SQL Server Master Data Services Considerations
By: Siddharth Mehta
Master Data Management (MDM) aims at managing a single inventory of master data in an enterprise using a standard set of tools and processes. As master data is very critical to an enterprise, it is easy to assume that MDM would be one of the fundamental processes that starts in an enterprise, but that is not the case generally and it is often employed to deal with federated master data within the enterprise. This raises questions like why and how does federated master data get created in an organized enterprise? Why is MDM critical and how does it impact the business? How do we implement MDM and what are the tools / processes to employ? We will discuss these and other similar topics in this chapter to understand the overall picture of MDM.
For the purpose of this discussion, consider a financial institution having separate lines of business (LOB) like deposits, savings, checking, credit cards, loans, mortgage, securities, mutual funds, etc. Different LOBs are the scale of small enterprises within itself and operate independently of each other. The master data generated by these LOBs would be critical to them and they would want complete control over this data. Different LOBs would have different master data important to them.
For example, a mortgage LOB would be interested in maintaining asset, earnings and demographics information of a customer, but a securities LOB may be interested in maintaining investments, earnings and demographic information. From this example its easy to make out that there would be some common as well as discrete master data available with different LOBs within an enterprise. To add further complexity, different LOBs may have evolved in the enterprise at different points of time. For example, a financial institution may have started a mortgage and loans department initially in a geography and after successfully stabilizing the business it may have started a securities and mutual funds business. This can create different versions of the same data within different LOBs updated at different points in time. So if they are compared, they may look different, but in reality they may be different versions of the same information at different points in time.
For example, a customer record is being maintained by the mortgage department, as the customer being a Manager in an IT organization with earnings of 100,000 USD annually and residing in California. The customer moves to New York to work for a different IT organization with a better pay package. The customer is now interested in the investments and securities and this LOB maintains a record of this customer with his present details. It is possible that the enterprise may already have a lot of valuable information on the potential customer, but still not sharing and using it across LOBs as the information is not identifiable due to a lack of master data management. The Mortgage and Securities LOBs may not be sharing their inventory of customer master data on a regular basis, which may lead to a waste of resources and even loss of business opportunities. These kind of scenarios make master data management critical for the success of an enterprise.
Master Data Management Considerations
Below are some high level steps / considerations towards implementing MDM.
- Identifying master data - The first step towards MDM is identifying master data based on different factors, as well as the source of the data.
- Collecting master data - Once the data is identified, it needs to be extracted and maintained in a central repository. The data might often need cleansing and standardization post extraction from the source system.
- Data governance - Once data has been extracted, a governance process is required under which data stewards would standardize the data.
- Implement MDM tools - To implement the master data management system, various vendors offer different toolsets. Master data services is Microsoft's technology stack for implementing MDM. In the upcoming chapters, we will discuss this in detail.
- Maintenance - MDM is a continuous process in which master data is regularly refreshed and maintained. A maintenance process driven by data governance needs to be implemented to maintain master data.
In the next chapter we will look at Master Data Services to develop our understanding before we start a deep dive into the implementation.
- Consider reading this article which shows what MDM means from different perspectives.
- Consider reading this article to understand more about MDM in detail.