Case Study: Propeller Ads achieves near-real-time reporting and 10x faster analytics with Vertica

Introducing Vertica allows Propeller Ads to continually improve the performance of its advertising network

Propeller Ads is a multichannel advertising platform serving over 100 million users in 195 countries, handling about 3 billion ad requests a day and generating 2 TB of new data daily. The company needed real-time analytics and high-performance processing to run machine‑learning driven ad optimization, but its MySQL‑based data warehouse was costly to scale and couldn’t meet latency and analytics demands.

Propeller Ads prototyped and then migrated to the Vertica Analytics Platform—implementing a lambda architecture with Vertica FlexZone, preserving Tableau and ETL integrations, and enabling TensorFlow‑based ML—over a six‑month rollout. The move consolidated a 90 TB data lake, cut query times by 10x to 100x, delivered near real‑time reports (≤1 minute for complex queries), expanded customer metrics to about 100 (a ten‑fold increase), and enabled continuous improvement of ad rotation and performance.

Propeller Ads

Jacob Becker

Vice President