Thursday, October 3, 2024

4 Popular Master Data Management Implementation Styles

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Master data management is a technology-enabled discipline to ensure that all shared master data assets are uniform, accurate, and consistent. Master data refers to the identifiers and extended attributes that describe the business’s core entities, including customers, prospects, citizens, suppliers, sites, hierarchies, and accounts—in master data management, business and IT units work closely together to create a single master record for each person, place, or thing in the organization to achieve a single source of trust. This article looks at four of the most popular master data management implementation styles— Registry, Consolidation, Coexistence, and Centralized—and details how they work and when each of them is typically used.

Master Data Management Implementation Styles

While the goal of master data management (MDM) is clear enough, achieving it is easier said than done. Data comes from a wide range of sources, and in many formats. That means there’s always going to be data quality issues and other challenges involved in consolidating, verifying, and aligning an organization’s data.

As a result, there are many different approaches to master data management. Some involve heavy centralization while others take a more federated or distributed mindset. Organizations should select the one that best suits their existing situation, including information architecture, strategic direction, corporate structure, regulatory constraints, and business preferences.

In order of complexity, here are the four most popular master data management implementation styles.

Registry MDM Implementation

The Registry approach is the dominant one among organizations that deal with many disparate data sources, particularly smaller and mid-sized ones. It works by placing data from all of those sources into one central repository where the data can be cleaned, consolidated, and aligned. Matching algorithms are used to identify and remove duplicates.

An advantage of this approach is that the original data isn’t altered—changes are made directly within source systems as opposed to a separate MDM repository. Anyone verifying the truth of data, therefore, can use global identifiers to track it back to the original unaltered source.

For example, an organization might draw data from enterprise resource planning (ERP), customer relationship management (CRM), human resources (HR), accounting, and other systems. In the Registry approach, instead of each system drawing its own sometimes differing conclusions, the MDM plan provides an aligned opinion, prediction, or trend based on all of them. This is a good way to avoid compliance or regulatory repercussions that demand data be preserved in its original form.

While this low-cost approach can be implemented rapidly without much impact on other key applications, there are some drawbacks, primarily reliability and time. Creating a registry hub that can receive, cleanse, and consolidate data from many different sources and types can be time-consuming. Modern platforms can automate these processes as well as the shunting of data and updates from one system to another, making them more feasible.

Consolidation MDM Implementation

The Consolidation style creates what is known as a “golden record” of all organizational data in a single place. Like the Registry style, it brings together data from multiple sources into a hub to develop a single version of truth, but in this approach, a human is involved in verifying accuracy of the golden record and analyzing it for errors. This leads to increased reliability over the Registry approach. In addition, it means the ability to bring experience to bear when evaluating the data, drawing conclusions, and making more informed decisions.

The golden record becomes the primary source of truth in the organization. As it is updated, any changes are pushed out to the original sources—ERP and CRM systems, for example. This is particularly beneficial for organizations that rely heavily on analytics functions, as cleansing, matching, de-duplication and integration functions can then be done in one place.

MDM Consolidation implementations are more expensive than Registry ones, but less expensive than the other types detailed here. The ability to synchronize data with original data sources almost in real time means that users of ERP, CRM, and other mission critical applications are not disadvantaged by long delays in receiving updates from the MDM hub. For these reasons, mid-sized organizations and those with a heavy analytics workload tend to favor this approach, using it to minimize the hassle involved in having multiple silos of information within the enterprise each presenting its own version of the truth.

Coexistence MDM Implementation

The Coexistence style of MDM implementation enables the MDM hub and the original data sources to all coexist fully in real time. Because there is no delay in updating records from one system to another, the golden record remains accurate at all times—as do the related applications that feed the data—leading to efficiency, timeliness, and complete accuracy.

This style is relatively simple for expanding businesses to upgrade to from the Consolidation style, as it takes only minor modifications to link centrally controlled data with their original sources. The benefits of doing so include ease and rapidity of reporting as well as enhanced data management overall.

As long as the central MDM hub and related data sources remain consistent, the golden record is highly unlikely to possess any disparities. Retaining all master data attributes in one place means overall data quality is enhanced, access is faster, and reporting becomes more facile.

Centralized MDM Implementation

The Centralized style of MDM implementation is sometimes also known as the Transaction style. It’s a step up from the others detailed here, as it makes it possible to link, cleanse, match, and enrich the data management algorithms for storing and maintaining all master data attributes. It’s all done centrally and then transmitted to the various sources that originally supplied the data.

In this approach, the centralized master system acts as the central repository. Surrounding systems and applications subscribe to it to receive updates so they remain consistent. This turns the MDM into a full-fledged system of record, which can then act as the primary source for the entire supply chain and customer base. Data creation among suppliers and customers can be done even in highly distributed environments, as it is now established as the system of origin for all information as opposed to being fed system first from other organizational applications.

The master data is always complete and accuracy is assured at all times, and it supports the implementation of advanced security and accessibility policies based upon data attributes—even in organizations with multiple locations, geographies, and domains. This architecture also increases data governance capabilities. As such, it is primarily used in large organizations with stringent data governance policies and deep enough pockets to afford the necessary investment of time and money.

Implementations can be lengthy and are often complicated, requiring a large implementation team, and help from external providers and consultants to execute. In most cases, organizations already have a consolidation or coexistence approach in place before they take the leap into the big leagues with a centralized MDM implementation.

Bottom Line: Choosing an MDM Implementation Style

When it comes to MDM, there’s no one-size-fits-all style that’s right for every business. But the need for the right MDM implementation style and the right data management approach has become increasingly evident following the rise of artificial intelligence and the associated massive capacity for growth.

“A strong foundation for AI necessitates well-organized data stores and workflows,” said Rich Gadomski, head of tape evangelism for Fujifilm Recording Media USA.

As a result, MDM tools are becoming more advanced to serve evolving needs, and now incorporate automated data quality, governance, compliance, and no-code/low-code configurations to meet changing customer needs.

Generally speaking, as the complexity of the data environment increases and the size of the organization expands, the implementation style should become more sophisticated—that means organizations should be moving from Registry to Consolidation to Coexistence and, finally, to Centralized. In some cases, the types of data sources used by the enterprise might mean using two styles simultaneously to lessen complexity, at least at first.

Each business must determine its own specific requirements based on such factors as data quality, access needs, security and privacy necessities, the regulatory environment, existing technology platforms, data types involved, and data governance mandates. By setting the right vision, strategy and policies for data management upfront, the path ahead often becomes clear.

Read next: 7 Data Management Trends: The Future of Data Management

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles