Data Governance

Data is spread across systems and teams, so without clear rules the same figure can mean different things in two reports. Data governance sets who is accountable for each data set, usually named owners and stewards, and who may decide on definitions, access and sharing. A framework normally includes a council that sets priorities, policies and standards, a business glossary, quality targets, classification rules and a process for handling issues. It supports regulatory duties such as privacy law and risk reporting, and it provides the basis for analytics and AI by making data trusted. Governance should be distinguished from data management, which carries out the daily tasks such as storing, cleaning and moving data; governance gives them direction and checks the results. Programmes tend to succeed when tied to a concrete problem, such as reporting errors, rather than launched as an abstract exercise.

A lender finds that its credit, finance and risk teams each report a different customer count. A governance council appoints one owner for the customer data set, agrees a single definition of an active customer, and requires all reports to draw from the approved source. Differences fall to a small, explained gap.