Data Governance Roles: Steward, Owner and Custodian

Data Governance Roles: Steward, Owner and Custodian


StockSnap_KPVOIKEPSG_DGRoles_BPData governance (also called data intelligence) consists of processes and tasks that staff perform within their organization. When these activities are executed well, all data users enjoy easier and faster access to the data they need to make decisions.   Earlier posts have discussed processes and tasks in data governance but now it's time to talk about people.  Recently we did a post on staff data quality responsibilities.  But what are the roles that certain individuals play in governance processes?

Data governance is a team effort, and the roles are unique yet interdependent. As with playing team sports, it helps to know who is the goalie, who is a pitcher, who is a linebacker, who is a forward, and who is the anchor leg on a relay? This post describes the key roles for data governance.

The three distinct roles in data governance are data steward, data owner, and data custodian. Briefly summarized, a data steward is concerned with the meaning of data and the correct usage of data. A data owner is concerned with risk and appropriate access to data. In comparing these two roles, often the data steward doesn’t care who uses the data as long as they use it correctly. Often the steward wants a lot of people to use the data! An owner, however, is concerned with who can access data, and tends to be more conservative with granting access. There is a natural conflict between these two roles, but in some organizations the same person plays both roles. A data custodian manages the actual data. This role manages servers, backups, or networks. This role may provision access per the data owner’s rules, and this role has mastery of a data schema and lineage. In comparison with steward and owner, a custodian has little knowledge of the types of decisions that are made using the data. In other words, a custodian knows exactly where data is located but does not know how to correctly use it.

We have used steward, owner, and custodian as the labels for these roles. Other labels are also popular, such as data trustee and data manager. Whatever label you choose, we think there are three distinct spheres of responsibility - defining the meaning of data (steward), controlling access (owner), and physically curating the data (custodian). Establishing roles for each of those spheres, and labeling them suitably for your organization, will cover the responsibility for the majority of the data governance work. In a small organization, the same person may play all three roles. Even in large organizations, sometimes the steward and the owner are the same person. Because of the unique nature of each role, it is helpful to articulate each role even if they are assigned to a single person. Each role makes unique types of decisions and brings a unique perspective and skill set to governance work.

Providing a written description of each role, and personally conveying that role to an individual, will help that individual perform the role successfully. Formally assigning roles makes it easier for colleagues to approach an individual playing a particular role and ask for assistance.

For additional data steward related resources check out our resources blog post on the subject.

If you need help in implementing data governance or data intelligence, remember that IData provides data governance services.  A data governance solution like the Data Cookbook can help in successful implementation of data governance at an organization or a higher education institution as well as improving data quality.  Feel free to Contact Us.

(image credit StockSnap_KPVOIKEPSG_DGRoles_BP #1095)

Brenda Reeb
About the Author

Brenda is a consultant in data management, data governance, and the information needs of users. She has over 20 years' experience providing services and solutions in higher education. Brenda has designed and implemented data management policies, established workflows, and created metadata. She is an experienced advocate for data management at all levels of an organization.

Subscribe to Email Updates

Recent Posts