Mastering Growth Metrics: Methodologies, Frameworks, and Real‑World Cases

Growth Analysis Basic Methodology

Douyin and Toutiao growth teams focus on user scale, tracking active metrics such as DAU and MAU. A mathematical model divides active users into three parts: newly added users, retained old users, and recalled churned users, emphasizing different treatment for new and old users.

Growth Theory Models

The classic AARRR (pirate) funnel—Acquisition, Activation, Retention, Revenue, Referral—guides growth work, while the newer RARRA model shifts focus to retention and activation first, reflecting the diminishing traffic dividend.

Acquisition: ads, offline promotion, landing‑page optimization.

Activation: guiding users to discover product value.

Retention: push, SMS, cold‑start protection, customer service.

Revenue: paid content, premium services, ads, virtual currency.

Referral: incentives, sharing mechanisms.

Organizing Growth Work by Function

Growth tasks (acquisition, activation, retention, revenue, referral) are mapped horizontally, while product, operations, and advertising roles are mapped vertically, each owning specific actions at every stage.

How to Build a Growth Metric System

The North‑Star metric (One Metric That Matters) is the most critical indicator for the current stage, aligning the team around a single goal and clarifying priority. Multiple North‑Star metrics may exist for different sub‑domains (e.g., product experience, user activity, business health).

Metric hierarchy includes three layers: T1 strategic layer (company‑wide goals), T2 tactical layer (business‑line objectives), and T3 execution layer (operational actions).

Metric decomposition uses three models:

OSM (Objective‑Strategy‑Metric) – ideal for end‑to‑start goal breakdown.

UJM (User Journey Map) – maps user actions across a defined usage flow.

Scenario‑based analysis – applies OSM/UJM to concrete business scenarios.

Metric Analysis Closed Loop

The analysis loop consists of four steps: discover problems via metrics, formulate hypotheses, validate hypotheses through A/B experiments, and roll out successful strategies while iterating back to problem discovery.

Case Study: Improving Video Submission Rate

The case focuses on the “quick‑shoot” version of a video app, where submission rates lag behind the standard version. By decomposing the submission funnel (start shoot → shoot page → edit page → publish) and instrumenting events for each step, the team identified low penetration at the start‑shoot stage.

Hypotheses included poor placement of the shoot button, insufficient entry points, and weak user incentives. Strategies such as moving the button to a prominent plus‑icon, adding entry points on multiple pages, and offering rewards were tested via A/B experiments, confirming the hypotheses and leading to a significant increase in submission rates.

New Growth‑Metric Analysis Scenarios

Emerging scenarios include multi‑model metric joint analysis, ultra‑fine‑grained real‑time cohort comparison, full‑journey data integration across apps, services, and CRM, intelligent alerting, and industry‑specific knowledge sharing.

Joint analysis of multiple models on unified dimensions.

Real‑time cohort comparison with dynamic audience definitions.

Cross‑system user journey analysis linking app, e‑commerce, and support data.

Smart anomaly detection and diagnostic alerts.

Industry templates for rapid onboarding.

Professional Edition Upgrade

The professional edition enhances data ingestion (including e‑commerce, call‑center, offline channels), analysis capabilities (enhanced space, aggregated tools), industry knowledge (templates for gaming, finance, automotive, retail), and data applications (CDP tag sharing, GMP integration, ABI visualisation).

Comprehensive data integration across touchpoints.

Advanced analytical models and dashboards.

Industry‑specific templates and best‑practice guides.

Seamless data application to marketing and product actions.