Knowledge owned, not held

Data offices fail in a predictable way: they become the place where everything escalates. Sources, flows, incidents, business roadmaps, AI questions — all of it lands on the one team that “knows the data”, and that team is soon funded by recharged problem-solving it can never refuse. The result is a crisis-mode organisation that owns nothing and fixes everything.

The alternative is a second-line office: standards, monitoring and challenge, with first-line ownership of data and incidents where the data is produced. The decision-rights matrix that goes with it has an explicit row the office usually resists writing: “does not own or fix the data”. Knowledge is owned by the business, not held by the office.

Stewardship then becomes concrete. Data Sharing Agreements between producing and consuming functions. An owner / steward / custodian model with named people. A quality ladder with levels, so that “good enough for carbon reporting” and “good enough for a regulator” are different, stated targets. In one group, that ladder took carbon-accounting data across five functions from unreliable to 95%, then 98%, then 99% of volume covered within a year, and the method was handed over at two to three months per additional use case.

Funding follows the same rule: no tooling before ownership, no work without a funded commitment. A budget built on service- and consumption-based recharges, mapped capacity and clear CAPEX ownership is the instrument that makes the second-line model hold once the consultant has left.