files and other systems of record may stay intact permanently or for a set period of time, after which they are sunset or repurposed. The option to have various applications interfacing with an MDM solution will likely also be phased, so not all applications will be integrated immediately -- just what makes sense for the business. Multiple MDM vendors may be implemented together in an environment for a period of a time, handling specialized areas before a consolidation phase begins. There are many scenarios across various industries. Organizations need to consider what the best solution is for their individual needs and draw up a plan that addresses an evolutionary approach.


As they do this, businesses must ask themselves what kind of ROI (cost take-out, for example, as opposed to revenue generation) they are seeking, both for the short term and the long term. Will the expansion of the project depend on showing this initial ROI? Do they have a vision to make this an enterprise MDM solution? Are there other projects that will tie into the MDM project down the road? And how many resources - if any - will the organization have to build and maintain today and in the future?


Considering these questions will help to better establish the right approach and determine what the phases of the project will be. It is important that businesses do not fall into the trap of choosing a solution that will not be adaptable to the future requirements of the business. There is always greater risk and cost associated with putting in a solution to meet immediate needs rather than putting in a solution to address what your needs of tomorrow will be.


When planned carefully, a right view MDM solution can provide all the benefits of traditional data management solutions - and the potential for enhanced, enterprise-wide efficiency and competitiveness. The prospect of putting a MDM solution in place may be daunting, espe­cially for larger organizations or those with limited IT resources. But with the right-view approach, this does not have to be the case.


That’s because at its heart, a right-view solution to MDM is a busi­ness strategy implemented via IT deliverables. Once the business strategy is in place, organizations can implement the IT components one domain at a time, as their overall business priorities call for them. The staged implementation process can also be coordinated with - and enhanced by - other initia­tives, such as implementation of an SOA.

For the bottom line, this staged-implementation approach allows the financial impact of a new solution to be absorbed over a longer time and tied directly to specific business benefits. Staged implementation also allows an organiza­tion to see the real-world benefits of MDM more quickly, which is a good thing. Without an MDM system in place, organizations everywhere will continue to be at the mercy of their information silos.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access