Case Study: 9GAG achieves 37% more post views and 22% higher interactions with Recombee

Real-time User Generated Content Personalization for a Global Entertainment Giant

9GAG, a global cross-platform entertainment network reaching 200+ million people and known for its infinite-scroll Home feed, needed to replace its in-house recommender with a solution that could serve personalized content under heavy traffic, deliver instant recommendations for newly added user-generated posts, and incrementally learn from continuous interactions. To meet these challenges, 9GAG selected Recombee and deployed Recombee’s Infinite Scroll personalization and real-time recommender engine.

Recombee implemented a custom ensemble of incrementally-trained models (collaborative filtering, content-based, deep learning, NLP and image processing, plus contextual bandits) optimized for multiple engagement KPIs and delivering recommendations in ~50ms, now powering 100% of 9GAG’s Home feeds. In a 50-day A/B test Recombee drove measurable gains: +37% post views, +22% overall interactions, +7.5% total session duration, +2.8% visits, plus increases in saves, shares and upvotes and improved user retention.

9GAG

Kristie Chen

Product Head