For practical purposes, master data management is a set of disciplines and tools enabling business data consumers to access a unified view of shared data about one or more specific master data domains such as customer or product. Yet, while the technical processes for extraction and consolidation drive the activity in most MDM programs, the actual intent of MDM is to satisfy business consumer needs for access and delivery of consistent, shared data.
Satisfying user needs means “working backward” by understanding how master data is employed within critical business functions and processes – and how that data is to be consumed by the users. This means identifying the key data consumers in the organization and soliciting their needs and expectations (both now and in the future). Information architects must work with the business teams to understand how the organizational mission, policies and strategic performance objectives are related to the use of master data. Finally, it’s necessary to understand how improvements in information sharing will maximize corporate value.
A gap analysis is performed to determine what must be done to the current data environment in order to satisfy future data consumption objectives. Performing this gap analysis helps in two ways. First, it isolates key strategic goals for data sharing that must be put into place before any technical MDM approach can add value. Second, and more importantly, it establishes the value of fundamental data management best practices that benefit the organization beyond the needs of a specific master data project. Effective data management practices penetrate all aspects of information use across the enterprise, such as:
• Data governance, which formulates the policies, processes and procedures to ensure that data use complies with explicit business policies; engages business data owners; identifies key data assets to be managed and shared; and delineates specific data requirements and quantifiable measures.
• Metadata collaboration, which defines standards for business glossary terms and definitions, representations for conceptual data elements, and alignment of models in ways that will reduce conflicts when data sets are merged into a unified master view.
• Data quality management, especially when deploying inspection and monitoring compliance with defined data standards and rules; and integrating services implementing data controls directly into the application infrastructure.
• Integration of identity resolution and management within business process model, which best satisfies the ongoing need for maintaining unified views for specific entities such as customer or product.
Use of these best practices does more than lay the foundation for improved information. These practices highlight the relationship among business processes, information and use. They also emphasize the need to adjust staff member behaviors as access to master data provides greater customer and product data visibility. And while the success of your strategic management plan for MDM must have milestones and deliverables aligned with these disciplines, the organization will directly benefit from each practice area independently.
How Data Governance Supports the Data Strategy
Because enabling comprehensive visibility into a composed view of uniquely identifiable entities will continue to be part of the information strategy, there must be a reliable approach for:
- Ensuring that proper validation and verification is performed and approved as new data is created or acquired.
- Confirming that enterprise requirements for the quality of shared information are satisfied.
- Accessing and managing the composed view of shared information within defined security controls.
- Guaranteeing the consistency, coherence and synchrony of data views composed from a variety of sources.
Data governance provides the foundation for mapping operational needs to the framework of a sound data strategy designed around unified master data domains. There are numerous aspects of instituting a data governance program, including these key practices:
- Data governance program management – Developing an ongoing program management plan that identifies roles, defines responsibilities, provides templates for key artifacts (such as data policies, data quality requirements and policy deployment plans) and specifies the types of tools to support the data governance program.
- Data governance operating model – Specifying the organizational structure for operationalizing data governance, the interaction and engagement model to create a data governance council, the development of ongoing agendas, meeting schedules, and the establishment of measures to ensure that progress continues.
- Definition and deployment of data policies – Developing a framework for the process and workflows related to drafting, reviewing, approving and deploying data policies, as well as integrating business rule validation within the application infrastructure.
- Data stewardship – Defining the operational procedures for data controls, inspection, monitoring and issue remediation related to data policy violation.
- Collaborative agreements. Introducing data governance and data stewardship opens the door for agreeing to existing or newly defined data standards, business glossary terms and definitions, data element concepts, and corresponding data types and sizes. A governance program facilitates these agreements, and collaboration tools can supplement the tasks associated with the metadata management life cycle.
- Data lineage. Data lineage needs to be mapped from creation (or acquisition points) to the various modification or usage touch points across the application landscape. An accurate lineage map helps in understanding the business application and process impacts of modifications to definitions or underlying data structures. It also enables the insertion of data validity controls for inspecting and monitoring data quality and usability
A reasonable investment in metadata management can add value to the organization by facilitating communication about shared concepts across business functions, while reducing variance and complexity. Metadata management also adds value by smoothing the differences between data sourced from functional silos – and paves the way for an effective MDM effort.