Maturity indicators: What levels 1–5 look like

This reference translates numeric scores into observable indicators you can validate during the audit. Use it to make scoring consistent across teams.

Score 1 — Fragmented / Ad-hoc

  • Data quality: frequent, unresolved errors; no reconciliation processes.
  • Ownership: no clear data owners; knowledge lives in individuals.
  • Tools: siloed spreadsheets or manual exports; no centralized storage.
  • Decisions: reporting is descriptive and rarely linked to action.

Score 2 — Emerging controls

  • Data quality: some checks exist but not automated; recurring issues persist.
  • Ownership: informal stewards; responsibilities not enforced.
  • Tools: basic ETL or dashboards exist but lack integration or clear SLAs.
  • Decisions: occasional experiments; limited measurement of outcomes.

Score 3 — Repeatable practices

  • Data quality: routine QA checks; some reconciliation and backfills possible.
  • Ownership: named data owners and stewards for important datasets.
  • Tools: standardized stack with documented processes; most users know where to find datasets.
  • Decisions: metrics linked to actions in some teams; learning loops starting to form.

Score 4 — Managed & proactive

  • Data quality: automated monitoring, alerts, and remediation workflows.
  • Ownership: clear role definitions with accountability and review processes.
  • Tools: integrated platforms, metadata catalog, and lineage information available.
  • Decisions: regular experiments, KPI owners, and measured outcomes across several areas.

Score 5 — Optimized & embedded

  • Data quality: high trust data used operationally; automatic self-healing where feasible.
  • Ownership: governance embedded in business processes; continuous improvement routines.
  • Tools: mature catalog, governed CI/CD for analytics, monitoring for model drift and pipeline health.
  • Decisions: analytics routinely drive prioritized action; measurable impact and value tracking.

Use these descriptions as a scoring rubric. If a domain sits between two levels, pick the lower score and note what is needed to reach the higher one — that gap becomes a candidate action.


Discussion

Comments and conversation will live here.