5.1 Business, Technical, and Operational Metadata

Introduction to Metadata Management

In the DAMA DMBoK2 framework, Metadata is defined as "data about data," or more precisely, structured information that describes, explains, locates, or makes it easier to retrieve, use, and manage an information resource. Without metadata, data remains a raw, uninterpretable stream of characters and numbers. Metadata provides the context that transforms raw data into a valuable business asset. The primary goal of metadata management is to ensure that an organization's metadata is cataloged, integrated, accurate, current, and accessible to both business users and technical professionals.

To pass the Certified Data Management Professional (CDMP) exam, candidates must understand the classification of metadata into three primary types: Business Metadata, Technical Metadata, and Operational Metadata. Each serves a distinct purpose, addresses a specific audience, and is captured from different points in the data lifecycle.

---

Business Metadata: Context and Meaning

Business Metadata focuses on the semantic meaning, business context, and governance rules associated with data elements. It ensures that business users interpret data consistently across the enterprise. It is primarily managed inside a Business Glossary and is championed by data stewards.

Key components of Business Metadata include:

• Business Glossaries: Standardized definitions of business terms, acronyms, synonyms, and abbreviations.
• Governance Roles & Ownership: Identification of Data Stewards, Data Owners, and system caretakers.
• Security & Privacy Classifications: Tags indicating compliance requirements, such as Personally Identifiable Information (PII), Payment Card Industry Data Security Standard (PCI-DSS), or GDPR sensitivity levels.
• Business Rules & Calculation Logic: The non-technical descriptions of how metrics are calculated (e.g., "Monthly Active Users" defined as users who logged in at least once in a 30-day billing cycle).
• Subject Area Mappings: High-level groupings linking business terms to specific domains (e.g., Customer, Finance, Product).

Worked Example: A financial analyst looks up the term "Gross Margin." The Business Metadata provides the definition ("Revenue minus Cost of Goods Sold divided by Revenue"), designates the Finance Director as the Data Owner, and flags the data as "Confidential."

---

Technical Metadata: Structure and Systems

Technical Metadata describes the physical and structural properties of data assets, including their design, physical location, and mechanical behavior. It provides database administrators (DBAs), developers, and architects with the details necessary to construct, optimize, and maintain data pipelines and storage systems. It is primarily housed in a Data Dictionary or system catalogs.

Key components of Technical Metadata include:

• Physical Schema Definitions: Table names, database views, partition keys, and schemas.
• Column Attributes: Physical column names (e.g., cust_id_pk), data types (e.g., VARCHAR(64)), precision, scale, and nullability constraints.
• Keys and Relationships: Primary Keys (PK), Foreign Keys (FK), unique indexes, and referential integrity constraints.
• Data Integration Logic: Physical mappings, Extract, Transform, Load (ETL) scripts, SQL queries, and API endpoints.
• Physical Storage Paths: File formats (e.g., Parquet, CSV), directory structures, and server addresses.

Worked Example: A database developer views a physical table named T_FIN_SALES. The Technical Metadata reveals that the column SALES_AMT is a DECIMAL(18,4) data type, has a default value of 0.0000, and holds a foreign key relationship pointing to T_DIM_CUSTOMER.CUST_ID.

---

Operational Metadata: Execution and History

Operational Metadata details the run-time history, processing events, and performance parameters of data systems. It describes how data was processed, when it was modified, and by what mechanism, helping operational teams monitor system health, audit data access, and debug failed execution runs.

Key components of Operational Metadata include:

• Batch Run Information: Job names, execution start and end timestamps, execution durations, and status codes (e.g., Success, Failed, Retrying).
• Data Volumes: Read/write record counts, bytes transferred, and delta size changes.
• Error and System Logs: Exception stack traces, CPU/memory usage statistics during ingestion, and queue latencies.
• Access Audits: Auditing details logging who read or modified a specific data store, when, and from which IP address.
• Data Ingestion Parameters: Version control hashes of ETL scripts used in a specific execution, connection strings, and active environmental variables.

Worked Example: The data operations team checks why a dashboard did not refresh. The Operational Metadata shows that job JOB_LOAD_FIN_SALES_DLY failed at 2026-07-18 04:12:00 UTC due to a database timeout, after successfully reading 450,000 records but writing 0 records.

---

Comparative Matrix: Business, Technical, and Operational Metadata

---

Harvesting Metadata Sources

Metadata cannot be manually input at scale; it must be harvested or ingested from a variety of enterprise sources automatically. Harvesting involves using software connectors to extract metadata from:

1. Data Modeling Tools: Captures logical entities, attributes, and relationships.
2. Database Management Systems (DBMS): Harvests system tables (e.g., INFORMATION_SCHEMA) to acquire schema layouts and constraints.
3. Data Integration (ETL/ELT) Tools: Extracts transformation logic, mappings, and pipeline routing.
4. Business Intelligence (BI) Applications: Captures reporting schemas, user access patterns, and calculated fields in visual sheets.

---

Enterprise Search and Data Discovery

To realize the value of harvested metadata, organizations deploy Enterprise Search engines integrated with a Data Catalog. A data catalog functions as an e-commerce catalog for data, allowing analysts to search, discover, evaluate, and request access to data assets.

Enterprise search utilizes indexes built from business, technical, and operational metadata. When an analyst searches for "Revenue," the catalog uses:

• Semantic Search: Matches synonyms (e.g., "Income", "Turnover").
• Faceted Search: Filters results by source system, business owner, data type, or quality score.
• Relevancy Scoring: Ranks search results higher if the asset has active stewardship, high operational use (popularity), and high data quality scores.

[!WARNING] CDMP Exam Trap: Do not confuse Reference Data (valid values, e.g., Country Codes or Currency Codes) with Metadata. Reference data is a component of data itself used to categorize or classify transactions, whereas metadata describes the structure, meaning, and administration of that reference data.