Data Governance cannot be relegated to a small team of people. Because Data Governance affects the enterprise, it requires an enterprise-wide organization to be effective. This does not mean the company needs a massive team of employees dedicated to governance, but it does mean that there will be governance roles filled by people across the company.
Horizontally, there are three bands of governance roles throughout the organization: Executive Board, Owners, and Stewards.
The Governance Executive Board is a consortium of leaders, likely VPs, that represent the entire business. This group is responsible for setting strategy, prioritizing initiatives, approving the data roadmap and reviewing key decisions. In addition, this board will approve the Data Owners that are part of the next group. The Executive Board should meet quarterly, though it may need to meet more frequently when governance initiatives are ramping up.
A Data Owner is a leader, likely a director, who is Accountable for the data. This person may not know the intricate details about the data, but they are the person who gets called by executives when there is an issue. This person understands and leads the business processes that impact their data.
Responsibilities include:
- Approving/Designating the audience for the data (specifying the Security Classification for data entities).
- Identifying Data Stewards and other Governance roles as needed.
- Collaborating with the Data Governance team to develop a roadmap to better manage their data.
Data Owners should meet on a semi-regular basis, perhaps monthly or quarterly, to discuss shared data initiatives and decide on issues escalated by Stewards.
A Data Steward is an individual who is Responsible for the data. Usually an individual contributor or supervisor/manager, the Owner relies on this person as an expert. They know the minutia behind the data and the processes used to capture and manage it. Moreover, they often know where the bodies are buried and what issues exist with the data.
Responsibilities include:
- Working with Data Governance to identify source and transformation logic for key data entities and attributes.
- Completing User Acceptance Testing (UAT) on their data in the Enterprise Business Model.
- Implement the roadmap developed by the Owner in collaboration with Data Governance.
Data Stewards should meet regularly to discuss methods of optimizing data processes, defining terms that span multiple groups and identify opportunities for improvement.
Data Families
In an immature organization, it can be difficult to identify the appropriate Data Owners for various data objects. It may be easier to break the organization's data up vertically into various subject areas or Data Families and manage it that way.
A Data Family should not align to the organizational hierarchy, but the underlying business processes.
Every organization will be different, but generic examples of Data Families include:
- Accounting
- Finance
- Supply Chain
- Operations
- Sales
- Marketing
- Human Resources
- Information Technology
- Legal
Data Families help because it is often easier to designate a Data Family Owner to be accountable for the whole Data Family. Then when a data object needs an Owner, the discussion becomes, "What Data Family does this object belong to?" and not "Who does this object belong to?".
As a caveat, it's very possible that a data object could be owned by multiple Data Families. That is ok. An object can have multiple owners. All of the owners need to agree on key decisions, but it works.
The end result is that we end up with an Executive Governance Board that presides over vertically organized Data Families, each with a Data Family Owner and multiple Data Stewards.
An established Data Governance organization, like the one discussed here, is key to ensure that the Governed Warehouse aligns to the business and drives value for the enterprise.