Case Study: Surfline achieves faster engineering velocity and full-stack experimentation with Split

From Feature Flags to Full Stack Experimentation, Surfline Accelerates Velocity with Split

Surfline, the world’s largest surf forecasting site, faced growing engineering and product complexity as it scaled and replatformed from a monolith to microservices. Their twice-weekly deployments were all-or-nothing, making rollbacks time-consuming (hours to a day) and consuming the equivalent of half a full-time engineer’s time; long-lived feature branches couldn’t get real production feedback, and analytics lacked a scalable way to measure feature impact across releases.

Surfline adopted Split to decouple feature release from code deployment, enable phased rollouts and full‑stack experiments, and push Split impression data into their analytics stack via Segment and RedShift. Split rolled out in days, required minimal engineering effort to configure flags, restored engineering velocity and code stability, and let the team run experiments on registrations, merchandising, search UX and more — with plans to put every code change behind Split as they complete their migration.

Surfline

Julian Ganguli

Principal Data Analyst