Case Study: Eventbrite achieves search-powered personalized event recommendations with Elastic

Eventbrite - Customer Case Study

Eventbrite faced a shift from an organizer-focused platform to a consumer marketplace for hundreds of thousands of new events every month. The core challenge was improving event discovery: search was good at matching queries but lacked personalization and context, while collaborative-filtering recommendations suffered from cold-starts for new events and were costly to compute at scale.

The solution was to serve recommendations from the search engine by combining content-based methods (Elasticsearch MoreLikeThis) with collaborative signals (matrix decomposition and count cooccurrence) and indexed affinity data. Techniques included boosting results by recommended event IDs, tagging events with related-events or users_who_might_like fields, and adding affinity tags so personalized search queries surface relevant events while respecting business rules like location and category. The result is a scalable, unified personalized-search experience that improves consumer discovery, handles new-event cold starts via content signals, and enables related events, newsletters, and broader discovery features.

Eventbrite

John Berryman

Aerospace Engineering