When data ownership works, data owners should disappear

Dinesh Mandolkar

06 Mar 2026 — 3 min read

Most conversations about data governance start in the wrong place.

They start with roles.

Who is the Data Owner?
Where do they sit?
What is their RACI?

But data governance doesn’t begin with people on an org chart. It begins with something more basic: data ownership.

Ownership is the simple, uncomfortable idea that someone must be able to answer for a decision made using data. Not fix it. Not document it. Answer for it. When data is wrong, misused, or misunderstood, ownership is the difference between confusion and clarity.

Only after an organisation accepts that responsibility needs a home does the question of Data Owners arise.

Why Data Owner roles appear

The Data Owner role exists as a transitional mechanism.

It appears when:

Decision rights are unclear
Accountability is diffused
Escalation is ad-hoc
Risk has no natural landing place

The role is created not because ownership is mature, but because it isn’t. It is a way of forcing the organisation to ask: who decides when rules meet reality?

At this stage, having named Data Owners is useful. Sometimes essential. But the role itself is not yet the solution it is a signal that the organisation is still learning how to carry ownership.

Where ownership sits changes how it behaves

Once Data Owners exist, organisations face an unavoidable design choice: where does ownership live?

Placing ownership high in the organisation gives it authority. Placing it lower gives it proximity. Neither is inherently right or wrong but each behaves very differently.

When ownership sits high, it carries weight. Decisions can cut across silos. Standards can be challenged. Escalations are heard. But distance from day-to-day reality becomes a risk. Without support, senior ownership turns abstract. Issues arrive late, filtered, and stripped of context.

When ownership sits low, it becomes grounded. Problems are visible early. Context is rich. But mandate becomes the limiting factor. Without explicit authority, ownership becomes procedural. Decisions stall. Risk is managed locally instead of addressed structurally.

Both models can work. Both fail if left unsupported.

How the two models are made viable

High-level ownership only works when it is assisted.

Senior Data Owners must retain accountability while delegating execution. This requires stewards who act with real mandate people trusted to investigate, enforce standards, and prepare decisions that can be made quickly and coherently. Accountability remains singular. Operational load does not.

Low-level ownership only works when mandate is delegated.

If ownership is pushed closer to operations, authority must follow. The ability to reject unsafe usage, enforce standards across teams, and escalate without dilution. Without this, decentralisation becomes exposure, not empowerment.

This is why decentralised ownership is viable only in very mature organisations. It assumes shared risk language, trusted judgment, and escalation paths that function without politics. Most organisations reach for decentralisation before they’ve earned it.

Maturity changes the shape of ownership

As organisations mature, something subtle happens.

Ownership stops being performed through roles and starts being expressed through behaviour. Decision rights become clear. Accountability becomes habitual. Escalation becomes normal instead of dramatic.

At this point, the Data Owner role begins to dissolve. Not because ownership disappears but because it has been absorbed. Product leaders own data decisions for their products. Domain heads own data risk for their domains. Compliance owns interpretation. Engineering owns execution. Ownership is no longer assigned. It is assumed.

The role was scaffolding. The structure now stands on its own.