Analytical MDM

In many organizations, data quality is not an entirely new issue, especially in the context of projects integrating data from multiple sources, i.e. Data Warehouses and Business Intelligence. After implementing Data Warehouse and Business Intelligence, the organization already knows a lot about data quality and problems with ambiguous interpretation of data. At this stage, central dictionaries must (or at least should) be built to allow for universal categorization and interpretation of business terms. This is a good foundation for building the first, easier step: building an analytical MDM.

With the above “data warehouse” experience, implementing an analytical MDM is an achievable task for any organization. It is necessary to add to the current solutions tools and processes that link dictionaries with data quality and consistency management procedures (Data Governance). It is very important to assign responsibility for these processes and their results. To be able to talk about MDM development they must be carried out directly by business users or dedicated, created for this purpose, units.

At this stage the business will be able to take responsibility for its own data at the analytical level, i.e., it will be able to identify problems and then create rules and corrective processes. Analytic MDM will not yet provide the full ability to prevent problems at the source – where the data is created. However, it will be partially possible where data quality analysis will enable diagnosis of causes and easy remediation.

The next step is operational MDM.