I Asked 50 Analysts What 'Data Governance' Means. I Got 47 Different Answers.

Over the past year I've asked every data person I've worked with — analysts, engineers, managers, consultants — to define data governance in one sentence.

I got 47 different answers from 50 people.

Three people gave me the same answer. The rest? A grab bag of "data quality," "metadata management," "data stewardship," "making sure data is correct," and my personal favourite: "that thing we keep saying we'll do after the migration."

Why this matters

When a team doesn't share a definition of governance, they can't agree on whether they're doing it well. And if they can't agree on whether they're doing it well, they can't prioritise fixing it.

I've sat in rooms where one person thinks governance means having a data catalogue and another thinks it means data security policies and another thinks it means a stewardship committee — and all three are signing off on the same "data governance initiative."

They're not working on the same thing. They think they are.

My working definition

Data governance is the set of rules, roles, and processes that determine who is allowed to do what with data — and what happens when something goes wrong.

Rules: what does "valid" data look like? What are the quality thresholds? What are the definitions?
Roles: who owns each data domain? Who can approve definition changes? Who resolves disputes?
Processes: what happens when quality drops below threshold? How does a new metric get defined? How do you handle a breach?

Everything else — the catalogue, the lineage tools, the stewardship meetings — is infrastructure for those three things. Not governance itself.

The test

Ask your team: "If the definition of 'revenue' changed tomorrow, who would know, who would approve it, and who would update every downstream report?"

If you can't answer that in 30 seconds, you don't have governance. You have documentation.