If you are just starting out and feel overwhelmed by all the various definitions, explanations, and interpretations of data governance, don’t be alarmed. Even well-seasoned data governance veterans can struggle with the definition and explanation of what they do day to day. It’s not that we/they (newbies and veterans alike) don’t understand the big picture that data needs to be managed effectively across the whole enterprise; it’s the complex nuances of making data governance successful that is far more difficult to explain.
When it comes to the implementation and execution of a data governance program, even the most experienced practitioners find themselves frequently answering questions with “it depends.” Because, ultimately, it does. Factors such as industry, available resources, organizational culture, executive support, and budget (just to name a few) all influence critical decisions and can significantly impact the success or failure of a program.
Yet even with all the complexities, there are elements of data governance that remain constant and can be defined. In the members-only section of the DGPO (Data Governance Professionals Organization) website there is an extensive glossary of data governance related terms, which includes a definition of Data Governance. This definition identifies the core elements of data governance and ties them back in with the overarching purpose – “to ensure that data is well-managed as an enterprise resource.”
Let’s look at the definition and break it down a bit further to really clarify what data governance is and is not.
“Data Governance is a discipline that provides clear-cut policies; procedures; standards; roles; responsibilities; and accountabilities to ensure that data is well-managed as an enterprise resource.”
This definition is packed with the core components and concepts necessary for any data governance program regardless of dependencies and complexities.
DISCIPLINE – The word discipline is powerful as it encapsulates some of the most difficult aspects of data governance. It identifies data governance as a practice; the application of a defined methodology. It indicates the necessity for rigor and consistency with the push for continuous development and improvement. It implies that the work of data governance is never finished, but rather a continuous process of refinement.
While discipline also indicates control, it should not be the abuse of power or authority. Too often governance is seen as red-tape and bureaucratic because the concept of control is implemented incorrectly. Disciplined control is not about the domination of decisions. It is about fostering an environment where decisions are made at appropriate levels of authority (the lower the better) within defined parameters and guidelines.
POLICIES – Policies are documented principles and guidelines that establish parameters and expected conduct for specific situations, conditions, or circumstances. Well-defined policies are concise and easily consumable. They demonstrate the intent and/or outline the problem being addressed. Effective policies are not only put into writing, they are also consistently monitored and reviewed.
The key to policies is that they reflect business needs and objectives and guide conduct accordingly. Policies do not however express the details of specific rules or standards. These are separate and distinct and may vary depending upon the application of the policy in context.
PROCEDURES – Procedures are repeatable processes that establish the specifics of “what and how” for activities necessary to achieve targeted outcomes. Procedures outline the steps necessary to reach the desired result. All procedures should be clearly linked to at least one policy. Like policies, they should not only be documented but also consistently monitored and reviewed for efficiency and effectiveness.
ROLES – Roles describe functions but not specific people. Yes, roles have associated “whos” with defined responsibilities and/or accountabilities (more on that in a moment) for specific activities but it is a common mistake to consider roles to be the “who” instead of defining the function – the “what” – that the role is supposed to do. The function of a role is to start or own a process, make decisions or act at specific points within a process, or to complete a process.
With the emergence of new technologies, especially with automation and AI, it is important to note that roles may assigned to machines as well as humans. (Repeat: roles are about the function, not about who executes). Roles may be owned and executed singularly or by a group. Likewise, a machine or specific individual may have multiple roles, be a member of multiple groups, or be a member of a group with multiple roles.
RESPONSIBILITIES – Associated with roles, responsibilities define who will execute specific tasks. More than one role, person, or machine may have responsibility for a step in a process. Responsibility is assigned and has oversight from the one with accountability. Responsibilities do not include authority for decisions or liability for outcome. Responsibility is the tactical execution of assigned activities.
ACCOUNTABILITIES – Also associated with roles, accountabilities are typically (with little exception) assigned to one person with ultimate authority for all decisions necessary to execute the task and all liability associated with completion and outcome.
Clear definition of tasks with associated roles, responsibilities, and accountabilities are best documented in assignment matrixes such as RACI charts.
It takes quite a bit to define what is necessary to achieve the overarching purpose of data governance – to ensure that data is well-managed as an enterprise resource. The members-only glossary of terms not only includes this definition, but many, many others. The glossary helps to clarify the fine lines between many similar data terms and serves as a great authoritative resource when needing to provide clarification to others. It is one of the many benefits of being a DPGO member.
Check out the DGPO website for additional content and resources. We want to help make sure you and your data governance program are successful. Not a member, yet? No problem, you can join here https://dgpo.org/membership/.
