Establishing a framework to maintain Governance of data is essential to the on-going health of an IT system; without it, any data quality activities will degrade over time. However, there is no value in establishing a Governance framework unless the framework developed is in alignment with what the business actually requires. Further, it is essential to define what governance actually is. The following definition is taken from www.dmreview.com
Data governance is the process by which a company manages the quality, consistency, usability, security and availability of its data.
Within the scope of a system the term data is taken to include not only the actual data, but also the system data (such as membership of a security group) and application data (such as the installation base). However, the most visible form of Governance is certainly in the area of data governance. Therefore, the majority of this document focuses on data governance; as it becomes possible to extend governance to security groups etc the principles introduced here may need to be expanded.
In order to undertake data governance, certain roles need to be defined. The following are proposed for use:
- Data Definition Owner
- Data Value Owner
- Data User
Data Stewards are policy people. Stewards have several crucial responsibilities, including defining data governance policies and advising data owners and managers on the implementation of those policies. Stewards are also involved in defining high-level information requirements. They develop and monitor control policies for data, and they serve as overall coordinators for enterprise data delivery efforts. Examples of these duties include:
- Defining preferred data source and hierarchy where multiple sources exist for a single data type
- Defining the priority for implementing managed data practices by Data Type, Spatial System Subsystem etc.
Stewards also define performance measures to help determine how well the data governance effort is working. Finally, they work with Data Owners, Data Managers and Users to continually improve corporate data flow. The role of Data Steward can be played by one person or multiple people working individually - depending on the size of the governance effort - or it can be played by a stewardship committee. Either way, the responsibilities are the same.
Data Owners, working with the Data Steward, carry the primary responsibility for defining corporate information requirements. There will be Data Owners in every business function throughout the company, although there is generally only a single Data Owner per data type. Data Owners provide the requirements for technical specialists to create standards for the storage (i.e. data model), retention (i.e. database procedures) and disposal (i.e. archiving) of corporate information. They also carry the ultimate responsibility of information quality (i.e. data values) and availability (security and permissions).
Data Owners control access to information for change management purposes; if requirements change, a Data Owner will oversee changes to data definitions required by the changes in requirements. Finally, the Data Owners work with Data Managers to actually deliver data to the company.
Data Managers work closely with the Data Stewards and Data Owners to implement data governance policies and carry out the data delivery function. On the technical side, Data Managers capture, store, retain and dispose of enterprise information in accordance with policies defined by Data Owners. Data Managers also design the technical infrastructure to meet Data Owners' information requirements.
As with Data Owners, there will also be multiple Data Managers throughout the company. Data Managers work with the user community to share information about current applications and technology, and they provide a conduit for process improvement ideas.
Data Users are crucial to a successful Data Governance implementation because all the policies, requirements, delivery mechanisms and technical architecture designs are created to meet the needs of the user community. Without the Data Users there would be no need for information to be governed!
Data Users must understand their information requirements sufficiently to assist in the data governance function and comply with information management policies. Because they are on the frontline, the Data Users play a critical role in communicating how information is used and how information processes can be improved. They also work with Data Managers to share information about applications and technologies - what works and what does not. In short, the needs of the user community drive the need for data governance.
The four Data Governance roles interact according to the following diagram: (missing) Although the diagram is a simplification of the overall situation, it does convey how the different roles interact.
This shows that the Data Manager is responsible for the values in a table. The Owner is also responsible for the values (in fact the Data Owner has overall responsibility for the values) but also defines the attributes that will be stored. The Data Steward has overall responsibility for the whole table by definition of the Policies, Procedures, Standards and Guidelines that the Owner and Manager use to perform their tasks. The User simply uses the data provided by the Owner.