Integrating data in MDM with other systems can present a challenge. In a heterogeneous IT landscape, different systems offer various levels of generalizations of data. For party data, these generalizations stem from multiple reasons including limitations of applications, legacy design decisions, prohibitive costs, etc. Recommended practices within MDM may not be supported by external applications that are upstream to or downstream from MDM.
Conceptually, while a legal entity-based model approach may seem ideal for a company, it may not be possible for external applications to support the data model. This is especially relevant in coexistence and centralized style implementations where synchronization between MDM and external applications are most prominent.
For example, consider a downstream system that does not have the concept of line of business dimensions and therefore has duplicate legal entities across different lines of business. The related MDM system, however, does not have duplicates since a legal entity contains all relevant lines of business dimensions as data containers.
In MDM, the generalization of data should be to the point of how the business wants to interact with the master data. However, when weighed against the capabilities or limitations of surrounding systems this is a balancing act.