Case Study: Conversant achieves real-time advertising performance and stability at scale with Cloudera

Empowering Real-Time Advertising with Performance and Stability at Scale

Conversant, a leader in personalized digital advertising that makes over a trillion real-time decisions daily for 160 million people, faced a scaling crisis: handling tens of billions of transactions every day with millisecond response requirements while growing ~30% annually. Legacy relational databases and an unstable big‑data platform caused performance problems, frequent outages and a lost year of progress.

Conversant partnered with Cloudera to deploy a modern data platform—one enterprise cluster for analytics and ML plus 100+‑node operational clusters for stream processing and storage of 150 billion daily events (powered by Spark, Kafka, HBase and Greenplum). The team migrated 60 billion consumer profiles in 24 hours, achieved millisecond query responses, improved governance and stability, and refocused engineering resources on new features and faster, smarter decision‑making.

Conversant

Patrick Jaromin

Director, Software Engineering, Ad Tech