Data Governance: Concepts, Goals, Methodology, Tools, and Enterprise Case Studies

What is Data Governance?

Data governance is the process of moving from fragmented data usage to unified data, establishing enterprise‑wide controls, and turning chaotic data into orderly assets.

It covers the entire data lifecycle from front‑end business systems, back‑end databases to analytical endpoints, forming a closed‑loop feedback system that supervises data acquisition, processing, and usage.

Guided by organizational strategic goals, it involves collaborative effort, process definition, and activities such as data asset inventory, collection, cleansing, structured storage, visual management, and multidimensional analysis to create business value, innovate models, and control risks. It is a continuous service, not a one‑off project.

Why Implement Data Governance?

After 30 years of information system construction, enterprises and governments have accumulated massive, diverse data assets that are difficult to use.

Business systems built for specific needs become siloed when the environment changes, hindering cross‑departmental collaboration.

Massive, scattered data leads to inconsistency and quality issues, limiting deep data utilization and preventing business model innovation and risk control.

Goals of Data Governance

Data governance itself is a means, not an end, to achieve organizational strategic objectives.

Goals differ by organization size and function:

Headquarters or government big‑data bureaus aim to define data policies, ensure data security, and enable seamless data sharing, focusing on strategic implementation.

Business units aim to improve information management, operational efficiency, decision‑making capability, and core competitiveness, emphasizing data value extraction, business model innovation, and risk control.

Contents of Data Governance

Based on GB/T 34960, the framework includes top‑level design, governance environment, governance domains, and governance processes.

Top‑level design defines organizational authority, governance objectives, and actionable paths.

Governance environment analyzes stakeholder needs, identifies support and resistance, and establishes supporting policies.

Governance domains set standards for data quality, security, and management systems, building sharing, service, and analysis layers.

Governance process follows a PDCA cycle: plan goals, design architecture, collect and cleanse data, store core data, manage metadata and lineage, and check alignment with objectives.

Data Governance Methodology

Recent industry practice emphasizes four platform capabilities—aggregation, governance, integration, and utilization—guided by the PDCA principle.

Aggregation: ability to gather data.

Governance: standards, quality, metadata, security, lifecycle, master data.

Integration: ability to connect dispersed source data.

Utilization: delivering data services that empower front‑end business.

PDCA: plan (standards, planning, processes), do (tool‑assisted implementation), check (dual technical‑business verification), action (continuous optimization).

The “PAI” methodology (process‑oriented, automation, intelligence) incrementally enhances governance capability.

Tools Required for Data Governance

From a technical perspective, data governance involves five steps: collect, store, manage, use, and govern.

Data asset inventory: cataloging organizational data resources.

Data collection & cleansing: visual ETL tools (e.g., DataX, Pentaho) to extract‑transform‑load scattered data.

Core and subject‑area database construction: designing tables based on business understanding.

Metadata management: managing attributes and linking business meanings.

Lineage tracing: linking data usage back to source for error resolution.

Data catalog: automated request and usage based on business scenarios.

Quality management: multi‑dimensional checks (null, range, uniqueness, etc.).

Business intelligence (BI): self‑service reporting and analysis.

Data sharing & exchange: table, file, and API based sharing mechanisms.

Meituan Delivery Data Governance Practice

1. Define Standards, Improve Quality

Business, technical, security, and resource‑management standards are established to align metric definitions, modeling practices, data‑security controls, and resource budgeting.

2. Implement and Enforce

Standards are applied to achieve a “from chaos to order” transformation in data models, metadata, and security, ensuring consistency and traceability.

3. Tool Overview

Key tools include the Data Map (Wherehows) for metadata search and lineage, and QuickSight for self‑service data visualization and reporting.

Ant Financial Data Governance Practice

Challenges and Quality Governance Approach

Ant combines technology, data, and algorithms, facing rapid business cycles and the need for strong governance across both development and data flows.

Architecture

System‑level controls cover data testing, release governance, and change management, while production focuses on quality monitoring, fault‑injection drills, and continuous audit.

Quality Governance Solutions

Pre‑, in‑, and post‑production measures include requirement grading, automated detection, attack‑defense drills, and metric‑driven audits to ensure data reliability.

Summary

Standardization across business, technical, security, and resource domains ensures compliant data production, management, and usage.

Architectural techniques such as bridge tables and time‑bucketization improve model flexibility and data consistency.

Security measures protect sensitive data and enforce fine‑grained access controls.

Metadata pipelines connect collection, construction, and application, delivering data maps and visualization tools that solve “find data”, “use data”, and “impact assessment” challenges.