Case Study: AppsFlyer achieves real-time analytics, 10× data aggregation and $20K/month savings with MemSQL

AppsFlyer Shares Journey to Real-Time Analytics Using MemSQL

AppsFlyer, an Israel‑based SaaS mobile marketing analytics and attribution platform, was aggregating roughly 8 billion events per day but struggled with unresponsive dashboards, faulty models, slow feature rollout and risky data-loss scenarios. They needed a fast, stable in‑memory database that scaled in real time and handled many more data dimensions; trials with Druid, MongoDB, Redis, Cassandra, Amazon Redshift and others failed to meet requirements.

Twingo introduced MemSQL, which AppsFlyer implemented for real‑time aggregation alongside batch processing with Apache Spark. MemSQL offered recoverability, the ability to reset to zero, supported 30× more data than their prior system, and enabled 10× higher aggregation with much faster response times—while saving about $20,000 per month and allowing AppsFlyer to store and monetize more data to meet product demands.

AppsFlyer

Yulia Trakhtenberg

Data Team leader