Why ThredUp’s CEO thinks its investments in AI have been ‘underhyped’

When he looks back at the investments ThredUp has made over the last 10 years, he specifically thinks the company’s investments in AI have been “underhyped.” 

“I [have] said that ThredUp is going to uniquely benefit from these investments in AI technology because of the problems that we’re trying to solve, which is that we sell 35,000 brands across 100 categories,” he said. “It’s a hard, hard problem, and AI has been able to accelerate our ability to deliver a great experience to customers.” 

Generally speaking, investments in AI have allowed ThredUp to serve up more relevant search results to consumers, Reinhart said. In turn, that has improved conversion rate; Reinhart said during the Wells Fargo conference that the rate at which new customers that are visiting ThredUp are converting has improved 18%. 

“There are very few things you see in consumer where you generate an 18% improvement in overall conversion from visits [from] new customers,” Reinhart said.

ThredUp is just one example of how brands and retailers of all sizes will increasingly be looking to talk up their AI investments to Wall Street. Over the past year, companies have released an onslaught of new AI-powered tools, some of which they have built in-house and others that they have partnered with third-party tech vendors on. These include AI-powered shopping assistants that help shoppers compare prices and come up with gift ideas, as well as sizing tools that help customers find the right fit without having to try anything on. 

Now that many of these tools are coming up on the six-month or one-year mark, analysts are increasingly looking for brands to talk about the payoff.

Reinhart believes that AI has been a particularly powerful tool for ThredUp due to the unique inventory challenges it faces as a resale company. He compared each item ThredUp sells to a snowflake — because ThredUp is buying inventory from individuals, it doesn’t have, say, 1,000 pairs of the same black leggings to sell to customers. Instead, it may have a handful of black leggings from Nike, Lululemon or Alo Yoga at any given time, all in different conditions.

Jessica Ramirez, co-founder and managing director of the Consumer Collective, said that, historically, “search is a very difficult thing that has been a problem across industries.” What she has seen with the ascent of generative AI, in particular, she said, is that it is “as if the search capability is on steroids.” Consumers have to do less digging to find say, the exact silhoutte of a dress they are looking for. But, she said, that’s only if companies properly keep up with the tagging of items, especially as style trends evolve. She gave the example of a beauty company that may have to be mindful of correctly tagging products that adhere to the “glass skin” trend. 

At the conference, Reinhart highlighted five new AI products or features that ThredUp has rolled out over the past 12 months that have driven growth. 

One of the most important investments, he said, has been using AI to tag products based on different attributes. That, in turn, makes it easier for customers to search for items based on a greater variety of queries. 

“We went from tagging six or seven attributes about a given item to using AI that tags items with more than 100 different attributes,” Reinhart said. ‘So those 100 attributes then feed into a data model that makes it easy for customers to search and pivot what they’re looking for.” 

Other AI-powered features ThredUp has added to improve search include the ability to search for a product based on an image in a customer’s camera role, and similar search, which allows customers to see what a particular item looks like in different colors with just a click.

It’s something that Reinhart readily admits is table stakes on most e-commerce sites, but has historically been difficult for resale companies like ThredUp to pull off, given the uniqueness of its inventory. 

Reinhart also believes that AI can help realize the true potential of social commerce. “One of the things we think is broken in commerce is how people get inspiration,” he said, noting that it has historically been too difficult for people to gather inspiration from their TikTok, Pinterest or Instagram accounts all in one place, and then use that to inform searches on e-commerce sites. 

In April, ThredUp launched a beta version of Social Shop, which makes recommendations based on a person’s social media activity. Right now, ThredUp has an integration directly with Pinterest where “you can log in with your Pinterest account and then, every time you go and load the app, we in the background are searching through all your Pinterest boards to pull in the style you’re curating and making that shoppable on ThredUp,” Reinhart said.

 Soon, he added, shoppers will be able to link their Instagram and TikTok profiles.

Finally, the last AI investment Reinhart highlighted was Style Chat, which the company launched last August. It’s a chatbot that makes recommendations to customers about what to buy based on whatever attribute they are looking for, whether it is price, product or style. 

AI isn’t the only thing that has driven ThredUp’s growth this year. Reinhart also said the launch of a premium selling kit has improved average selling prices. And, he said, the company has had “small tailwinds with CACs coming down because of the closure of the de minimis loophole.”

But, what’s powerful about AI is that he believes it is only in its “second or third inning.” There’s more room for AI tools to get a lot better and, in turn, build a more powerful flywheel for ThredUp, he said. 

“Where I see the world going is that, for the customer, shopping secondhand online – the experience, the emotion, the quality – is indistinguishable from shopping new,” Reinhart said.