CSA Exclusive: Resale retail giant ThredUp gains insight into unique items with AI

A popular omnichannel resale retailer is identifying “one-of-a-kind” product attributes – at scale.

ThredUp, a digitally native apparel resale retailer that operates an online marketplace as well as in-store shops at Madewell, J.C. Penney, and Macy’s stores, has a constantly growing inventory of more than 3 million items. Ryan Moser, director of engineering, recently spoke with Chain Store Age about how the retailer is leveraging the Lily artificial intelligence (AI) platform to efficiently tag each product with multiple attributes.

“Providing our customers with more windows into our inventory is useful to ensure that they can find exactly what they are looking for,” said Moser. “Having previously used a combination of manual tagging and technology, we chose Lily as a partner since they can help provide more tags very efficiently, in a way that ultimately enhanced the customer experience in the pilot program.”

The Lily platform uses AI and computer vision technology to define all of the elements of a product consistently, for use in customer-facing and operational retail management systems. Starting with recommendations and customer experience improvements, Lily applies image recognition and machine learning to apply deep attributes and use them to better understand why shoppers purchase specific items. 

“We send Lily relevant inventory that could benefit from tagging, and they are able to tag items in a matter of seconds,” said Moser. He went on to describe how ThredUp can offer an enhanced customer experience as a result of improved inventory tagging.

“With millions of products on ThredUp, the Lily AI solution provides our shoppers with more ways to surface the exact items they want, further reducing the rack-digging typically associated with thrifting,” explained Moser.

Unlike traditional retail, every garment that comes to ThredUp is a one-of-a-kind item with unknown metadata and value. ThredUp assigns attributes to each garment, such as cut and color, which makes it searchable online and factors into its resale value.

Lily is now helping ThredUp efficiently tag each article of clothing with more attributes than ever before, according to Moser. The platform’s filters and facets identify specific accents, patterns and materials, and account for factors such as style and occasion, and fit details such as shoulder cut and sleeve length. Because ThredUp carries 35,000 brands across 100 categories, a customizable search experience is critical. Lily AI helps ThredUp customers discover the items they want faster and more efficiently.

Moser says Lily is helping ThredUp provide customers with more ways to discover exactly what they are looking for online through search and filtering.

“These allow customers to narrow their search to exactly what they are looking for,” said Moser.