The traditional approach to Master Data Management (MDM) involves extracting relevant data from multiple systems and storing it in a centralized MDM database. This database becomes the master for certain data types, and changes to the master data are managed in the MDM system that are propagated to up / downstream systems. Efforts, costs, and time required for this approach tend to be very high due to the initial challenges involved in setting it up, as well as for on-going maintenance. Change management takes place in MDM, outside of the realm of the system of record (SoR), which may not be the most optimal approach.