What are the different architectural approaches to master data management (MDM) and customer data integration and the benefits of each?
What factors should I consider in choosing an implementation approach?
Creating value from a master data management (MDM) or customer data integration implementation is not a one-time event, but rather a journey. Customers may choose to start with one style of architecture, but over time migrate to a different approach. The analysts, the buying audience, and even the software providers have all accepted the three styles outlined by leading analyst group Gartner and Forrester Research – registry, hybrid or co-existence, and transactional.
The primary differences are found in integration styles and service enablement. The biggest critical difference between registry and hybrid is data ownership. That is, what are the “point of entry” and the “point of maintenance” for core and common customer data?
Registry
- Non-invasive approach – data remains in the contributing systems
- Faster and easier to implement
- Links to where master data object is physically held in various source systems
- Both read / write
Hybrid or Coexistence
- Master data is authored and stored in numerous locations
- Includes a physically instantiated golden record in a central hub
- Synchronizes master data across the application portfolio; degree of synchronization varies based on business requirements
Transactional
- Customer data is owned and maintained in a centralized hub.
- Master data is authored in the central hub
- Requires ongoing synchronization between the CDI hub and the operational systems
- Requires strong transactional performance characteristics
- Requires strong organizational commitment
A transactional hub model offers several benefits and serves as a foundation for a move to a service-oriented customer management architecture; however, because of its more invasive approach with existing mission critical applications it is not always the near-term or even long-term end state for MDM or CDI. The ultimate right choice for a customer depends on a few key factors:
What is your time frame for achieving a return on investment? Complexity associated with transactional hub deployments leads many enterprises to consider registry style implementations because of quicker deployment times and ROI. An enterprise can benefit immediately and transition to a transactional hub over time.
Who or what systems own the data?
Nearly every system deployed in or offered to current corporate IT infrastructure operates under the assumption that the data about the customer used in that system is managed directly by that system. This ownership assumption is embedded in the business processes, the internal operations and the technical core of these systems. Wresting ownership of the customer data away from any system poses serious challenges. The reality is that reprogramming, re-architecting or customizing an existing system so that it provides customer data to (and retrieves customer data from) a separate transactional hub is complex.
Are there political, privacy or security concerns prohibiting you from centralizing all the data?
Distribution of customer data within an organization’s own IT systems is a critical roadblock to obtaining a single view of the customer. However, this roadblock is compounded when the customer channel is extended beyond the enterprise into a larger network of affiliated specialists, vendors and service providers. Such an extended channel is becoming the norm rather than the exception in an increasing number of industries, including healthcare, financial services, banking, hospitality, transportation and retail.
Will you be adding additional source systems over time?
Companies exist in a state of constant change to corporate IT infrastructure resulting from system upgrades, acquisition of new companies and divestment of business units. As the timeline for reaching a project’s goals lengthens, the project will have to absorb more of these types of changes. New applications selected by existing business units or added to the mix from acquisitions will simply add to the workload for rewiring to the transactional hub. A MDM or CDI approach that delivers value in rapid, concrete steps that build on a resilient foundation can help an organization absorb changes more readily.
The deployment of a MDM or CDI solution may start in one of the three styles of CDI-MDM and evolve to a different style or a hybrid of two or more styles. Initiate Systems has customer implementations covering all three architectural styles and our experience over the last 10 years has led to the conclusion that successful deployment of CDI-MDM infrastructure follows an evolution. We advise customers that start with a registry style approach and choose to evolve their implementation to a transactional hub to consider the following.
- Data analysis and migration – Assess the obstacles in creating and persisting a single, golden record. For example, if you find 3 records all thought to be the same customer, and 2 different addresses and 3 different home phone numbers are represented across the 3 records, which values should survive and persist on the golden record?
- Source system clean-up – Will the customer data continue to reside in the original source systems in addition to the MDM or CDI hub, or will it simply point to the single golden customer record based on the new, centralized customer identifier?
- Service integration – A MDM or CDI hub, especially in the transactional hub style must be tightly integrated with connected applications. Any service centered on the customer such as search, create, read, update, delete, merge or unmerge will need to be integrated into or made available via exposed services (e.g., Service-Oriented Architecture) to any application requiring customer information to function.
- Data integration – Initiate recommends a publish and subscribe data integration model for ensuring that systems that maintain a local cache of customer core and common data stay in sync with the MDM or CDI hub.
Regardless of approach, Initiate software provides a comprehensive feature set necessary for effective MDM or CDI across all three styles with the following features and capabilities:
- Flexible architecture: Allows companies to organize customer data according to their business requirements and priorities, whether they want to centralize all or some data and store the rest remotely.
- Accurate identification: Resolves duplications and discrepancies among disparate and redundant data sources, and retrieves accurate details to provide a complete, single customer view.
- Standardization: Provides consistent information about each customer regardless of the computing platform or application that uses it within the enterprise.
- Real-time capability: Delivers responses in real time, enabling employees and support personnel to deliver fast, accurate information to customers and determine customer data quickly.
- Scalability: Scales to support hundreds of millions and even billions of records to accommodate future growth and expansion while providing accurate information as the system scales.
Return to the questions you must consider when evaluating a master data management or CDI solution.