Data governance is the discipline of ensuring that data is accurate, consistent, secure, and available to the people who need it. Most organizations need it. Very few have the dedicated data team that enterprise data governance frameworks assume. The gap between the need and the assumed resourcing leads many organizations to treat data governance as something they will address when they are bigger. The problems data governance is designed to prevent — duplicated records, inconsistent definitions, orphaned data, inaccessible data — accumulate in the absence of governance and become more expensive to fix the longer they are allowed to persist.
Governance Without a Data Team
Data governance without a dedicated data team is governance distributed across the business — which is actually how most data governance problems should be addressed anyway. Data problems are fundamentally business problems, not technical ones. The finance team owns financial data. The sales team owns customer data. The operations team owns inventory and production data. Effective data governance assigns accountability for data quality to the business functions that produce and consume the data, provides them with the tools and process to exercise that accountability, and establishes a lightweight coordination mechanism to handle cross-functional data issues.
The Minimum Viable Governance Model
A minimum viable data governance model for an organization without a dedicated data team has four components. Data stewards — designated individuals in each business function who are responsible for the quality of the data their function produces. A data dictionary — a shared definition of key data entities and the rules that govern their quality. A data issue process — a mechanism for reporting, tracking, and resolving data quality problems. And a periodic review — a monthly or quarterly meeting where data stewards review data quality metrics, surface cross-functional issues, and make decisions about data standards. This is not a full enterprise data governance program. It is the foundation that prevents data problems from becoming data crises, at a cost that organizations of any size can sustain.