Mastering Data Governance: A Complete Guide to Metadata, Standards, Quality, and Security

Metadata

Metadata is information about data organization, domains, and relationships; essentially, data that describes other data.

Metadata management ensures correct creation, storage, and consistent definition of data across the enterprise.

Types

Business metadata, technical metadata, and operational metadata, with their collection details shown in the accompanying table.

Metadata Management Five Steps

(1) Define metadata strategy: plan, collaborate with stakeholders, assess existing resources, and set strategic goals.

(2) Understand metadata requirements: capture frequency, synchronization, history, access control, storage, operations, management, and quality needs.

(3) Define metadata architecture: choose centralized, distributed, or hybrid models based on technical frameworks.

(4) Create and maintain metadata: integrate technical metadata with business and operational metadata for unified management.

(5) Query, report, and analyze metadata: provide front‑end visualization tools for querying and analysis.

Master Data

Master data describes core business entities such as customers, partners, employees, products, and materials, and is reused across multiple systems.

Master data management ensures completeness, consistency, and accuracy of these entities.

Implementation Architecture

Four steps: current state analysis, planning management system, building implementation plan, and platform deployment.

Key Implementation Phases (10 Important Links)

(1) Master data standardization system, (2) Classification design principles, (3) Coding design, (4) Attribute standards, (5) Governance process design, (6) Historical data integration, (7) Data migration strategies, (8) Production and maintenance strategies, (9) Distribution strategies, (10) Integration examples.

Data Standards

Data standards ensure consistency and accuracy for internal and external data use and exchange.

Management combines policies, processes, and tools to standardize definitions, classifications, and codes.

Classification

(1) Business standards, (2) Technical standards, (3) Management standards.

Implementation Steps

(1) Define goals and scope, (2) Conduct standards research, (3) Clarify organization and processes, (4) Draft and publish standards, (5) Promote standards internally, (6) Deploy standards platform and operate.

Data Quality

Data quality means data meets the needs of its consumers and business scenarios, covering both data itself and process quality.

Data quality management identifies, measures, and monitors issues throughout the data lifecycle to continuously improve quality.

Common Quality Issues

Missing data

Abnormal data

Inconsistent data

Duplicate or erroneous data

Six Quality Dimensions

Based on GB/T36344‑2018, the six dimensions are illustrated in the accompanying diagram.

Data Quality Management Seven Steps

(1) Define high‑quality data, (2) Define data‑quality strategy, (3) Identify key business and quality rules, (4) Conduct initial quality assessment, (5) Identify improvement directions and prioritize, (6) Set improvement targets, (7) Develop and deploy quality operations.

Data Assets

Data assets are data resources owned or controlled by an organization that generate economic value.

Data asset management involves planning, controlling, and delivering activities to protect, deliver, and enhance asset value.

Asset Inventory

(1) Top‑down business view: analyze policies, processes, and documents to build a hierarchical catalog.

(2) Bottom‑up technical view: trace from IT systems to database tables and structures.

Asset Catalog

Provides visibility into data location, ownership, and usage.

Four Asset Management Steps

(1) Overall planning, (2) Management implementation, (3) Audit and inspection, (4) Asset operation and value realization.

Data Exchange

Data exchange enables sharing of data between different information systems through defined principles and technologies.

Methods

Electronic or digital file transfer (e.g., FTPS, HTTPS, SCP)

Portable storage devices (e.g., USB, DVD)

Email attachments

Database sharing

File‑sharing services (e.g., Dropbox, Google Drive, OneDrive)

Five Principles

Consistency: source unit ensures accuracy.

Black‑box: consumers need not know technical details.

Agile response: subscribe to services instead of rebuilding integrations.

Self‑service: providers focus on supply, not consumption.

Traceability: providers can track who uses the data.

Data Security

Data security protects digital information assets from unauthorized access, disclosure, modification, or theft.

Security governance combines organizational, policy, technical, and operational capabilities.

Organizational Governance

Five‑layer structure: decision, management, execution, supervision, and participation.

Policy Governance

Four‑layer policy system covering strategy, standards, compliance, and risk.

Technical Capabilities

Measures across the data lifecycle: classification, database auditing, encryption, leakage prevention, masking, watermarking, behavior analysis.

Operational Capabilities

Mechanisms for risk assessment, incident response, monitoring, and audit to build a long‑term security operation framework.

Data Lifecycle

The data lifecycle includes collection, storage, integration, presentation/use, analysis/application, archiving, and destruction.

Lifecycle management governs data flow from creation to deletion.

Common Models

Various academic and industry models are illustrated in the diagram.

Four Phases

“In” phase: planning, standard definition, and governance before data creation.

“Store” phase: choose appropriate storage and compute engines based on structure, timeliness, performance, and cost.

“Use” phase: emphasize data reuse to reduce cost and increase efficiency.

“Out” phase: archive to low‑cost storage or securely destroy data following approved procedures.

Platform Coverage

A comprehensive data‑governance platform can integrate all the above modules, enabling efficient, end‑to‑end governance.