Share with your CDO
York University built its data governance program around capability delivery rather than policy enforcement, a deliberate choice that took six years and produced 830 common data definitions, 310 classified data elements, and double-digit improvements in user-reported data findability and access. Director of data strategy Patrick Cernea started in 2020 with a two-year foundation phase, added a CDO and a full team in years three and four, and only pushed a formal governance policy through York’s board in December 2025, after users were already bought in.
What this means for your business
The story York tells is really a sequencing argument, and sequencing is where most enterprise data governance programs die. The recurring failure mode looks like this: a governance team arrives with a policy, users experience it as obstruction, and adoption collapses before any metadata repository or data catalog earns enough trust to matter. If your organization is earlier than year three in a governance buildout, or if you have a policy sitting unsigned while user sentiment sours, this case is directly about you.
The specific insight worth stress-testing is Cernea’s claim that data literacy investment buys policy acceptance. York ran quarterly community-of-engagement sessions on practical topics, dashboards and student retention analysis, not governance theory, and only formalized policy after that groundwork was laid. That ordering matters because it inverts the default enterprise instinct, which is to lock down data access and then educate. When AI and analytics adoption is accelerating, as York’s $2 million digital transformation context makes clear, a locked-down data estate doesn’t just slow governance, it slows every model that depends on clean, findable, trusted data. The literacy-first sequence is also the AI-readiness sequence.
The CDO who should be most uncomfortable reading this is one whose team has delivered tools, a catalog, a glossary, access controls, without a corresponding investment in the human layer. York’s 6% trust improvement and full-capacity CoE sessions aren’t vanity metrics; they’re the leading indicators that a governance program has moved from being a cost center defending data hygiene to being infrastructure that AI initiatives can actually rely on. I’d revise this assessment if York’s AI adoption outcomes, not just governance survey scores, showed no corresponding improvement once the foundation was in place.
Based on reporting from York’s data governance program eyes capabilities, not control, originally published 2026-07-30 15:41:00.
