The MDM Institute predicted that in 2008, the master data management (MDM) market will continue to transition from early adopter to mainstream, with more than 95 percent of some industries actively exploring replacements to their homegrown MDM solutions.1

 

As with all hot initiatives, there's often a rash of hype and interest from vendors looking to jump on the bandwagon. Suddenly, every vendor has an MDM success story to tell. Analysts and vendors contradict each other as they highlight which route they think you should take.

 

When overwhelmed by market information, companies need to evaluate the needs of their organizations and tailor an MDM program that will best meet those needs. Before you even begin to look at vendors and technology, you need to appreciate that MDM is a long-term program - a corporate discipline - not a one-time project. Think of MDM's function as similar to the finance function in your organization, except it manages data assets rather than financial assets.

 

Getting Started

 

First, you need to set an expectation: MDM will never be easy. No amount of technology will make it a turnkey operation. There is no single approach that will work for all - or even for most. By nature, MDM needs to be highly customized to the needs of your organization. And while IT organizations have to be involved in the adoption of MDM, the content of the data is inherently the responsibility of the business. So the real challenge is - how do you get started?

 

First, stop worrying about implementing "master data management." Take stock of your individual environment and business needs, figure out what will make the biggest impact on your organization and focus on addressing that problem.

 

Recognize that regardless of where you start, you'll need to address inherent data governance issues. Unlike automated data quality, master data requires human judgment (and therefore governance) at some level. A business can tackle a small piece of the data in the organization - finance data, perhaps - to learn how to develop data governance practices and use MDM tools to effectively manage the process in a way that works for it.

 

You hear a lot about the different kinds of MDM, but there are really two primary use cases: MDM for operational purposes and MDM for analytical purposes. Both manage master data, and both seek to get a single view of the data. However, they differ in where and how they plan to use that golden copy of master data.

 

Operational MDM

 

The focus of operational MDM is to make sure that data in multiple operational systems that should be the same is actually the same. The goal is to synchronize operational systems data so that you have consistency at the front end, such as in customer-facing systems, which is particularly important for organizations with a lot of customer contact.

 

Companies that favor adopting an operational MDM approach first are often focused on improving their operational efficiency or creating synchronized records. They may have a customer service issue, where an individual customer's account information is scattered across multiple systems, leading to a disjointed customer experience. Or they may be concerned with process consistency and are seeking to reduce process errors (such as delivering the wrong products to the wrong location). They may be actively looking into a service-oriented architecture (SOA), which needs consistent master data to work properly. Or they may be looking for better supply chain consistency in working with partners and throughout their product line.

 

In theory, regardless of the primary driver, once a company has synchronized their master data across all operational systems, they won't need to harmonize that master data for business intelligence (BI) because the data will already be consistent. In practice, however, MDM is such a new market that few, if any, companies have truly synchronized all their systems, and to get to this stage will take many years, probably decades. Most companies have smartly tackled a small portion of that to increase the chances of success. As a result, many who wish to deploy a holistic MDM program find that they need to tackle components of the solution separately in the near term.

 

Organizations need to take into account two important considerations when embarking on operational MDM. The first is risk. Operational MDM requires significant customization and custom coding of your operational systems in order to enable the synchronization. A bit like open heart surgery, this process can be costly and time-consuming, and the stability of the operational system is at risk from such changes.

 

The second consideration is breadth. Because operational MDM tends to be implemented to address a critical subject area (usually customer or product), operational MDM tools are specially designed to provide a great deal of depth in that one subject area - a bonus when it comes to customizing attributes and designing the master data itself. However, that depth also means that operational MDM tools sacrifice breadth, and companies that seek to manage more than one type of master data will find their customer data integration (CDI) tool ill-suited for other data like finance, product or geography. Few solutions are available for operational MDM that are not customer- or product-oriented. It will take many iterations over multiple years to cover all master data subject areas for a company.

 

Operational MDM can deliver significant gains in the form of operational efficiencies and process consistencies, but companies should recognize the potential risk, costs and time to enterprise-breadth data management of such a program and seek the executive sponsorship and funding required to make it a success. In short, the program should be treated in the same way development of a new operational system would be treated.

 

Analytical MDM

 

In contrast, analytical MDM focuses on improving the quality and accuracy of BI reports, enabling managers and executives to act upon better information when making decisions. Companies that use an analytical approach to MDM first need to manage more than one kind of master data, such as financial, customer, location, product and supplier data, so that they can deliver analysis on a cross-section, such as customer profitability by product. To get this breadth across multiple subject areas, companies may need to sacrifice out-of-the-box depth within a single subject area. Analytical MDM tools won't have the same product detail off the shelf as an operationally focused product information management (PIM) tool and therefore may need to be more customized. However, unlike a PIM product, analytical MDM products can work equally well with other subject areas with the same amount of effort.

 

Those that tackle the analytical approach first are often focused on maximizing the effectiveness of BI systems and want to see the MDM project deliver benefits in a short period of time. Unlike operational MDM, analytical MDM does not try to synchronize the master data in the operational systems. Instead, analytical MDM seeks to create a harmonized golden copy of data in a repository, mapping the data and correcting inconsistencies in the master data repository rather than in the operational systems. Not having to deal with costly and time-intensive synchronization processes results in a shorter time to value at less risk because you aren't seeking to change the source systems themselves. In fact, most organizations already have much of the infrastructure in place to do this through their existing data warehouses. They can consider a first step in an MDM program to be an upgrade to the master data flows into their existing data warehouses to include business ownership and governance of the data.

 

Approaching analytical MDM first, organizations benefit from insight offered from high-quality BI very quickly. They can focus on establishing the data governance processes required with effective MDM programs without having to simultaneously tackle the difficult technology task of synchronizing the master data once managed.

 

For companies that need better BI or who are new to MDM, analytical MDM can be hugely beneficial. Generally, analytical MDM costs considerably less, making it easier to get funding approved. The harmonization of master data happens at the analytical layer, where it has a direct impact on business decisions. As a result, higher-level business executives immediately see the value of the program. This makes a stronger case for funding the much more expensive operational MDM program in the future.

 

Regardless of the approach, MDM programs are most successful when tackled iteratively, starting with the data subject area that will deliver high value to the organization in a short time. Monolithic MDM projects encompassing all of a company's operational and analytical systems are inconceivable as projects, hence the market (and its associated vendors) has come up with a virtual potpourri of MDM types.

 

When faced with PIM, CDI, analytical MDM, operational MDM and a wealth of other MDM flavors, companies are understandably confused and overwhelmed, particularly as vendors and analysts advocate one approach over another. The key is not to worry about implementing MDM itself. Remember that successful MDM isn't about the technology at all - it's about the data governance processes required to improve master data. All the products support various aspects of the requirement.

 

Focus on the area where consistent, accurate data will provide the most business benefits. Are you looking for better BI and more accurate reporting? Then you probably want to turn to analytical MDM first. Are you looking to obtain operational efficiencies or gain process consistency across your source systems, and do you have the necessary MDM funding and organizational will? Then you may want to look at operational MDM.

 

MDM in Action

 

Whether an analytical or operational approach is adopted, MDM needs to be a business-led program and should be driven by your particular business problem. For example, after several reorganizations, a global producer of industrial products had a long list of central products along with an even longer list of local products in each of its subsidiaries. However, many of these products performed identically and only differed by name or slight variation. This lack of a single global product portfolio resulted in inefficient purchasing, labeling, documentation and R&D as well as global contracts with big customers that were difficult to manage.

 

To remove these inefficiencies, the company embarked on an analytical MDM project that could provide them with a harmonized source for master data, mapping the local products to the global ones, without sacrificing the local idiosyncrasies and controls put in place.

 

The initiative resulted in a single globally consistent product portfolio for the company. This initiative catalogued all centralized product instances and created a single source of supporting product data - all under a sustained data management process. This enabled the company to better support customer, e-business and competitive initiatives at both the headquarters and the local market levels. In addition, managers’ ability to quickly access vital global product master data helped the company improve the efficiency of its research and development department. Having created a multisubject repository to feed analytic systems, the base data in this repository could then be used as a master to update operational systems as a second step.

 

An important point to note is that the catalyst for this enterprise-wide effort to optimize the company’s product management processes was competitive pressure. As the market for the company’s products consolidated and major competitors globalized, the company realized that it needed to address its master data challenges if it hoped to maintain its competitive edge. This clear connection to improved business results helped secure support for the project from the company’s high-level executives.

 

Regardless of the various choices available, most companies will need to implement both analytical and operational MDM eventually. However, you should recognize that success lies in starting small, establishing governance practices and getting early wins. You can then use that knowledge and the best practices gained from your experiences to improve the second and subsequent increments in the program.

 

Reference:

  1. "Milestones to the Master Data Management Road Map." The MDM Institute, November 12, 2007.

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