Case Study: Manage achieves 10x faster data freshness (2 hours → 10–15 minutes) with MemSQL

How Manage Accelerated Data Freshness by 10x

Manage is a programmatic mobile marketing company that processes over a terabyte of data and 30 billion bid requests daily to deliver contextual ads for clients like Uber, Wish, and Amazon. As data volumes grew, their MySQL and later Hadoop/Hive pipelines introduced multi-hour delays, preventing the real-time freshness needed for timely bidding and analytics.

Manage implemented MemSQL with a Streamliner Apache Spark pipeline that ingests logs from Apache Kafka into the MemSQL columnstore, de-duplicates and aggregates data into summary tables, and exposes results to dashboards and APIs. The new architecture cut data freshness from about two hours to 10–15 minutes, enabled ad-hoc queries in seconds, and allowed the team to react to marketplace changes in real time.

Manage

Kai Sung

Co-Founder & Chief Technology Officer