Data Core · Data Governance · Module

Ownership

Who is accountable for each data set being right, current and appropriately used — one person, in the business rather than in a technical function.

The idea

How it works

Data quality is nobody’s problem where nobody owns it. Errors persist through several reports and are corrected separately in each, indefinitely.

Ownership belongs with whoever understands what the data means rather than with whoever stores it. The technical function is a custodian and cannot judge whether a value is right.

Working with it

In practice

  1. 01

    Name one owner per data set

    In the business, not in the technical function. A committee is not an owner.

  2. 02

    Separate owner from custodian

    The owner decides what is correct; the custodian keeps it running. Confusing them means nobody judges correctness.

  3. 03

    Give the owner authority to change it

    Accountability without authority produces a name on a document.

  4. 04

    Include ownership in handovers

    It lapses silently when someone changes role, and nobody notices for a year.

One level in

The components of ownership

A component is something that exists afterwards which did not exist before — a deliverable or a mechanism, not an intention.

  1. The owner

    Who is accountable for the data set being right.

    Learn
  2. The custodian

    Who operates the system it lives in.

    Learn
  3. The handover

    What happens to ownership when someone changes role.

    Learn

Ownership belongs with whoever understands the meaning, not with whoever runs the system it sits in.