Page 85 -

46 Enterprise Data Governance

master data, by definition, is validated before use by
transactional systems. There are more consumers of data
than producers, at least in 80% of cases.

In the end, this bank, without modifying its existing IT
platform, will be able to reduce the opacity of its reference
and master data management by putting a data repository
and adequate governance functions into the hands of its
business users. This type of project lasts for more than just a
few weeks.

First of all, the bank must start a semantic modeling of its
data. This modeling provides a business view of data, i.e.
independent of the physical data structures of existing
databases. This model is used by the MDM system and
allows an automatic generation of administrative functions:
consultation screens and data updates, version management,
audit trail, etc. The access rights on governance functions
are clarified in order to decide what the data repository
management responsibilities should be. For instance, if the
bank so wishes, it could be possible to delegate the
management of certain product configurations to regional
agencies. The aim of this management is not to directly
update the production environment of the banking platform.
Consequently, a validation process (undertaken by the
bank’s headquarters) of configurations proposed by the
regions is put in place. The MDM system then supports the
approval process of data updates (workflow).

2.2.2. Second stage: the business rules repository is
added to the data repository

The second stage of the transformation is concerned with
business rules. We will not describe here the use of business
rules for reference and master data validation in an MDM.
This is useful when rules become complex and can no longer
only be registered in the data models.