Metric System Analysis and Growth Practice: Building Indicator Systems, Designing Optimization Strategies, and Real-World Case Studies

Today’s topic is metric system analysis and growth practice. Volcengine, ByteDance’s cloud service platform, shares its accumulated experience through growth methods, technologies, tools, and capabilities to help external enterprises achieve continuous growth during digital transformation. Two data products are introduced: the behavior analysis platform (DataFinder) and the A/B testing platform (DataTester).

The session covers four main parts: (1) building an indicator system with a real case to identify bottlenecks; (2) designing strategic growth optimization methods; (3) illustrating growth thinking with a client case; and (4) a Douyin Group case study.

1. Building an Indicator System

Key steps include defining a North Star metric that reflects core product value (e.g., total GMV for a cashback business), decomposing it into related factors (traffic, activation rate, retention, etc.), and mapping each factor to responsible departments. A concrete example shows how GMV is broken down into e‑commerce and food‑delivery components, further split into user segments and conversion rates.

Based on the decomposition, a comprehensive indicator system is designed covering basic activity, user experience, user operations, and product revenue. After the system is built, data collection points are defined and SDKs are used to report the required events.

2. Designing Strategy Growth Optimization

The session introduces the Lift model from "Measuring Conversion Rate" and outlines six principles for effective A/B testing: value proposition, relevance, clarity, attention, anxiety, and urgency. Each principle is illustrated with visual examples and practical guidelines.

3. Indicator Growth Thinking

Using the earlier example, three A/B experiments are described to improve the low delivery page arrival rate: enhancing the entry tab, optimizing the landing page with clearer value propositions, and highlighting the cashback offer. Experiment results show measurable improvements.

4. Douyin Group Case Practice

Two case studies are presented: (1) Dongchedi, which increased short‑video playback by 300% through metric decomposition and A/B testing; (2) Douyin, which raised login rates by targeting older users and improving privacy consent prompts, achieving a 0.5% lift that translates to millions of users.

The overall workflow is: build a small growth team, define the North Star metric, decompose indicators, locate bottlenecks via dashboards, run A/B experiments, analyze results, and iterate. Regular growth retrospectives and continuous optimization complete the cycle.