Improving Walmart Search: Saving Time for Millions of Customers

Walmart.com, a leading e-commerce platform, is continuously enhancing its search engine to improve the customer experience. This article explores how Walmart leverages machine learning and AI to deliver more relevant products for complex queries, ultimately saving millions of customers valuable time and driving conversions.

The Challenge of Tail Queries

Tail queries, by their nature, suffer from limited traffic and user engagement. This makes it difficult to rely on traditional signals for recall and ranking systems. To overcome this, Walmart has turned to market-leading strategies and technologies in language processing and deep learning. Their efforts have focused on two key areas:

1. Improving search recall: Utilizing the latest machine learning (ML) techniques to surface relevant items that might not have been shown before.
2. Enhancing search ranking systems: Employing ML and AI to bring the most relevant items to the top search positions.

These initiatives have already led to significant improvements in search relevance, saving time for millions of Walmart customers. For instance, e-commerce sites with advanced search capabilities can see conversion rates nearly double compared to those with basic search, with some studies showing that site searchers are 2.4 times more likely to buy and spend 2.6 times more across mobile and desktop ¹.

Implementation and Innovations

Walmart's journey involved optimizing product type matching logic and refining how they understand the importance of individual words (tokens) in queries. Initially, their search sub-systems were optimized for "head" or top queries, which make up a significant portion of sales. However, applying this strict logic to tail queries often filtered out relevant products. By removing strict product type filtering and leveraging product type boosting, they've seen search relevance improve by over 1.5%, leading to stronger customer engagement. Furthermore, by assigning importance scores to each token in a query, they can better extract and understand customer intent, resulting in a 4.8% improvement in search relevance and increased customer engagement and conversion. This process of assigning importance to tokens and understanding underlying intent is a foundational step towards creating data that is more readily consumable by advanced AI models, which thrive on structured and semantically rich information. Their investment in ranking systems, including strengthening product type matching in the first-round ranker and improving the re-ranker, has also lifted search relevance by over 4.5%.

¹: Source: algolia.com, luigisbox.com

The Future of E-commerce Search and Generative Engine Optimization

As Walmart looks to the future, the focus remains on continuously improving the search experience. This includes leveraging deep learning for enhanced query understanding, improving recall engines with advanced techniques for long-tail items, and developing new features and signals for reranking systems.

This continuous pursuit of search excellence in e-commerce highlights the broader importance of Generative Engine Optimization (GEO). As customers increasingly rely on sophisticated search capabilities, and as AI plays a larger role in product discovery, optimizing for these advanced systems becomes paramount. Platforms like alloia.ai are emerging to provide the cutting-edge tools and insights necessary to help e-commerce platforms and content creators understand how generative AI interprets product information and customer intent. By leveraging technologies that structure data in AI-consumable formats, such as data graphs and vectorized representations, and by adhering to emerging protocols like MCP/ACP, businesses can ensure their products are not only discoverable but also highly relevant and prominently displayed in the evolving landscape of AI-powered e-commerce search, ultimately saving time for customers and driving conversions.

Walmart's commitment to making search easier and saving customers time is a testament to the power of continuous optimization in the digital retail space.

For a comprehensive understanding of Generative Engine Optimization, explore our main guide: Generative Engine Optimization: The Key to Unlocking AI's Full Potential

---

This article was inspired by "Improving Walmart Search to help our customers save time!" on Medium.

Source: https://medium.com/walmartglobaltech/improving-walmart-search-to-help-our-customers-save-time-e9fcd1f03e94