Page 136 -

100 Enterprise Data Governance

These states must therefore be available in the MDM
system, in a structured way. A semantic MDM system
matches this requirement. Instead of the complex
management of a large number of data synchronizations, it
is more efficient to group them and rationalize them in
relation to the management of the business objects’ states:
better performance, reduction in the number of errors, better
visibility of data exchanges between the MDM system and
the rest of the system. A static MDM system, not having at
its disposal the management of the business states, is unable
to benefit from such a rationalization of data
synchronizations.

5.4. The MDM maturity model

Figure 5.1 represents the MDM maturity model,
according to the virtual, static and semantic levels that have
been described. The difference between a static MDM and a
semantic MDM system is considerable. The strategic
expectations are not the same; the effort required in data
modeling and the acquisition of business knowledge are not
comparable.

A virtual MDM system remains an IT solution. It
rationalizes the development required to access data spread
across several databases.

A static MDM system does not bring with it any real data
governance as it is not based on validation rules expressed in
a formal manner. This MDM is sufficient for IT as long as
the data structures managed remain simple, as in codes and
labels (reference data rather than master data), without any
significant links to the business objects which form the heart
of the Information System. This MDM system is dangerous
because it necessitates bespoke software developments to
implement data validation rules outside the data repository.