Case Study: Komli Media achieves 50% lower processing costs and 30% faster job times with Qubole Data Service

Komli Media Improves Utilization with Premium Big Data Platform Qubole

Komli Media, a digital advertising platform built on Hadoop, was struggling to turn more than 100 TB of ad and behavioral data into timely insights. Collecting roughly 700 GB/day (≈21 TB/month), the company faced ad‑hoc, spiking workloads on fixed clusters that caused queries to take 10–15 hours, forced frequent denials of analyst requests, and drove monthly cluster costs to about $15,000.

Komli moved to Qubole Data Service for managed cloud processing with true auto-scaling and S3 I/O optimizations, which let clusters expand and contract automatically with demand. The switch cut job times by ~30 (bringing typical turnaround to 1–2 hours), halved monthly processing costs, scaled to meet all user requests (15 analysts/engineers), and delivered the flexibility and performance Komli needed to expand its analytics capabilities.

Komli Media

Shailesh Garg

Head of Analytics