Master data represents the business objects which are agreed on and shared across the enterprise. It can cover relatively static reference data, transactional, unstructured, analytical, hierarchical and meta data.
Let me start by telling you what it is not. Master Data Management (MDM) is not a pure technological solution. It’s not a quick, pre-developed GUI to maintain data. To ensure sanitized master data, you need to include fundamental changes which are required in business processes. Often, the most complex master data management challenges are more policy related than technical (and policy is politically sensitive).
What is important to note, is that MDM consists of both creating and maintaining your master data. Creating a sanitized, consistent master data set is a waste of investment in function of time, money and effort, if the solution does not include tools and processes to this master data through updates. To do so, a Data Governance structure is usually set up (really mastering the Data) as part of a Master Data Management program. This process involves the roles of Data Owners, Data Architects and a Data Steward.
In order to implement a successful MDM program, it is important that the focus on people, processes and technology is in balance. Also archiving and keeping track of historical data needs to be part of the Master Data Architecture.
Each organization holds master data, basic information consisting of customers, products, employees, processes, etc. This master data is essential for an organization in order to guarantee an optimum functioning of the organization. Most of the time, this master data is stored in different (duplicated) ways and systems. Every system has its own version of the truth, in totally incompatible ways. This makes it impossible to combine all data, and have a 360° view on the data. There are also other types of data that don’t fit in the existing systems, like configurations, pictures, logos, PDFs, etc. Usually, these are stored somewhere locally in each division.
Depending on the type and structure of the organization, the purpose, sources and targets of specific data, the same data can be master data for one company, and non-master data for another company.
Organizations are engaged in business intelligence, (re-)designing their systems, optimizing business processes, creating a single view of the customer and complying with external regulations. An adequately equipped MDM for this, is an crucial prerequisite. For that reason, MDM has gained a lot of importance within many large and medium-sized organizations.
Managing master data involves a whole process, which can be facilitated by specific technology. If we take this to the Microsoft SQL Server Stack, we are talking about MDS of course, but also SSIS, SSRS and DQS.
Here is an overview of what this process can look like:
Important in the whole process is also that: