Case Study: Spring boosts product discovery and real-time search for 500,000+ products with Elastic

Spring - Customer Case Study

Spring is a fashion marketplace serving 2,000+ brands and 500,000+ products with a single checkout experience. The company faced search challenges across messy, multi-sourced product data: keeping Postgres and Elasticsearch in sync, surfacing accurate prices and sale information, resolving ambiguous queries (e.g., “blue dress”), and reducing noisy results so customers could find the right product quickly.

Spring built a Catalog Ingestion System feeding a central SPRING database with both pushed and pulled supplier data, then populated Elasticsearch via nightly repopulation and real-time ES syncs. They layered on a Sale Tagger and price computation, a Color Categorization API, Twiggle NLP for product ontology and query interpretation, an Attribute Tagging System (with NSQ) for richer metadata, and Consumer API-driven visual browse and personalization using user events and preferences. The result: more accurate prices and sale flags, better-filtered and visually similar results, and personalized search experiences that surface the products customers actually want.

Spring

Julie Qiu

Senior Software Engineer