Wayfair achieves 99% category win rate and 7-point clickthrough lift

Wayfair, with 40 million products and 22 million customers, had inconsistent supplier data hurting search, and manual labeling was far too slow (eight weeks for just 6,000 images), while generic models like CLIP failed to capture design nuances in 95% of cases. Working with Snorkel, Wayfair built a programmatic data workflow using weak supervision, producing 46 tag models within days and using model-guided error analysis and a custom foundation model for visual tag extraction. The result was a 98.97% win rate over prior category models, a 7-point clickthrough lift, a 5-point add-to-cart lift, and 10x faster model development, saving an estimated six years of employee effort.

Snorkel AI is a data-development and evaluation company whose programmatic labeling tools build training and benchmark datasets used to stress-test whether a model approach is viable before a full build.

This sits in the Design & Stress-Test stage of the AI use-case lifecycle: Named examples from the tools pharma teams use to scope, prototype and benchmark an AI use case before committing to build it — data-catalog, sandbox and evaluation platforms.