The resale retail sector, once considered a niche market fueled by thrifty shoppers and vintage enthusiasts, has emerged as a multibillion-dollar industry.
As just one data point, ThredUp, a quiet powerhouse in the space, has processed over 220 million secondhand clothing items to date. Behind the digital racks of its 4-million-item inventory lies a sophisticated data operation transforming how consumers shop for pre-owned fashion.
“We are a very data-driven operational company,” ThredUp Chief Product and Technology Officer Dan DeMeyere told PYMNTS.
Not all data, however, is created — or unlocked — equally. Making sense of vast amounts of disparate data has long been one of retail’s persistent challenges.
“One of the problems that we had … was that there’s so much data,” DeMeyere said. “It’s not always in the same easily accessible place. When you are trying to go after a problem for the customer or for the business, the data itself ended up being like one of the bigger hurdles.”
Fast forward to today, and the resale market’s meteoric rise is being borne upon a technological backbone with artificial intelligence and advanced analytics transforming how retailers operate.
AI and data analytics require structured, non-siloed data to meet their potential as a technological enabler capable of reshaping operations and consumer experiences.
ThredUp’s partnership with Databricks, initiated in 2017, marked a turning point for the company. Before the collaboration, ThredUp struggled with the common retail problem of data silos, DeMeyere said. Customer behaviors, inventory details and transactional data lived in separate systems, making it difficult to generate meaningful insights. The implementation of Databricks’ data lake architecture changed that dynamic, creating a unified platform that democratized data access across the organization.
This year, ThredUp has deployed AI in three ways, each aimed at enhancing the customer experience.
“We effectively overhauled our search engine to leverage generative AI so that people could start searching for things like ‘ugly Christmas sweater,’” DeMeyere said. “And even though in our database, there was no item that was tagged as ‘ugly Christmas sweater,’ we could have results.”
The company has also launched an image search feature.
“Take a picture of anything, and if there’s clothes in it, we will detect it, and we will find those clothes on our marketplace for you,” he said.
The technology extends beyond customer-facing features. Every night, AI algorithms analyze customer engagement patterns to personalize search results, while in distribution centers, AI assists with productivity and automation of previously manual tasks.
While some companies remain cautious about AI adoption, ThredUp has taken a more aggressive stance.
“Our trust level is very high, not 100%, but we do believe it is the future,” DeMeyere said. “If we can’t figure out how AI is going to change the shape of resale, someone else will, and they will disrupt us.”
Perhaps most intriguingly, ThredUp has ventured into what it calls “resale as a service,” partnering with brands like Madewell to power its resale operations. This white-label service uses the same AI and data infrastructure that powers ThredUp’s core business.
Looking ahead, DeMeyere said he envisions even more personalized shopping experiences.
“I think we’ll get to a place where your homepage will be different than everyone else’s,” he said, suggesting that AI could eventually generate custom product photography based on individual style preferences.
However, challenges remain. With 4 million items available, customer overwhelm is a real concern. A search for a basic item can yield 50,000 results, which DeMeyere said “is not helpful.” The company is working to use AI and data analytics to make the shopping experience more manageable and personalized, even for first-time visitors.
As 2025 approaches, ThredUp is focused on solving one of retail’s most persistent challenges: real-time personalization with minimal data. The company is developing systems that can create instant user profiles based on browsing behavior, potentially revolutionizing how online shopping adapts to individual preferences.
“We can create a graph [in] real time of when you click, where’d you come from, how long are you spending on each page, what are you clicking on each page, and in real time create this vector representation of who you are,” DeMeyere said. “Every time you click and the page refreshes… the personalization is real time getting smarter and smarter.”
The post ThredUp CTO Says Data Turns Secondhand Retail Into Top-Shelf Experience appeared first on PYMNTS.com.
