You propose data contracts. An application team says "we are not a data team, our database is our own business". How do you respond?

What the interviewer is testing

Whether you can make the case in terms of the other team's interests rather than governance principle.

Acknowledge what is true

They are right that their internal schema is their own business — and that is precisely the argument for a contract rather than against it.

Today, downstream systems read their internal tables directly. That means their internal schema is already a public interface they never agreed to, and every refactor risks breaking things they cannot see. They have all the constraint and none of the control.

A contract inverts this: they publish a deliberate interface, decoupled from their internal design, and behind it they can refactor freely. This is the argument that works, because it offers them something they want.

The concrete offer

They publish an explicit contract — a defined set of fields with types, semantics and a change policy — and downstream consumers use only that.

In exchange, nobody reads their tables directly. That is a real benefit they do not currently have, and the data team should be willing to do the work to make it true.

The platform does the plumbing. Change data capture, the outbox, the schema registry, the validation in their pipeline — built by the data platform team, not added to the application team's backlog. Their obligation is to keep the contract's promises, which is bounded.

The obligation being asked for

Be honest: a breaking change now fails their build. That is the cost. It is smaller than the current cost of discovering breakage through someone else's incident, and it is bounded because the contract covers a handful of fields rather than the whole schema.

What a strong answer adds

Starting with one team and one high-value dataset, demonstrating the refactoring freedom in practice, and letting that team advocate. Governance programmes rolled out estate-wide by mandate produce compliance artifacts; one working example produces demand.

Common weak answers

Escalating to a mandate. Arguing from data governance principles, which are not this team's concern.