Case Study: Reckitt achieves demand forecasting at scale and 10x capacity with Databricks

Achieving demand forecasting at scale

Reckitt, a multinational consumer goods company serving millions of retail customers—especially through a highly fragmented “traditional trade” channel of small neighborhood stores—struggled to forecast demand due to massive, diverse data streams, hundreds of disjointed pipelines, and a costly, hard-to-scale Hadoop infrastructure that slowed experimentation and delivery of insights for 16,000+ field reps.

By adopting the Databricks Lakehouse (with Delta Lake, collaborative notebooks, and automated cluster management), Reckitt consolidated pipelines, compressed storage, and sped up processing. Results included shrinking data from 80TB to ~2TB (≈98% compression), halving 24x7 pipeline runtimes (about 24 to 13 hours), lowering operational costs, and increasing capacity to support stores by roughly 10x (from ~45,000 to nearly 500,000).

Reckitt

Atif Ahmed

Director of Advanced Analytics