CONCEPT Cited by 1 source
Dual-owner metric governance¶
Definition¶
Dual-owner metric governance is an ownership model where every governed metric has two mandatory, team-based owners with distinct responsibilities:
| Role | Who | Responsible for |
|---|---|---|
| Business Owner | Data Analyst / Data Scientist team | Metric health — provides written + SQL definitions, maintains the definition on a regular cadence, defines allowed dimensions/granularities, gathers stakeholder approval |
| Operational Owner | Data Engineer / Data Scientist team | Data health — ensures quality of underlying tables, provides on-call support, creates/maintains DQ checks, runs backfills, communicates delays/changes |
Both owners must approve any change to a metric's definition. This requirement serves two purposes: 1. Informed and coordinated action — all relevant parties are aware and can make reactive adjustments 2. Accountability and quality control — each party may be aware of critical nuances the other is not
(Source: sources/2026-06-10-lyft-metric-semantic-layer)
Design decisions¶
- Teams, not individuals — owners must be teams to survive org changes, rotations, and attrition
- Dual sign-off — prevents silent drift from either the business-logic or data-engineering side
- Distinct expertise — the Business Owner knows what the metric should measure; the Operational Owner knows whether the underlying data can support that measurement reliably
Relationship to separation of duties¶
This is a data-platform-specific instance of the broader concepts/separation-of-duties-data-governance principle, applied at the metric-definition level rather than at the data-access level.
Seen in¶
- sources/2026-06-10-lyft-metric-semantic-layer — canonical instance in Lyft's Metric Semantic Layer