Data Governance
What is the difference between data governance and data quality?
Last reviewed
Direct answer
Data quality asks whether a value is complete, valid and accurate within its system. Data governance asks what the value means, which source has authority for it, and who decides when sources disagree. Two values can both pass every quality check and still conflict, because they are accurate under different definitions. Quality is necessary, and it is not enough.
Why this happens
Quality tools catch missing dates and invalid codes. They cannot decide whether billable census or clinical census should drive staffing.
Good governance improves quality, because owners have clear definitions to check values against.
Example: one metric, three definitions
Operations
Occupancy 92%
Residents divided by licensed beds.
Finance
Occupancy 90%
Billable resident days divided by available unit days.
Ownership report
Occupancy 91%
Occupied units divided by units in service.
- Which number governs?
- Operator Authority
- Governing Record
- SeniorCRE Intelligence
- Execution
Illustrative figures. Governance first. Intelligence second. Execution last.
Integration can move all three numbers into one place. Only governance decides which definition is approved for which purpose.
The governance question
Governance answers four questions for every important measure: what it means (definition), which system is authoritative for it (authority), how disagreements are resolved (reconciliation), and how anyone can trace the result back (lineage).
Evidence → Disagreement → Authority → Governing Record
What good operating infrastructure should do
- A register of approved definitions, owned by named people.
- Explicit source-authority rules per measure and per decision.
- A reconciliation process that records outcomes, not just corrections.
- Lineage from every reported number back to its sources.
- Change control when a definition or authority rule changes.
SeniorCRE perspective
SeniorCRE treats governance as the operator’s function, not the software’s. Its architecture encodes Definition, Authority, Reconciliation and Lineage as first-class records, so integration supplies evidence and the operator decides what governs. Unified data is not the same as governed truth. Status: the Governed Operating Record is designed and not yet implemented in any community. What has been built, validated and deployed is published on the evidence page, and nothing here should be read as a production result.