Founded in 2011, GameAnalytics was the first analytics provider specifically for the gaming industry, and it now collects data from over 100,000 games played by 1.75 billion people monthly. Initially, GameAnalytics faced scalability challenges with their data architecture, which relied on Amazon DynamoDB for historical queries and an in-house system for real-time queries. The limitations of this setup, particularly with ad-hoc analysis, prompted a switch to Druid, a database designed for flexible data exploration without the need for pre-computation. Druid's architecture, which separates ingestion and queries into different nodes and supports multi-dimensional filtering, enabled better scalability and faster query responses, with response times averaging 600 milliseconds. Additionally, Druid's integration with AWS tools and ability to unify real-time and historical data allowed GameAnalytics to streamline their infrastructure, reducing their data footprint and improving cost efficiency. After transitioning to Druid, GameAnalytics now ingests 25 billion events daily and can handle up to 100,000 queries per hour, enabling them to focus on delivering valuable features to their customers.