Ecommerce leader taps Sama to improve search relevance

Overview

A market leader in online retail delivery struggled to maintain the accuracy of its extensive product catalog, which included millions of items from multiple retailers. They relied on their foundational model to generate new product listings, including images and details, but it often produced errors. These inaccuracies, such as misleading descriptions and irrelevant search results, negatively impacted the customer experience. The frequent onboarding of new products and retailers also added to the challenge, causing fluctuations in volume that strained their processes.

‍

Scale, Quality, and Speed

The e-commerce leader needed a partner who could handle the scale, high quality requirements, and short turnaround times needed for their project. For these workflows, the Sama team:

• Added new products to the catalog, labeling multiple attributes per UPC, including brand, product name, size, unit cost, category, & images
• Validated product information, descriptions, and images, recommending new types, attributes, and tags to optimize search relevance
• Enriched data sets by adding detailed product information and standardizing size and unit specifications
• Proactively identified and flagged potential issues or inconsistencies in the product data

‍

The Results

Our team of experts helped the company achieve a 98% quality SLA, exceeding their 95% target. Sama significantly improved the performance of their Gen AI model and:

• Delivered more than 4 million products within a 12-month period
• Increased the accuracy and consistency of product descriptions, creating better customer experiences and fewer returns
• Improved product taxonomy and search relevance, enabling customers to discover products more efficiently
• Enriched product data to empower recommendation engines, driving deeper customer insights and delivering highly personalized shopping experiences

Better personalization strategies helped improve the average order value by 1% year-over-year. The company hopes to increase this to 4% in the coming year.