Page 120 -

82 Enterprise Data Governance

master data, from its configuration. But it also exists via
business processes, i.e. the transactions which support the
activity of the company, like, for example, order entry
processes and replenishment processes within the company.
In particular, stock information is not included in the master
data definition (see Chapter 1); rather, it is transactional
data.

Therefore, during the modeling of reference and master
data, certain patterns of this modeling will be valid for the
reference and master data as well as for transactional data.
1
These are business objects and their business states .
Consequently, the foundation of the reference and master
data modeling is, in reality, nothing more than the
enterprise data architecture, including both reference/master
data and transactional data. This data expresses all the
information that the company handles, beyond the sole scope
of master data.

If this aspect is not properly understood, then the
company runs the risk of modeling reference and master
data that cannot be used as a launching pad for the re-
appropriation of transactional data. This re-appropriation
comes, after the implementation of the MDM system, as the
progressive transformation of the IT system gets under way.

At this point in time, these modifications require the
setting up of transactional data models. If other business
objects, with other business states, are used, this provokes a
distortion between the reference/master and transactional
data.

The financial valuation of the MDM approach also goes
through the return on investment of the modeling effort. If

1. These states are necessary in order to make the validation rules of the
master data more reliable. For more detail on this, see Chapters 8 and 9
dedicated to semantic modeling.