High-Quality Labels Power Accurate Search for Walmart’s 385 Million Online Visitors

High-Quality Labels Power Accurate Search for Walmart’s 385 Million Online Visitors

Retailers invest millions of dollars every year in omnichannel ads designed to whet shoppers’ appetite for their products. To see a return on these significant investments, leading retailers complement their spend with a seamless search experience on their e-commerce websites.

When shopping online, consumers expect the search experience on an e-commerce site to be as reliable as a Google search. Retailers who fail to “Googlify” their online search are leaving money on the table: by some counts, 12% of users will bounce to a competitor’s site after an unsatisfactory search.

Walmart has always understood the importance of providing a good search experience online for their 385 million visitors. When they set out to improve site search relevance and account for their ever-growing catalog, they needed a labeling partner who could produce high-quality datasets for their algorithms.

Crucially, Walmart needed a labeling partner who could scale over time to account for SKUs being added every day without sacrificing the high standard of quality required to provide their customers with accurate search results.

Search relevance labeling challenges

For every SKU in an online store, there are a handful of associated attributes that a customer might search to find what they’re looking for. Labeling many dimensions per product is time-intensive and requires skilled annotators.

Walmart needed a labeling partner that could validate images and descriptions for their extensive product catalog, with feedback loops to ensure a high standard of quality.

Sama was engaged to classify items, validate descriptions and images, and provide suggestions for new product types and tags. Additionally, because new SKUs are added to their online store every week, they needed a partner who could account for a growing catalog. As Walmart explained:

“When an item can’t be classified by the current model, for example, when a completely new product is introduced like the fidget spinner, Walmart needs Sama to classify the items manually.”

Sama’s directly managed workforce of annotators remains in constant communication with the team at Walmart, to ensure that expectations and needs are being met. The team size can be scaled on-demand to accommodate growing datasets, with consistent and ongoing throughput via API deliveries.

Better customer experiences thanks to more accurate search relevance

Sama assembled a team of product attribute experts for Walmart. By 2019, Sama had covered more than 2.5 million items, improving Walmart’s retail item coverage from 91 to 98 percent.

Tight feedback loops resulted in high-quality labels that visibly enhanced their product descriptions, images, and categorization. As Vamsi Madabhushi, a Product Manager at Walmart Labs put it:

“Quality is important to them. You really get the sense they are there for more than just the financial transaction. They are a true partner.”

Today, Sama continues to deliver accurate annotations for Walmart’s growing catalog of products. If you’re ever purchased a product through their online store, then you’ve likely benefited from the shortened product discovery time that high-quality labels can bring.

Better quality labels lead to better customer experiences. And in the world of e-commerce, happy customers reward you with their loyalty — bringing better conversion rates, higher average order values, and ultimately, increased revenue.

Learn more about how Sama can help you improve search relevance and shorten product discovery time.

Related Resources

Orbisk is Using Accurate AI to Help Restaurants Reduce Food Waste Up to 70%

Swift Cost-Effectively Scaled Annotations Without Compromising Accuracy

Accurate Data Labeling Powers the Volumental Shoe Sizing App