Case Study: Rokt achieves real-time reporting and lower storage costs with ClickHouse

NYC Meetup Report Real-time Slicing and Dicing Reporting with ClickHouse (Rokt)

Rokt, a global marketing technology company, needed a better way to power reporting for internal analysts, account managers, external users, and other systems. Its existing architecture, which relied heavily on Elasticsearch along with Kafka, Spark, S3, Redshift, and batch pipelines, made it difficult to support flexible slicing, dicing, joins, and custom reporting without adding burden to the reporting team.

Rokt evaluated several database options and chose ClickHouse for its SQL support, real-time analytics capabilities, strong ingestion from Kafka, and more efficient storage and performance. With ClickHouse, Rokt achieved predictable query performance, supported up to 200 queries per second with a maximum query time of six seconds, and reduced storage needs versus Elasticsearch by about 6x; the company also found ClickHouse about 3x less expensive than Redshift in its benchmark.

Rokt

Vadim Semenov

Rokt