Target improves product search with hybrid lexical and vector system
Retail Product Search: A Practical Approach at Target
Information RetrievalMachine Learning
Summary
Searching for products online is tricky because people might know exactly what they want or they might just be exploring. Traditional keyword searches don't always understand natural language well, while newer vector-based methods can miss important details. The authors designed a system for Target that combines both keyword and vector search to get better results. This hybrid approach showed improvements in customer engagement and reduced cases where no products were found.
What this means in practice
- •For e-commerce platform teams: Enhance product search by combining keyword and vector methods to increase customer interactions and sales conversions.
- •For online retail operations: Deploy hybrid search systems to reduce zero-result queries and improve shopper experience at scale in production environments.
Authors
Darshan Sonagara, Qujiaheng Zhang, Ankit Singh, Alex Li
Abstract
Search is one of the most important features in e-commerce, directly driving customer engagement and business growth. A good product search system must show both relevant and desirable results. However, retail search presents unique challenges. User intent can range from exact matches to open-ended discovery. Search systems must also balance multiple goals, such as relevance, revenue, and profit, while keeping response times low. Traditional keyword-based methods often fall short in handling natural language or semantic queries. Vector search helps alleviate these issues, but it can miss key intent signals or return low-precision results. In this paper, we present the design of a hybrid search system at Target that combines lexical and vector search. We describe our approach to data processing, embedding training, precision control for the final result set, multi-channel result fusion (where we compared fusion strategies and adopted weighted interleaving), and the performance optimizations used to maintain low latency for production deployment. Our method improves offline evaluation metrics, and in online A/B testing it raised click-through rate by 0.97%, order conversion by 0.98%, and demand per visitor by 1.10% over lexical-only search, while roughly halving zero-result searches. The resulting system is deployed at scale and serves millions of guests daily.