How Kuaishou E‑Commerce Built a Data Metric System to Drive Growth

Introduction

This article shares Kuaishou E‑Commerce’s exploration and practice in building a data metric system.

Key Topics

Necessity of metric system construction

How to build an effective metric system

Kuaishou E‑Commerce metric system product practice

Metric system management

Summary

Q&A

1. Necessity of Metric System Construction

In the context of Kuaishou E‑Commerce, constructing a metric system is essential to provide data support for business decisions. Precise analysis of sales profit, user behavior conversion and other key indicators helps optimize product strategy, market positioning and overall operational efficiency.

1.1 Kuaishou E‑Commerce Business Overview

The business follows the “people‑goods‑scene” model. The content scene consists of live streaming and short videos; live streaming drives most transactions, while short videos focus on “planting grass” to attract users. The shelf scene uses product cards that connect with the content scene, enabling traffic sharing through store search, mall pages and “you may also like”. A complete user lifecycle moves from potential user to content consumption, then to planting, first purchase and repeat purchase, which can be abstracted into three stages: audience discovery, cultivation and conversion.

Based on the relationship between operating scenes and users, five major operational scenarios are identified:

Potential consumer identification and excavation

Content and commerce integration for recommendation

Persona and IP‑based marketing interaction

Merchant self‑sale and influencer distribution to drive conversion

Member operation and old‑customer recall to increase stickiness

2. Business Challenges

During data production, querying and usage, various roles encounter problems:

Data product managers evaluate metric reuse for new business needs, but low efficiency slows project speed and resource allocation.

Data engineers may lack clear definitions and scopes for technical metrics, leading to uncertainty in development.

After development, product managers must decide appropriate visualization styles for metrics.

Business users (e‑commerce operators) often face inconsistent metric names or vague definitions, hindering data understanding and decision‑making.

Managers at different levels may lack a holistic view of business data performance, limiting strategic planning.

Conclusion and Q&A

The article concludes with a summary of the metric system practice and a Q&A session (details omitted).