Case Study: Audiomack achieves 206% increase in monthly plays and 67% increase in weekly follows with Recombee

One-On-One Content Recommendations for the Global Music Sharing and Discovery Platform

Audiomack, the global music sharing and discovery platform used by millions, needed highly personalized, geolocation-aware recommendations at scale — with near real-time updates, sub-100ms delivery, robust cold-start onboarding, and the ability to leverage hierarchical relationships between songs and artists across 100+ cultural regions. To meet these requirements they partnered with Recombee and adopted Recombee’s recommender engine to replace and augment in-house recommendation capabilities.

Recombee built a custom ensemble (collaborative filtering, content-based, contextual bandits) with real-time data ingestion, Item Segmentation, localized recommendations, and an advanced cold-start onboarding flow, delivering recommendations in ~50ms on average. The solution drove a 206% increase in monthly plays from recommendations, a 67% increase in weekly account follows, made Recombee-powered rows the best-performing module in the Discover tab (46% of plays), and Item Segmentation alone accounted for ~10% of total follows with strong follow-on behavior.

Audiomack

Christopher Dalla Riva

Senior Product Manager