Glossary/Data Governance & Quality

Metadata Management

Metadata management is the systematic collection, organization, and maintenance of metadata (data about data) to enable discovery, governance, and understanding of data assets.

Metadata is information about data: table names, column names, data types, ownership, lineage, freshness, quality metrics, and business definitions. Metadata management is the practice of systematically tracking and maintaining this information. Without metadata management, metadata is scattered: in source systems (schemas), documentation (wikis, spreadsheets), and people's heads. Metadata management centralizes it into organized systems where it's discoverable and enforceable.

Metadata management emerged because metadata is valuable but often neglected. Organizations spent resources building pipelines and warehouses but didn't systematically document what they built. This created information silos: teams didn't know what data existed, who owned it, or whether to trust it. Metadata management solves this by making metadata as important as the data itself.

Metadata management includes both technical metadata (schema information, data types, lineage) and business metadata (definitions, ownership, quality). It also includes operational metadata (when data was last refreshed, error counts). Modern metadata management is often automated: lineage systems infer relationships from code, data catalogs auto-discover schemas, quality tools capture metrics. This reduces manual effort and keeps metadata current.

Key Characteristics

  • Systematically collects metadata from multiple sources
  • Maintains technical, business, and operational metadata
  • Enables search and discovery through catalogs
  • Tracks lineage and data relationships
  • Automated collection where possible
  • Version-controlled and auditable

Why It Matters

  • Discovery: Metadata enables finding relevant data quickly
  • Trust: Complete metadata builds confidence
  • Governance: Metadata enables policy enforcement
  • Compliance: Audit trails and ownership tracking
  • Efficiency: Documentation reduces time to understand data

Example

Metadata management tracks: table customers (owner: product team, SLA: daily update, last update: 2 hours ago), column email (type: varchar, nullable: false, lineage: from CRM system), and metric monthly_active_users (definition: distinct users with activity last 30 days, fresh within 24 hours).

Coginiti Perspective

Coginiti manages metadata across both technical and business layers. CoginitiScript's #+meta block captures authoring, versioning, and descriptive metadata for transformation logic. SMDL files store semantic metadata: entity definitions, dimension types, measure aggregation rules, and relationship cardinality. The Analytics Catalog organizes these metadata-rich objects across workspaces, making them discoverable. Query tags on Snowflake, BigQuery, and Redshift add operational metadata by tagging execution with project and department identifiers.

See Semantic Intelligence in Action

Coginiti operationalizes business meaning across your entire data estate.