Data Governance Practices at Meituan Hotel Travel Platform

1. Background

With the rapid growth of mobile internet, offline commerce has moved online, generating massive data. Companies, especially large internet firms, recognize data as a core asset and face quality, efficiency, and security issues throughout the data lifecycle, prompting the need for systematic data governance.

Meituan's hotel‑travel business became an independent unit in 2014 and saw exponential growth in data production tasks and volume between 2017‑2018, leading to five major problems in 2019: severe data‑quality issues, rapidly rising data‑costs, low data‑operation efficiency, insufficient data‑security controls, and lack of development standards.

2. Data Governance Practice

2.1 Governance Strategy

The governance covers the full data lifecycle and is divided into organization, standards, technology, and measurement. Standardization and organizational guarantees are prerequisites; technology implements the strategy, and a measurement system monitors effectiveness.

2.2 Standardization and Organizational Guarantees

A full‑link data standard is defined from collection to disposal, and a Data Management Committee (involving business and technical teams) ensures cross‑team alignment.

2.3 Technical Systems

Four unified systems are built:

One Model : unified warehouse modeling to enforce standards and remove redundancy.

One Logic : centralized metric definition and management to guarantee consistent metric logic.

One Service : a unified data‑service layer decoupling logic from APIs, providing monitoring and usage control.

One Portal : a single entry point for different user groups (decision analysis, sales, data‑asset queries) to locate data products easily.

Additional tools such as Dataman support standard enforcement, configuration‑driven ETL development, and rule‑based validation.

2.4 Data‑Operation Efficiency

By systematizing data‑asset information and building a knowledge base, the team reduces repetitive QA, improves data discoverability, and cuts engineers' operational time.

2.5 Cost Optimization

Cost reduction targets compute, storage, and log‑collection resources through task cleanup, long‑task optimization, better resource scheduling, storage compression, cold‑data handling, and lifecycle management.

2.6 Data Security

Security controls span encryption at source, tiered masking in the warehouse, and strict permission management, following a three‑layer system and five usage principles.

2.7 Measurement Indicators

A PDCA‑based metric framework monitors quality, cost, security, usability, and efficiency across data‑flow stages, enabling continuous improvement.

2.8 Governance Outcomes

Data quality improved, with a 60% reduction in task growth rate.

Data‑cost per unit dropped ~40%.

No major security incidents after encryption and masking.

Engineers’ daily QA time reduced by >60%.

3. Future Planning

Governance evolves through three stages: passive (ad‑hoc), proactive (systematic, standardized), and automated (intelligent, policy‑driven). Meituan is transitioning from proactive to automated, aiming to embed metadata‑driven policies and automatic remediation.

4. Author Information

Jian Shu, Wang Lei, and Luo Xi are data engineers and product managers in Meituan’s Data Science & Platform team, contributing to the described governance work.