Think big but evolve in small steps. This is what one of our customers, one of the largest telecom operators in Europe, mentioned as their key strategy for implementing a master data management solution at a recent Gartner MDM Summit in London.
An MDM initiative involves streamlining the collaboration of people, processes, and IT systems. Requirements for an MDM initiative are bound to evolve as business needs and priorities change. MDM implementation can become very complex and endless if a clear strategy and strong foundation are not carefully considered.
Expose and Access Data as a Service
With convergence of social, mobile, cloud, and on-premise enterprise applications, the IT landscape has become highly fragmented and poses challenges to almost every organization. A holistic view of overall data across the IT landscape is critical in achieving operational, tactical, and strategic business goals. Doing so requires communication between heterogeneous systems using a robust integration platform. An integration layer based on service oriented architecture (SOA) allows you to expose and access data as a service from various systems across your enterprise. Some of the immediate benefits that are realized with an enterprise service bus are: standardization of shared data structures and ability to rapidly connect to new and existing data sources to create, update, read, and search data.
Create a Strong Foundation to Manage Data
Once the integration layer is established as the foundation, it will be much easier to implement, consolidate, and manage master data. Every organization is different in terms of IT complexity and maturity. Based on the organization’s unique goals, they can start with any of the following MDM implementation styles, which can be evolved from one to another:
- MDM as registry implementation: This style offers limited benefits but it’s faster to implement, as it requires no additional changes to existing systems and interfaces. An MDM system maintains references to source systems for an entity, and relationships among these source systems and audit logs for all changes to these entities.
- Coexistence implementation: This is the next step toward an ideal MDM implementation. Here, entities and attributes co-exist in a source system and an MDM system. All changes in the source system are propagated to MDM and can be bi-directional also. This architecture uses MDM capabilities such as de-duplication, data cleansing, data enrichment, and searching along with data governance capabilities.
- Master data in one place: This is the ideal MDM implementation, but may not be valid for every organization (especially large global enterprises) because it might be difficult to have a single MDM system for the entire enterprise. In this scenario, the MDM hub becomes the main centralized source of master data, providing better control over data governance standards, security, and maintenance of master data objects.
There is no one right way to implement an MDM solution; however, the whitepaper “The Evolutionary Steps to MDM” discusses an approach that provides flexibility for IT and business users responding to changing business requirements.
