Solving App discovery through smarter user data analysis
summary
Company claims real results: built lock‑screen and launcher integrations (blink‑to‑selfie, glance‑pause video) that run on OEM phones and power app recommendations across 8 million devices. Pipeline uses Gaffer for collection, Spark for batch profiling, Redis for real‑time serving and an item‑item collaborative filtering model. Outcome is a monetization layer that reportedly reduced uninstall rates to 30–40% and drives targeted app installs rather than vague topical suggestions.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:23Introduction
- 1:59Data Collection
- 5:01App-app aka item-item Collaborative Filtering
- 7:51Store : App/Game/Deal Discovery
- 8:55Internet of Things
related talks
22mGamelight: Power of Game Recommendation Apps
42mPanel: The Ultimate Growth Hack: The Smart Approach to Paid & Organic User Acquisition Campaigns
20mWinning the Game: Leveraging Contextual Bandits for Creative Optimization
11mPersonalizing your app in real-time: Powered by Firebase and ML
16mThrough the looking glass – Following the rabbit down the data… Whole
31mLevelling up – How to move from a small indie to an established publisher
23mThe unreasonable effectiveness of data for better monetisation
06mINTERVIEW