Navigating the ocean of Android devices
summary
Plarium builds an API-driven pipeline that turns Google Play and public data into device targets: it groups ~16,000 models by SOC and GPU so SOC bucketing lets ten representative GPUs cover ~80% of GPU market, uses GFXBench/3DMark to pick the weakest GPU (Mali‑400) as baseline, and combines Device Catalog + Android Vitals to decide whether to fix a crash or exclude devices. Example finding: top‑10 devices only cover ~15% of market; a 1GB‑RAM cohort can be 13% of users but drive two thirds of revenue.
AI-written summary · about
chapters (11)
- 0:00Intro
- 3:10AN OCEAN OF DEVICES
- 3:56HOW DO WE MEASURE ALL OF THAT?
- 6:03WHAT IS COMMON TO ALL OF THESE?
- 7:13THE SOC APPROACH (System-on-a-Chip)
- 7:58GPU MARKET FRAGMENTATION
- 10:22OS MARKET FRAGMENTATION
- 12:41SCREEN SIZE FRAGMENTATION
- 14:42GOOGLE PLAY CONSOLE: DEVICE CATALOG
- 17:06EVEN BETTER WITH DEVICE CATALOG
- 20:05TEST IN PRODUCTION WITH ANDROID VITALS
related talks
17mPlan and optimize your game for success on
24mBuilding an awesome game for Android
42mSupercharging Android Testing: Scalable Environments for Larg…
37mNew and advanced Google Play tools for game developers
17mBuilding the business case for quality
07mWhat's new in Android graphics optimization tools
40mAndroid vitals: Debug app performance and reap rewards
14mWhy you should be using Google Play Games services