Case Study: Gaia achieves a 50% lift in subscriber engagement and stronger data-driven decision-making with Qubole

Gaia Boosts Subscriber Engagement And Data-driven Decision-making With Qubole

Gaia, a member-supported streaming service available in 185 countries, relied on a slow, rule-based Postgres recommendation engine and an on‑prem single‑instance architecture that couldn’t handle mixed data sources, heavy workloads or frequent overnight jobs. The fragile system caused outages, long processing times, and limited analytics, preventing effective content personalization and timely business insights.

By moving to a Qubole data lake on AWS and rebuilding recommendations with Apache Spark and XGBoost, Gaia gained autoscaling, multi-source querying, and validated data pipelines. The new ML-driven engine and 66+ TB data lake boosted subscriber engagement (a 50% lift in average minutes watched), reduced engineer troubleshooting time by at least threefold, and enabled more reliable, data-driven decision-making.

Gaia

Patrick Lawlor

Product Data Analyst