Case Study: Fynd achieves real-time customer insights and millisecond-latency analytics with Rockset

Fynd Uses Kafka and Rockset to Respond to E-Commerce Consumer Behavior in Real Time

Fynd, an O2O fashion e-commerce portal serving 17 million customers and ingesting 20–30 million events per day from over 9,000 stores, needed to respond to consumer behavior, inventory and fraud issues in seconds during high‑traffic, limited‑time sales. Their existing analytics flow required a minimum 30 minutes to prepare and load data from Kafka, so Fynd turned to Rockset (a real‑time analytics platform that connects directly to Kafka) to meet strict real‑time requirements.

Rockset continuously syncs Fynd’s Kafka JSON event streams, auto‑schematizes and fully indexes the data for millisecond‑latency SQL, and Fynd built a serverless microservice (AWS Lambda + Rockset client libraries) to refresh dashboards multiple times a minute. The result: immediate, actionable insights on 20–30M daily events, over 40 real‑time metrics tracked, faster detection of anomalies and fraud, reduced infrastructure management, and eliminated the previous 30‑minute lag — all enabled by Rockset.

Fynd

Amboj Goyal

Principal Engineer