Data and Analytics Assessment

Metadata, Catalog and Lineage Maturity Assessment

A maturity assessment of metadata management, from metadata standards and the business glossary through catalog coverage, lineage, operational metadata, adoption and the use of metadata by AI.

Already have an account? Log in to start

At a glance

Questions: 44
Domains: 7
Working time: Initial assessment: 8 to 10 hours of facilitated sessions over one to two weeks with the data office, two or three stewards, data architecture and engineering, analytics and AI leads, and security. Evidence gathering: the metadata plan and standard, the glossary and critical data element list, catalog exports and coverage reports, lineage samples, change requests and usage analytics. Start with a 30-minute session with every area for the three all-area questions. Reassessment: 4 to 5 hours.
Repeat: Annually, and after a catalog platform change, a major data platform migration, a new AI program or a significant data incident traced to missing or wrong metadata.
Roles: 6
Updated: October 2026

Who it is for

Data Office
Business Data Owners and Stewards
Data Architecture and Engineering
Analytics and AI Teams
Security, Privacy and Compliance
Process Owner/Full Model Assessor

Primary: head of data governance, data architecture lead and the catalog product owner. Secondary: business data owners and stewards, data engineers, analytics and AI teams, security and privacy teams and internal audit.

What you will gain

Shows whether metadata is complete, trusted and used, where the catalog exists only as a tool rather than a working practice, and which gaps to close first. Gives the data office a prioritized roadmap for the glossary, catalog and lineage, and a baseline that Master Data, Classification and AI teams can rely on.

About this assessment

Many organizations buy a data catalog, scan their main databases and then watch usage fade: definitions are missing or disputed, owners are blank, lineage stops at the warehouse and nobody checks whether the metadata is still true. This assessment tests whether metadata is managed as a product: a plan and metamodel, clear naming and definition rules, an approved glossary with critical data elements, a catalog that covers the estate and is curated and certified, lineage that is captured automatically, validated and used before change, operational signals that show whether data is fresh and fit, and adoption that is measured. It also tests the newest uses of metadata: dataset cards, AI-drafted metadata with human review, governed access for AI agents, metadata as trusted context for AI assistants, and controls driven automatically by metadata tags. It is a deep dive in the ITSG data family. Ownership, critical data criteria and literacy come from Data Governance and Management; quality rules stay with Master Data and Data Quality Management; classification levels stay with Data Classification and Handling. The model is neutral to jurisdiction and vendor.

What we assess: 7 domains

Each domain is a group of related questions, scored on its own and rolled up into your overall result.

01 · 6 questions

1 Metadata Strategy and Standards

Metadata management plan, roles and ownership of metadata, metamodel, naming and definition rules, required attributes, tool strategy and open standards

02 · 7 questions

2 Business Glossary and Semantics

Term workflow, approved terms, critical data elements, synonyms and hierarchies, glossary use in delivery, alignment with the enterprise information model, semantic model for AI

03 · 7 questions

3 Data Catalog Coverage and Curation

Catalog scope and coverage, automated harvesting, asset record completeness, term-to-asset links, curation workflow, certification, deprecation

04 · 7 questions

4 Lineage and Impact Analysis

Lineage requirements, automated capture, column-level lineage for critical data elements, lineage validation, impact analysis before change, report lineage, lineage to AI models

05 · 5 questions

5 Operational Metadata and Observability

Freshness, volume and run metadata, profiling statistics, quality status and data contracts in the catalog, schema change alerts, quality of the metadata itself

06 · 6 questions

6 Discovery, Adoption and Performance

Search and discovery, classification shown on assets, access requests, role-based adoption, usage analytics and feedback, performance reporting and improvement

07 · 6 questions

7 AI and Active Metadata

Dataset cards, AI-generated metadata with human review, agent-readable catalog, metadata as AI context, metadata-driven automation, cataloging AI-created data

KPIs to track

Measures that show whether the improvements hold.

KPI
Catalog Coverage - Share of in-scope systems with their data assets registered in the catalog. Target: at least 90%.
Asset Record Completeness - Share of catalog assets with all required attributes, including a named owner. Target: at least 90%.
Critical Element Lineage - Share of critical data elements with complete column-level lineage from source to consumer. Target: 100%.
Changes with Impact Analysis - Share of data-affecting changes with a lineage-based impact analysis before approval. Target: at least 95%.
Active Catalog Users - Share of target data users who used the catalog in the last 90 days. Target: at least 70%.

Framework alignment

Questions are written in plain language and mapped to:

ISO/IEC 11179-1:2023
ISO/IEC 11179-3:2023
DAMA-DMBOK2 Revised
COBIT 2019
W3C DCAT Version 3
OpenLineage specification
ISO 15836-1:2017
ISO 8000-150:2022
ISO 8000-61:2016
ISO/IEC 8183:2023
Open Data Contract Standard v3.1.0
ISO/IEC 42001:2023
ISO/IEC 5259-1:2024
MLCommons Croissant 1.1
NIST AI RMF 1.0
OWASP Top 10 for Agentic Applications 2026
ISO/IEC 27001:2022

Works best with

Practices this assessment relies on. Assessing them too gives a fuller picture.

Data Governance and Management (v1.9.6): owns the Data Ownership Register, critical data criteria, charter, strategy and literacy program; this model uses them as Inputs (1.1, 1.2, 1.5, 2.1, 2.3, 3.3, 6.3, 6.4).

Enterprise Architecture (v7.9.6): owns the Enterprise Information Model; glossary terms align with its subject areas (2.6). EA cites the Business Glossary, Approved Terms section owned here.

AI Governance (v11.9.6): owns the AI Dataset Catalog; this model links AI lineage and dataset cards to it rather than duplicating it (4.7, 7.1).

AI Pipeline and MLOps (v11.9.6): captures catalog entries and lineage for ML pipelines using this model's Metadata Standard (4.2, 4.7).

Data Classification and Handling (planned rebuild): owns classification levels and labels; this model tests that they are shown on assets and drive controls (6.2, 7.5).

Master Data and Data Quality Management (planned rebuild): owns quality rules and uses the critical data element list held in the Business Glossary (2.3, 5.3).

Data Engineering, Platforms and Observability (v1.9.6) and Analytics, BI and Insight Delivery (v1.9.6): own pipeline observability and metric definitions; this model tests that their signals and lineage reach the catalog (5.1, 5.4, 4.6).

Recommended: Run after Data Governance and Management; run before Master Data and Data Quality Management and Data Classification and Handling so they can rely on the glossary, catalog and lineage.