Case Study: Nielsen achieves real-time audience analysis and scalable COUNT DISTINCT performance with Imply Druid

How Nielsen Marketing Cloud Uses Druid for Audience and Marketing Performance Analysis

Nielsen Marketing Cloud, a Nielsen division, needed to analyze massive datasets to provide its customers with real-time audience sizing and marketing performance insights. They faced a significant challenge with their previous Elasticsearch-based system, which was unable to efficiently handle COUNT DISTINCT queries across over 10 billion devices, leading to slow ingest times of 10 hours per day and query timeouts.

They implemented a solution using Imply's Druid, a real-time analytics database. This allowed them to ingest dozens of terabytes of data daily and leverage Druid's rollup and ThetaSketch capabilities for fast, approximate distinct counts. The move to Imply's Druid resulted in massive scalability, the ability to handle concurrent queries, and a significant reduction in storage needs, condensing a year's worth of data into just 40 terabytes. This provided their customers with the real-time, complex audience analysis they required.

Nielsen

Itai Yaffe

Big Data Group Lead