所有作品
案例 11
GA4, Meta Pixel & Custom Monitoring System
Implemented GA4 and Meta Pixel (including Meta Conversion API) across products, then built a custom Apps Script monitoring system that pulls GA4 data across multiple property IDs, sends conditional warning emails, monthly reports, and API error alerts to different PMs and myself.
GA4Meta PixelConversion APIApps ScriptGA4 Data APIMulti-propertyWarning EmailMonthly ReportProduct Health
- GA4 and Meta Conversion API giving more complete conversion data than pixel alone
- PMs receiving daily/conditional alerts for the specific products they own, not a shared inbox
- Monthly reports automatically generated and sent without manual effort
- Product type
- Analytics implementation + custom monitoring system
- Platform
- Web (GA4, Meta Pixel), Apps Script, GA4 Data API
- Timeline
- 2024–2026 · tracking and automation program
- Role
- Analytics implementation, Apps Script development, monitoring logic
- Team
- Self-directed + PM stakeholders
- Tools
- GA4, Meta Pixel, Meta Conversion API, Apps Script, GA4 Data API, Google Sheets
為什麼這個專案存在
Beyond designing product UX, I also took ownership of analytics implementation and product health monitoring. This included integrating GA4 and Meta Pixel (plus Meta Conversion API for server-side tracking) across products, then building a custom Apps Script system that actively monitors product health — pulling GA4 data across multiple property IDs and routing alerts to the right people.
我負責的部分
- GA4 implementation and event tracking setup
- Meta Pixel integration (client-side)
- Meta Conversion API integration (server-side pixel, bypasses browser blockers)
- Apps Script development — pulled GA4 data via GA4 Data API across multiple property IDs
- Built conditional alert logic: different events and thresholds trigger different warning emails
- Built trigger-count monitoring: tracks how many times each event fires, alerts on anomalies
- Built zero-activity (cold-spot) warning emails for purchase and event silence
- Built monthly report emails sent to different responsible PMs per product
- Built self-monitoring: API trigger error emails sent to myself when tracking calls fail
困難之處
- Each product has its own GA4 property ID — the monitoring script had to handle multiple properties and route alerts to the right PM, not a single inbox
- Meta Pixel alone is blocked by ad blockers and browser privacy settings — Conversion API was needed to ensure server-side event delivery for accurate conversion data
- Defining meaningful alert thresholds is a judgment call: too sensitive creates noise, too loose misses real problems
- Zero-activity needs to be tracked as a cumulative signal, not just a one-time check — a single silent day might be a timezone gap, three consecutive silent days is a real alert
- Apps Script has execution time and quota limits — the monitoring queries had to be efficient enough to run within those bounds
- Verifying that tracking and API calls are actually firing correctly is its own problem — needed a self-alerting system for when the monitor itself fails
如何完成
- 1Implemented GA4 event tracking across product surfaces (web, web view)
- 2Integrated Meta Pixel client-side and Meta Conversion API server-side for complete conversion coverage
- 3Wrote Apps Script to query GA4 Data API across all product property IDs
- 4Built conditional alert logic: event name + count + threshold combinations trigger different email templates
- 5Built zero-activity accumulator: tracks consecutive days of silence per product, escalates after threshold
- 6Built monthly report generator: summarises key events per property, sends to assigned PM
- 7Built API error handler: logs failed tracking calls, sends alert email to myself for investigation
- 8Iterated on alert thresholds and email copy based on real traffic patterns and PM feedback
背後的思考
Treated alert routing as a product design problem — 'who needs to know what, and when?' is a UX question as much as a technical one. Different PMs own different products; a general 'something went wrong' email to everyone would be ignored. Routing specific events to the right person made alerts actionable.
精選視覺
抽象示意圖——原創 mockup,非客戶截圖。
改變了什麼
- GA4 and Meta Conversion API giving more complete conversion data than pixel alone
- PMs receiving daily/conditional alerts for the specific products they own, not a shared inbox
- Monthly reports automatically generated and sent without manual effort
- Zero-activity alerts catching product-health issues before they surface in business reviews
- Self-monitoring catching tracking failures that would otherwise go undetected
我的收穫
This project crossed what most designers would call their scope — but it sits exactly at the intersection of product understanding and technical reach. Knowing which event to track, which threshold matters, and who needs to know is a design decision. Writing the code to make it happen is just closing the loop.