Establishing a business case for a Master Data Management System (MDM) has become easier as more enterprises recognize the vast benefits of linking its critical data to a single reliable point of reference, which is typically referred to as a master file. With a centralized MDM, an enterprise can easily and consistently tap into its core data including employee, supplier and customer data, for analytics in support of advantageous and cost-saving business decisions. Formally, MDM is considered an endeavor involving both business and IT to ensure data uniformity, accuracy, accountability and stewardship. Difficulties initiating and maintaining a sustainable MDM are inherently problematic; in that sought after data is typically scattered throughout the organization, unstructured and housed within various systems and applications. For this reason, it is imperative to establish cross-functional governance that will effectively initiate, establish and maintain the MDM initiative components throughout IT and business units alike.
When it comes to MDM, metadata is king, and consistency his queen. However, consistent metadata is one of the largest hurdles to overcome in identifying, gleaning and maintaining core data. Various applications across enterprises, often times numbering into the thousands, can run the gambit in metadata naming conventions without established normalization procedures. Metadata inconsistency is a particular recurring issue with the introduction of cloud applications, which are often imbued with a unique semantic technology altogether. In an attempt to circumvent inconsistent metadata, enterprises often extract core data by leveraging data lakes or central repositories already rich with semantics by technological means. Alternatively, or in conjunction with the aforementioned solution, metadata normalization can also be achieved through establishment and implementation of governance for terms, definitions and standard processes.
The following recommendations provide guidance in laying the proper foundation for MDM, whether to be considered an independent initiative or as part of an existing Data/Information Governance plan:
Set Initial Business Parameters: Prepare the scope of businesses likely/currently using key core master data on a consistent basis. Initial efforts may focus on just one MDM domain (customer data, for example), rather than multiple attributes. This decision may hinge on the implementation style, data locations, how intrusive the effort will be for businesses, etc. Few companies have neither the time nor resources to commit to an enterprise-wide MDM.
Form a Cross-Functional Governance Council: The council should be comprised of key stakeholders with the capacity to make high-level decisions, implement policies and procedures, delegate and/or control accountability and claim ownership over the MDM initiatives. These individuals should include business owners of the relevant core master data.
Set the Standards for Metadata Normalization and Semantic Modeling: Task the council to scour business feedback regarding key terms and definitions that are relevant to core master components and the ultimate business objective; and create business rules and procedures that establish relevant standard metadata terms and definitions. Also, aim to use definition techniques that sufficiently separate master data from data specific to applications. Part of this process should include education concerning the established framework, underscoring the value of consistent metadata.
Plan for Change: Master data definitions will likely change as business parameters, systems, applications and associated processes evolve along with the organization. A key to long-term master data management requires adaptation and flexibility to address these changes.
Start Small: Start with a test case prior to organizational implementation. Depending on an organization’s size, this may involve one business or a project. Use incremental assessments to gauge successes or failures in relation to business objectives. The lessons learned from this initial test case will assist in the full implementation and initial success will provide leverage during implementation across the organization.
Implementing the proper foundation for metadata management is a worthwhile endeavor that can be achieved through proper planning. Most importantly, recognize the necessity of a framework that incorporates consistency that is understood and adopted cross-functionally and has the flexibility to adapt with the enterprise as it inevitably evolves.