How to Build an Effective Data Governance Framework: Steps & Best Practices

Big data is booming in China, leading to the emergence of data governance as a key concept for digital transformation.

Enterprises are establishing dedicated data management departments to define data policies, standards, ensure data quality, maintain architecture, and provide platforms and tools.

Many big‑data vendors offer data‑governance products, but the real challenge lies in planning and consulting on enterprise data architecture, building a unified data‑control system, and establishing data standards and services to support business.

Data Governance Organizational Structure

The governance structure typically includes a Data Governance Committee, Data Management Department, Data Production Departments (business), Data Consumption Departments (business or customers), and a Data Development Department (often outsourced). The responsibilities and processes are illustrated in the diagram.

Data Asset Inventory and Standardization

Data asset inventory determines what data the enterprise has, its location, value, and ownership. It answers questions such as how much data exists, what types, where they are stored, and who is responsible.

How many data assets does the enterprise have?

What data does the enterprise possess?

What is the value of the data?

Where are the data stored and which are most valuable?

Who owns the data and who is responsible?

The result is a data asset catalog that provides a global view of data assets, supporting effective data utilization, management, and security.

Six‑Stage Data Asset Inventory Process

Build data standards: define unified data definitions and value frameworks based on industry and business specifics.

Data discovery: scan the enterprise to locate all data assets and create a storage distribution map.

Data definition: identify and describe data, forming a database‑table‑field hierarchy.

Classification and grading: classify data by business and grade by value, guided by standards.

Ownership clarification: determine business ownership and responsible parties for each asset.

Data asset catalog: compile a comprehensive map showing content, volume, value, location, and ownership.

Data Standard Development

Data standards consist of basic standards (data elements, code sets, encoding sets) and indicator standards (metric systems). Code sets define permissible values for a data element (e.g., gender). Encoding sets specify naming conventions. Data elements are the smallest indivisible units with defined attributes.

Common Issues in Data Standardization

1. Handling duplicate data across systems: Identify an authoritative source (e.g., public security for population data) and align other systems to it.

2. Forming data sets from multiple data elements: A data element can belong to multiple data sets across tables.

3. Categorizing data sets into domains: Map resources to departments and systems, then create domain‑based standards and services.

4. Defining data quality rules: Separate syntactic rules (length, type, range) from semantic rules (consistency, completeness) and implement them at the database level.