Metadata Management and Data Governance - Data Governance: Creating Value from Information Assets - page 54

that it cannot implement a single version of the truth because some of
its divisions utilize the same terms, though they have different meanings.
So, it adopts a Fit for Purpose model. This presents the company with a
serious issue. It created a data Governance Board represented mainly by
the vice presidents of the various business areas, with IT as a technical
adviser. IT is asked for the number of versions of patient stay defined in
the model and where they are defined. IT quickly does a search in the
metadata repository. It determines that there are five different versions of
patient stay. The data Governance Board asks how much time and effort
would be required to reduce the number from five versions to two versions
of patient stay . IT quickly does an impact analysis through its metadata
tool and determines that it would require modifying 50 easy ETL pro-
grams, 100 medium ETL programs, and 10 hard ETL programs at a cost
of three man months. The board approves the changes.
MASTER DATA MANAGEMENT
Master data management (MDM) provides for special data quality pro-
cesses on a limited number of objects. MDM utilizes metadata to assist in
the cleansing of MDM data. However, it is a separate concept from meta-
data. All objects cannot be given MDM treatment. Usually, the MDM
scope is limited to between 5 and 10 objects. These objects are very critical
and need special treatment to ensure a higher data quality standard. For
example, we need to know that a patient is the correct person in our data-
base. We may have more than one Mary Smith. It is important to know the
correct Mary Smith in order to deliver the correct treatment. One Mary
Smith may be allergic to penicillin and another Mary Smith may not. Do
we give Mary Smith a penicillin treatment? MDM repository contains one
true version of each master entity created from multiple source systems
and this “golden copy” is used by downstream systems. MDM tools are
designed to identify duplicates, handle the variations of key entities across
source systems, and standardize the data. Without MDM, the impor-
tant data that links the subject areas of the data warehouse could not be
relied upon. For example, are Michael Schrader, Mike Schrader, and MT
Schrader the same person? With MDM, we can identify Michael Schrader,
Mike Schrader, and MT Schrader as one person. MDM requires special