Data governance that survives contact with implementation
A policy nobody follows is worse than no policy. What makes the difference.
Every organisation that has been through an audit finding has a data governance policy. Most of them are PDFs. The question is never whether the policy exists. It is whether anyone's Tuesday looks different because of it.
Why most policies fail
They are written for the audit, not for the work. They describe an ideal state in the passive voice, assign ownership to committees rather than people, and define no moment at which a real decision gets made differently. A policy nobody follows is worse than no policy, because it creates the appearance of control where there is none.
What makes the difference
Governance that survives implementation has three properties. It names a person, not a forum, for each data domain. It defines controls at the points where data actually moves — ingestion, transformation, reporting — rather than in the abstract. And it is small enough that following it is cheaper than ignoring it.
Built, not bolted on
The frameworks that last are the ones built alongside the systems they govern, with the rules written into how the data flows rather than appended afterwards. That is the difference between a policy and a practice.