Data governance
Maturity assessment, policies, roles, lineage and data lifecycle using the DMBOK framework.
Three reports, three different numbers.
The committee asks for one indicator and gets three versions, each of them defensible. Nobody is lying: every area calculates with the data it has at hand, and none of that data has an owner.
This is the exact point where everything built on top breaks. A model fed with that data does not fail visibly: it confidently answers something that is not true.
The order matters.
Using DMBOK, and starting where almost nobody wants to start: deciding who answers for each piece of data.
- 01Maturity assessment — Where the organisation actually stands, backed by evidence rather than a perception survey.
- 02Owners before tools — Data with no owner is not fixed by buying software. Domains and named people are assigned, and what each one decides is agreed.
- 03Policy and lifecycle — From the moment data is created until it is deleted, deletion included. With the retention the law requires and not a day more.
- 04Lineage and quality — Where each figure comes from, how many hands it passed through, and what may have broken it. Without this, no indicator survives a regulator.
What changes.
The indicator stops being argued about. Not because the number is prettier, but because there is one way to calculate it and someone answers for it.