How Dawn’s VTree Revolutionizes Cross‑Platform Event Tracking

Background

NetEase Cloud Music open‑sourced Dawn , a cross‑platform (iOS, Android, H5, React Native) full‑link event‑tracking framework. Dawn introduces a virtual tree (VTree) that represents the UI hierarchy as declarative objects, enabling stable, hierarchical tracking data for user profiling, real‑time and offline attribution, funnel analysis, and recommendation scenarios.

Industry Pain Points

High development cost – each tracking point requires repetitive parameter definitions.

Low precision – exposure tracking lacks a unified strategy across the app.

Unstable models – mspm or x‑path based schemes are fragile and hard to maintain.

Weak link tracing – client side does not capture end‑to‑end user actions, forcing costly offline analysis.

Quality control difficulty – QA effort is large while the benefit is limited.

Key Features of Dawn

High‑performance VTree : Views are declared as objects; the original view hierarchy is exported to a VTree with minimal overhead. The VTree serves as the backbone for all downstream tracking.

Structured tracking content : Because VTree is hierarchical, each node carries self‑describing metadata, allowing automatic generation of stable Super Position Model (SPM) and Super Content Model (SCM).

Unified exposure strategy : A cross‑platform rule starts exposure when a node becomes visible and stops when it is hidden, eliminating per‑screen exposure code.

Full‑link tracing : Client‑side tracing records 100 % of user‑log relationships, supporting unlimited depth analysis for attribution, funnel, and recommendation use cases.

Iteration management platform : A dedicated console stores historical VTree versions, supports parallel versioning, and visualizes changes for rapid app iteration.

Multi‑mechanism quality assurance : Offline validation removes the need for manual QA, while online automated audits detect exposure or data anomalies in real time.

Broad platform coverage : Native support for iOS, Android, RN, H5, and extensibility to Flutter, Weex, etc.

Using Dawn

Analysts design the tracking VTree in the EasyInsight platform. Client developers then generate the VTree with a few declarative statements, integrate it into the app, and rely on built‑in validation and online audit features to ensure data quality.

Industry Comparison

Dawn vs. X‑Path Series

VTree visualization – Dawn provides end‑to‑end visual flow; X‑Path series does not.

Visual tracking – Dawn supports H5 visual tracking; X‑Path series only in limited cases (e.g., Meituan).

Content structure – Dawn yields hierarchical, highly customizable data; X‑Path series produces flat, redundant parameters.

Model stability – Dawn’s SPM/SCM are highly stable; X‑Path relies on post‑processing and is less stable.

Development cost – Dawn uses a declarative approach with moderate effort; X‑Path is low cost but lacks client‑side tracing.

Full‑link tracing – Dawn offers unified tracing; X‑Path depends on downstream big‑data processing.

Exposure strategy – Dawn unifies exposure for lists and non‑lists across all platforms; X‑Path provides a weaker exposure model.

Validation & audit – Dawn includes visual validation and online audit; X‑Path offers limited support.

Dawn vs. mspm Series

VTree visualization – Dawn supports full‑process visualization; mspm series does not.

Content structure – Dawn delivers hierarchical, customizable data; mspm series uses flat, meaningless strings.

Development cost – Dawn’s declarative approach is moderate; mspm series incurs high client maintenance cost.

Full‑link tracing – Dawn provides robust client‑side tracing; mspm series lacks it.

Exposure strategy – Dawn unifies exposure across platforms; mspm series cannot.

High‑precision exposure – Dawn handles occlusion and achieves high precision; mspm series is weak.

Validation & audit – Dawn offers strong visual validation and online audit; mspm series provides weak support.

Dawn vs. Four‑Stage SPM

VTree visualization – Dawn supports full‑process visualization; four‑stage SPM does not.

Iteration management – Dawn enables visual, parallel version management; four‑stage SPM is limited.

Content structure – Dawn offers hierarchical, highly customizable data; four‑stage SPM is flat with redundant parameters.

Development cost – Dawn’s declarative approach is moderate; four‑stage SPM is low but sacrifices flexibility.

Full‑link tracing – Dawn provides powerful unified tracing; four‑stage SPM relies on downstream processing.

Exposure strategy – Dawn unifies exposure across all platforms; four‑stage SPM weakens exposure tracking.

High‑precision exposure – Dawn achieves high precision with occlusion handling; four‑stage SPM is weak.

Validation & audit – Dawn offers strong visual validation and online audit; four‑stage SPM is weak.

Adoption at NetEase Cloud Music

All major Cloud Music applications have migrated to Dawn, covering virtually all P0/P1 business scenarios across dual‑client, H5, and RN. The stable VTree structure and automated validation have eliminated QA labor and dramatically reduced data‑related bugs, especially those caused by inaccurate exposure tracking.

Community and Resources

GitHub repository: https://github.com/eventtracing/dawn

Official site & documentation: https://eventtracing.github.io

Code example

坑位mspm系列