Master Data Management (MDM)
In recent years, ODS-style feedback systems defined for a specific purpose — reference data — have emerged. All systems are packed with reference data. This data can include the set of data you use to describe the stage of a sale opportunity (for example, a lead, a qualified lead, an opportunity, a forecasted opportunity, and so on).
Additionally, this reference data can be far more complex. In the financial services example from the preceding section, the integrated customer profile can also be considered reference data.
At a basic level, MDM seeks to ensure that an organization doesn’t use multiple (potentially inconsistent) versions of the same reference (or master) data in different application systems or parts of its operations.
Here’s a common example of poor MDM: A bank at which a customer has taken out a mortgage begins to send credit-card solicitations to that customer, ignoring the fact that the person already has an account relationship with the bank. The customer information used by the mortgage-lending section within the bank lacks integration with the customer information used by the credit card organization of the bank.
Often, this problem exists because the applications were written to optimize internal processes — not to unify customer information. Similar situations exist for products, locations, employees, and vendors — all considered reference data.
Processes commonly seen in MDM solutions include source identification, data collection, data transformation, normalization, rule administration, error detection and correction, data consolidation, data storage, data distribution, and data governance.
These processes are also used in the logical evolution of the ODS feedback loop. A set of tools that have been put together to enable the efficient interchange of reference data between various systems — including run-the-business and monitor-the-business solutions.
What entities you consider for MDM depends somewhat on the nature of your organization. In the common case of commercial enterprises, MDM might apply to such entities as customer (Customer Data Integration), product (Product Information Management), employee, and vendor.
MDM processes identify the sources from which to collect descriptions of these entities. In the course of transformation and normalization, administrators adapt descriptions to conform to standard formats and data domains, making it possible to remove duplicate instances of any entity.
Such processes generally result in an organizational MDM repository, from which all requests for a certain entity instance produce the same description, irrespective of the originating sources and the requesting destinations.
With the emergence of MDM, data warehousing teams can resolve problems such as issues with the quality of data, consistent classification and identification of data, and data reconciliation.