The chief data officer leaves before the definition of "customer" does.

By the time a new chief data officer learns where the bodies are buried, the role has a way of relocating them to someone else's org chart. The definitions, meanwhile, do not move at all.

The job churns. MIT Sloan put the average chief data officer tenure near 30 months — half a CEO's run — and noted a quarter of organizations have no single point of accountability for data. Wavestone's 2024 executive survey, by Randy Bean and Thomas Davenport, filed the same report: turnover high, tenures short. The role arrives with a mandate and leaves with severance.

The diagnosis is rarely the person. Commentators argue that data and AI leaders are set up to fail — handed accountability for outcomes without the authority, budget, or clarity to deliver. The title says chief; the reality is convener of meetings about whose number is correct. When the role owns everything and controls nothing, short tenure is the design as drawn.

Here is what the revolving door reveals. Accountability is not a chair; it is a written decision right attached to a defined term. If "customer" and "revenue" each mean six things across five teams, no one can inherit responsibility, because nothing is on paper. Each new CDO re-litigates the vocabulary, ships a glossary nobody adopts, and departs. The terms outlast every nameplate.

Watch the handoff, not the hire. When a data leader exits, ask what survives them: which definitions, owners, and decision rights are written down and would still hold next quarter. If the answer is a deck and a committee, the next hire inherits the same vacuum. Tenure stabilizes when the organization stops recruiting a person to be the definition and writes it down instead.