ThredUp Inc. (TDUP) Earnings Call Transcript & Summary

Good morning, everyone, and thank you so much for joining us. My name is Nathan Feather. And I'm Morgan Stanley's small and mid-cap Internet analyst. I'm excited to be joined today by James Reinhart, Founder and CEO of ThredUp; and Sean Sobers, ThredUp's CFO. Thank you both so much for joining us today.

Thanks for having us.

Thank you. Good to be here.

Now before we begin, a quick housekeeping item. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

And with that, let's start it off. So it may be helpful to start off with an intro, James. For those that may be less familiar with the company, can you walk through a quick overview of your strategy and the competitive differentiation?

Sure. ThredUp's a managed marketplace for women's and kids' clothing. So we help women, primarily, we don't do men's, clean up their closets, get rid of all the clothes that they're no longer wearing in a convenient service. And then we put all those items online through our distribution network to allow women to shop this incredible assortment of secondhand clothing. And I think that uniquely differentiated across, say, other marketplaces is that we manage the entire experience. And so rather than sort of an eBay experience where you would list your own things and sell them and fulfill them and do customer service, we take all of that in-house. And so we run a network of distribution centers. We have enormous amounts of data and processing power to put all those clothes online. And we think, ultimately, what that does is expand the market opportunity. And so we think the TAM is much larger for a managed marketplace, although they take longer to build, longer to scale, they require more CapEx, but we ultimately think incredibly defensible over time, and that's the journey that we've been on.

Okay. Great. Well, let's dig a little bit more on the TAM, and you put out a lot of work on this. So can you help investors size the market opportunity for ThredUp across U.S., Europe and then globally?

Sure. I mean the resale market is the fastest-growing part of the broader retail apparel environment. And depending on the report that you read, it's growing anywhere from 10 to 15x faster than traditional retail. And it's really being driven by young people. And I think the way that the market sort of studies work is they talk to people about how are they selling and are they selling in the market? How are they participating? And what are they going to do in the future? So a lot of the growth rates make -- that you kind of see in the future is really consumer surveys around how they're going to participate in the future because what you tend to see in these types of markets is once a consumer participates in secondhand, they become pretty sticky. And I often talk about like the woman who first gets the J.Crew dress for $24, and it's amazing. It's kind of hard to go back and buy $178 J.Crew dress again. And so there's real stickiness on the engagement side. And so the market is growing in the U.S. I mean there's been some volatility, obviously, with the pandemic and interest rates going up. But if you take kind of the long arc, we think this is a huge market over time in the U.S. And then in Europe, it's a similar size. And so we think it's a massive TAM over there. As some of you might know, we have a business that operates in 7 countries in Europe. And we think both of those are great vectors for growth.

Okay. Great. Well, you haven't navigated, I think, a challenging near-term macro environment. So it'd be interesting to hear what your current read of the consumer is. How that's trended over the past few months? And then are you still seeing pressure in especially some of the value buyers that you have in the platform?

Now look, I think that the consumer is just living through the compounded effects of inflation. I think all of the news media talks about how inflation is coming down. But like you can't forget that since 2019, it's elevated quite a bit. And so I think the consumer is dealing with the lingering effects of that. If you look at credit card balances and credit card default rates, we can debate how big an indicator they are, but like I think it speaks to an indication that there's a little bit of stress on the consumer. You look at sort of rent costs, groceries, these things are just -- they hurt the pocketbook. And even if you're sort of a wealthy consumer and you're here at the Morgan Stanley like tech conference, like you can't ignore the grocery bill. You're like, "Man, groceries are more expensive." And like dining out, that's more expensive. And airline tickets, they're more -- and so all that stuff, I think, puts pressure on discretionary categories. And apparel is a discretionary category, by definition. None of us need a new pair of pants tomorrow, right? But we do need to get to eat. We do need to get to work. You need to pay for our kids' things in school. And so I think that's where we're living in. And I think consumer confidence, while improving, is still not in a position where I think people are feeling particularly like they want to lean in on the discretionary side. What I would say is that I think that, that is -- the trend line is more positive than it was over the last couple of years. And I think the catalyzing events are going to be things like interest rates starting to come down because I think there's a psychological effect there that's really, really powerful. So -- but until that happens, I think consumers will feel like they're getting better, but I don't -- but I think their purchasing power for discretionary is going to be squeezed.

Okay. Well, any impact on some of those value buyers? Or is that cohort behaving differently than some of the more like tried and true retail buyers?

Yes. I mean, I definitely think it's hurting the value buyers more because I think they just have less discretionary dollars to spend. And I think the Q4 holiday season is interesting because I actually think people did spend a lot, but everything was on sale and deeply discounted. And so you might have had some pull forward of demand, which is normally typical, but I think things were really discounted in the back half of last year, specifically in apparel.

Okay. Great. Well, and I want to dig in a little bit about that later. But first, I want to talk to you about one of the strategies that you put in place to try to address these concerns, which is shifting upmarket. And so first off, what's that strategy behind that shift? Is it primarily macro? Or what are the other kind of impacts in the marketplace? And then can you help people understand, I think people would think upmarket, they often go immediately to lunch or -- and that's how we were talking about here. So -- and what do you mean by -- what is now the kind of bread-and-butter item on the platform and ASPs you're looking for?

Yes. I mean I think the shift has been subtle, but it has been upward. And the distinction would be that I don't think there's a lot of room to compete with Old Navy, right, in a resale context. I think it's hard to compete with a private label brand at Target, right? It's -- and so I think what we've tried to do is actually move up to a brand assortment where there is actually some opportunity. And that's brands like a J.Crew. It's brands like a Ralph Lauren. It's a brand like Banana Republic who's trying to move, right? And so there's like a lot of those brands that I think we're trying to -- we sort of inched up. And I think the lag is that you have to both -- in a managed marketplace, you have to acquire both supply and demand. And if you acquire all the demand without the supply, then you have a lot of unhappy buyers. You acquire all the supply without the demand, you unhappy happy sellers. And so we've had to methodically move that up. But I think now we're at a point where average prices on ThredUp are in the mid-20s. 5, 6 years ago, they were in the high teens. And so we've definitely made that shift, and I think it's been super positive for our business. And -- yes.

I think -- and Nathan just add to that too is I think that people sometimes misunderstand how we take 4.5 million items, which we currently have listed today, and transition the ASPs from, let's say, $19 to $25. But I think people don't understand the magnitude of our operations. We're processing on average 100,000 items a day. So it's not that long for us to go through the full cycle to be able to make that shift in transition on ASPs.

Okay. Well, it's great. And I want to talk about something maybe a little bit more exciting. It's in the comments you were talking about, James, in the earnings call in gen AI. And so interested to hear specifically on the buyer experience, what are some of the key projects you're working -- are working on? How far are they along in the development process? And generally, when should we expect to start to see some of the impacts from those projects?

Yes. It's super exciting for us. I think we all know like AI has gotten a lot of hype. I actually think it's underhyped, right? I think it's transformative. I was a History major in college, like it's like the printing press, right, like -- or the invention of the Internet. I think it fundamentally will change the way the world works in so many profound ways. And so as the founder, I have spent a lot of time over the last 6 to 12 months, like rethinking how it will impact the customer experience. And so what we've done today, we fully launched it last week, we talked about it on the earnings call earlier this week, is launch a fundamentally new AI search product. And you might say, okay, that doesn't sound super interesting. But like in secondhand, where you have 4 million items and you're adding 100,000 a day, one of the biggest things we hear from customers is, "Oh, I can't find something I want." And this is a dramatic -- I mean it's not a step function. It's like it's 2 step functions improvement in how it works. And because what AI is able to do from us now is catalog -- we used to be able to tag maybe 4 to 6 attributes that a customer might search by and we would have to use human intuition to sort of do that, right? And we would sometimes use some computer vision, but it was kind of a dumb -- it was a dumb way to do it, which is a computer before this could see stripes, they could see color, they could see some pattern, but it couldn't pull out Golden State Warriors. It couldn't pull out that that's a four-leaf clover for St. Patrick's Day. And those are 2 like of the most simple ones. The AI approach now is tagging hundreds of attributes and then it's clustering those attributes in ways that it can intuit what the user is actually looking for. And so already, we're seeing consumers do more searches, find more product. And right now, what we could do is educate. What's funny about this is the hardest problem has been, we have all these customers who are used -- I'm going to be brutally honest, are used to the ThredUp search kind of sucking, right? And so they figured out like they know how to use filters, they don't necessarily trust the search results. And so there's this massive education of like, "Oh, no, no, no. The new search is amazing. Let us show you how it works." And the users that like are getting on that journey are converting at higher rates, they have higher add to cart, they get to -- like you can clearly see it. And the way we see that is for new customers. So new customers who don't have any legacy baggage around what search used to be. New customers are using it in ways that are like quite amazing. So that's live. And then we're working on a number of other things across outfitting and merchandising that I think will also be really powerful in the front end. I would say -- to even add a little bit deeper because I think the concept of finding the Golden State Warriors is almost logical, we can understand that, and it's capturing the image and all this. But a simple thing we did with an investor yesterday, they said, "Search Stevie Nicks." And I was like, "I don't know if we're going to get Stevie Nicks concert shirts or whatever we going to get?" But it actually was Stevie Nicks' style. So a bunch of boho dresses. And so it's so much more intelligent than we find the word Stevie Nicks, we find who Stevie Nick's style and how that works. And so it's really interesting how it works. And you can then take -- we'll keep -- just one more thing on this. So like you can actually then take what's become trending and trending, whether it's across social or Instagram or Vogue, like pick whatever sort of influencing sort of style, you can sort of take what's happening on the Internet and magically inject that into what's shoppable on ThredUp and give the customer something that's highly relevant and shoppable. Like there's a trend right now around sort of urban Black cowboy like it's like a theme, like you can now take that from a merchandising perspective and deliver customers a search experience, where like when you type that in and create that -- you're like blown away with the results. You'd get cowboy shirts. You'd get boots with -- you'd get cowboy boots. You'd get denim that would fit a cowboy. You'd get a whole bunch of cowboy hats. You'd get some bolos. Like it's kind of -- I mean it's magical in a way that I think customers are going to love.

We have cowboy hats?

Well, it turns out. Well, this is a great example. You could never find cowboy hats because the fine cowboy hats, you have to like find the filter, and you put cowboy in and Cowboy is a brand. And so there's a lot of like technology behind the scenes that's changing. Yes, hats. It's one.

It's really exciting. So I also want to focus not just on the front end, but also on the back end of how you're thinking about leveraging AI within your operations. So I mean the efficiency of your DC network, it's long been a real competitive advantage for you all. How do you believe that gen AI can help deliver that next stage of efficiencies?

Yes. So we're processing 100,000 items a day in the DC. We're taking a couple of hundred thousand photos a day with highly automated but still manual, like process, like there is somebody who is hitting a button that makes a camera work. I think in the future, you're going to be able to radically alter sort of the flow of inbound process to be able to do 2 things: one, essentially take any item that -- in any -- so if we had an item up on stage, Nathan and I just pulled it up, and I just turned it around and there were cameras hitting it, I could immediately stitch that on to a human. And not only that, but I could stitch that on to a human and it could be, if you're an African American who's shopping, I could stitch it on to an African American woman so you could see. If it was a plus size item, we could stitch in onto somebody whose plus sized. If it -- and then we could take that item and say, "Hey, show me that in the context for going out on a date in New York City," and we could generate that. And then we could say that -- the same process could say, "Hey, and show me some boots that I can wear with that dress," and we could do that. And so all of that stuff done, right, with the computing power of AI. And so, so many things today that we have to do manually, I think -- and we've gotten better, but this is -- again, it's a step function or 2-step function changes and how we can do that, which lowers our cost and I think radically improves the customer experience.

What I think if you tie in what James is saying and get precise on maybe what the opportunities are, we process 20-plus million items in the U.S. So if we're able to use AI in a way where we're not putting items on a mannequin and then off a mannequin, which we physically do today, you're going to save some amount. But assume you save $1 and everything else stays the same, so you're going to drop $20 million to $22 million to the bottom line. So it's pretty amazing and it's super impactful.

And then everything scales because you're not building more of those stations to do more of those things, some more of your space can be used for inbound and outbound processing. And so -- so yes, we believe it's fundamental transition. We -- or the smartest people at the company are spending 110% of their time on it.

Okay. Great. So it's all really exciting. I want to flip over to another real growth area of the business, which has been Resale-as-a-Service or RaaS. So I mean, '22 -- so this really dramatic increase in adoption, I think it was up 3.4x or something like that. And there was still growth in '23, although a little bit more of a digesting period. And so given that you now have a lot of brands that have had RaaS partnerships for a year or 2 years. How are their wants and needs evolving over that timeframe? And how are you meeting those needs? And then in terms of scaling RaaS further, how should we think about the '24-'25 roadmap?

Yes. I mean I think you -- it kind of maps sort of your Gartner Hype Cycle, right? I think there was a lot of anticipation and hype around how brands could leverage resale. And I think a lot of brands launched. And it turns out products don't distribute themselves. And so I think now we're in the phase of, okay, brands understand how the customer is responding. They understand -- has a platform like ThredUp can power or there are some startups there doing this as well. And so now like, okay, well, how do I continue to evolve that customer experience? And so I think what you're going to see this year is more brands sort of tweaking how their model works. And so I'll give you just an example. Like there are a number of brands who said, "We love the idea of branded resale. We don't want to do anything like really complicated. We want to just create a peer-to-peer market, like we want to connect buyers and sellers in this retail channel." It all sounds really good. But we all know, like at the end of the day, you need to drive liquidity in that market. And so like there's a huge handbag brand that has a peer-to-peer site, I think it's a $4 billion company. It's got 70 handbags on its peer-to-peer platform. It's hard to imagine anybody at that brand cares at all about this, right? And so what I think you're going to see in '24 and into '25 is some real -- a reality check around what's actually like working and what's not working. And I think ThredUp is well positioned to help these brands scale this, deliver for the customer. But I still believe over the long term, it will be an important part of what we do, and it will be an important part of what brands do.

Right. Well, thinking about -- another area that's been growing quite nicely in Remix, really strong since the acquisition. I think you have grown the U.S. business, it's been improving on the profitability...

Yes. Doubled over the last couple of years, yes.

Yes. So I guess what have been really the key improvements that enabled that growth? And then how much low-hanging fruit do you see remaining to really just leverage the ThredUp playbook at Remix?

Yes. I think that -- I think in some -- in many, many businesses, you go through these cycles of sort of growth, optimize growth, optimize growth, optimize growth, optimize. I think we've been in a growth base for the last couple of years. And I think we're much more right now in a let's optimize. So the business is double, right? There's a new distribution center there. They've learned a bunch of new processes. There's new talent in the team. So I think what -- this year is really an optimization year and primarily around that mix of goods. And so consignments -- we've been very public around the consignment shift in Europe. I think it's critical to the success of that business. The way it's been critical to the success of the business in the U.S. And so we're really focused on that new competency. And it's going -- kind of going as planned. We sort of built all the technology. The difference in Europe relative to the U.S. is you're dealing with 9 countries, and you're dealing with different supply chains and aggregation. And so we built most of that technology in the middle of last year, and then we started to roll it out in countries like over the course of the end of '23. And then we launched a few more countries in January and February. And I think we'll be out to almost all of our current markets on the sourcing side by April or May. And so -- but remember, Nathan, like it takes -- you have to launch into a country, you need to get people to order bags, then they need to fill those bags, they need to come back to the DC, they need to be processed and then they'd be sold for it to kind of hit the P&L. And so sometimes, it's -- it can be a 60-, 90-, 120-day lag from when you launch in a country from when you actually start to recognize revenue. And so a lot of the consignment revenue transition is backloaded just because the time is a real thing, right? It's not an instantaneous switch. And I think we do have real opportunity to take some of the things we've done in the U.S. and bring it to Europe. I think when we first were talking to Remix pre-acquisition, they brought us down to their room that was doing pricing. It was like a very dark room and a bunch of people in there. I was like, who are these people? It was like the equivalent of a basement. It was very dark and interesting.

With a bunch of spreadsheets.

Yes. It was very wild. So -- but being able to take them further along on our journey in the U.S. and how we do pricing and make it sense in the style of is it different per size? Is it different per season? Is it different per item? And the answer is yes. How do they be -- are they able to implement that and make a better environment for their consumer, but also for the business itself, you're able to sell things more appropriately if you're able to get the right items in.

Okay. Awesome. Well, do you want to talk about the full year '24 EBITDA margin guide? So exciting to see that positive at the midpoint, I think, 1%. Sean, what are the key building blocks that get you to breakeven and then importantly past?

Yes. So I think it's kind of like this year and then beyond this year is the way I think of it. And I think it's fairly simple in a sense that it's as simple as revenue growth is a big driver to this. and it's going to drive EBITDA as we have had it in 2023, but that's also going to expand based on our guidance. And I think if you look at -- we don't share the unit economics in detail, but if you look at the U.S. economics by the unit, they're great. Like we need to just do more selling of what's happening in the U.S. And sometimes people get hung up on Europe is going through this change and they need some improvements happen. But the U.S. -- it's like a reminder, the U.S. is 80% of the business. And so the more we do there, life is better across the board. And getting to EBITDA breakeven, I'm not saying everything is easy, but this is -- we feel very comfortable with this. Then I think we get leverage on marketing. We get leverage on operations, product and technology, SG&A. But I think it's like a very simple walk from where we are as we close '23 to breakeven in '24. And then I think you're headed towards like where do we go from here? I think our long-term target model hasn't changed since the IPO, and that includes acquiring a business that had different economics, but the goal is still the same. So 20%, 25% EBITDA, we want gross margins in the 75% to 78% revenue. If you look at that particular in the U.S. business, we're already surpassing that, right? We've done greater than 78% a couple of times. And so now it's bringing up the European business to get there. And part of it is what we've already talked about, the consignment transition, and part of it is bringing kind of what we know and how we do things in the U.S. to the European business. And we talked about pricing, but there's just all kinds of things like data science and intelligence that they just haven't used before, and we know how to do that on the U.S. side and bring it to that business.

And I would just add, like, Nathan, like the other thing is I think if you look at the -- over the last 2 years, we've had sequential improvement in EBITDA every quarter -- 7 of the 8 quarters. In 1 quarter, it was just a little flat, right? And it was just a -- sort of a timing issue. But 8 quarters of quarter-over-quarter EBITDA expansion, right? Similar story for Q1. And on -- if you look at a yearly basis, '23 over '22, it's almost 1,000 bps a month. So business grew double digits with 1,000 bps of margin expansion. And the guide for this year, if you think about the consignment transition, like the guide is essentially another double-digit growth with like 600 bps of margin expansion. So like where we sit, we feel like we're running the business in a very effective way around continuing to grow top line, continuing to show leverage on the bottom line. And so we expect that to continue. We have extreme confidence in the full year EBITDA number. And then free cash flow kind of comes from there, right? Because we have very minimal CapEx. We don't expect any CapEx spend beyond sort of maintenance until at least 2026. So we're 2-plus years of living in a world with minimal CapEx and expanding EBITDA. And so I think we feel good about that.

I mean you add in the transition, too, as we migrate completely away from owned goods in the U.S. and we make that transition in Europe, there's less of a need to buy inventory, obviously, as well. So that's a little bit of a tailwind on the cash side also. So I think there's a lot of tailwind as it relates to breakeven EBITDA that's going to just start generating cash and then all the things on the balance sheet that will help us from an inventory side.

James, you noted minimal CapEx over the next 2 years. So I want to refresh what has been the core of the CapEx, which has been your DC base. And so with Dallas partially built out, what's your current capacity over the entire network utilization? And then given that, what are the key priorities for the DC base in terms of improving efficiencies over those kind of next 2 years, given you're not spending that much on CapEx?

Yes. We have about 9 million slots, right, 9 million capacity. We think -- we believe we have -- we can essentially double the business inside of that 9 million without requiring any incremental CapEx. So just, right, in the U.S. So lots of room to kind of grow into that, primarily because we become more efficient on turns. So we have less items that are hanging in those slots because we're turning them faster, right, which I think speaks to all the sculpting and pricing and work that we've done. So we feel like really good about the network build-out. And that was very expensive over the last couple of years, but it's kind of behind us. And so I think it's important for people to feel good about that. And then on the optimization side, I think we will always -- we -- our team thinks about the operations process in seconds. So it's like how many seconds does it take to move a good through inbound and outbound. And we know every second is worth x. And we have plans to take lots of seconds out of the process this year, excluding the conversation we had about AI, right? I think the AI piece isn't in an R&D track that -- we're investing real resources to unlock that because we think it's profound. But you'll see it's kind of progress from us on the way we've expanded it over the last few years, and then I think there's some big unlocks.

Okay. Well, it's also been a few quarters since you introduced the bag fee for sellers. And there's been different ways you try to tweak monetization on the seller side. So I guess, talk through what that per bag fee has led to in terms of inventory quality, unit economics. And then going forward, what are the levers you can pull to either ramp faster or slower in terms of inventory, especially given the excess capacity you have within the DC network?

Yes, it's a great question. It's funny as a founder, you do something like this, you're like, s***, we should have done that earlier, right? Because it's not just the monetization of the bag fee, right, which is helpful, it's that it fundamentally changes how sellers act in the market. And so when you have to pay a fee, people tend to put more items in the bag. They tend to make sure the items are higher quality. They tend to read the instruction on it, I think that we say, around like what we want. Like it's such a -- it's so obvious, right? But I think we had such resistance to keeping it free. And so the effect of that has been we get more items per bag, that we leverage over our logistics, we get greater acceptance, greater quality, more in-season inventory. So it has been successful on every metric that we have. And the lever is we can mess with the fee. So we have thought about it as something where if our logistics costs on inbound bags go up, can we have the fee float with that? So it's a very effective hedge around increasing inbound logistics. If labor increases, we can flex the fee. So it's a natural hedge in the supply chain that we think is really important. And then for our best suppliers who we want to maintain total control of their selling, we can waive the fee and people love that. It's like I remember when 2-day shipping was just starting to become a thing that the brands would do. I remember on Bonobos, like back in the day, you would order something and then a day later, they're like, "We gave you 2-day shipping." And it would come like the next day, and it was a surprise and delight moment. And so we're using fees as a surprise and delight moment in a powerful way.

Okay. Great. Now you've also touched a couple of times the shift from product to consignment, particularly within the Europe business. So firstly, I'll just confirm, U.S. business, effectively fully consignment at this point?

So think of it this way. We ended '23 there, probably about 80% consignment. And we expect them to be -- or them, us, the U.S., however you want to look at this, it's about 95% consignment. And we...

By the end of the year.

Yes. '24.

Okay. Great. And then on the Remix side, how much is still within the product business? And then how should we think about both the gross margin benefits of switching over to consign, but also critically, the cash benefits of not having to acquire the inventory and hold it?

Yes. So the European business, essentially for '23, was less than 1%, so call it 0. And we'll do -- we'll transition to about 20% for all of '24. So the exit run rate will be higher than that. But it -- for the full year, it will be about 20%. From an impact -- I mean there's 2 types of impact, I think, from the shift to consignment. There's just the traditional accounting shift, which means you -- the payouts associated to the item move from COGS, up to contra revenue. So inherently, you just get a better looking contribution -- or a better looking gross margin at that point, but your gross profit dollars don't change at all. And that is the big reason why we push people to go think about -- or we think about the business on gross profit dollar growth. That's a true measure of what the growth of the business is in this transition period. I think the other side of it is, is you basically eliminate inventory risk. So you don't buy something that you're not going to be able to sell. You aren't able to change the purchase price of the item because now you're actually buying it with the selling price known. So I think there's a big piece there. And if you look at inventory on the balance sheet, I think something like, I don't know, 75% of it is the European inventory. Very little piece is the U.S and the U.S. is going to continue to shrink. And now Europe is going to start to shrink. So thinking how that's going to work on net working capital, those are going to both be beneficial as we go through '24. And it will continue in '25 on the European side, but it's just going to continue to get smaller.

Okay. Great. And outside of that product, the consignment shift, what other levers do you have to pull on gross margin in '24 and '25? And how should we think about that ramp?

I mean I think there's the traditional stuff we've been talking about, what I call is like our normal automation and process that we've been working on. And I think there's these huge step functions of opportunity, I would say, on gen AI. I mean it's -- from my point of view -- I called our Head of Ops, I was like, when are we going to start using AI in inbound ramp up process? He's like, we've been using AI since you've been here. So -- but that's -- it's like their thought process has always been far ahead there, but it's also this step function opportunity is huge. I don't know if there's any more specifics we can get into yet until those kind of start to take provision.

Stay tuned.

Okay. Great. Now I want to touch on one of the hot topics for investors lately, which is the potential impact of some Chinese-based exporters. And so first off, are you seeing any impact there on either the demand side or the marketing side, particularly performance?

Definitely, definitely. I think it's been one of the things that has slowly accumulated, right, in its impact. But I think we would be -- we would have our head in the sand, right, if we didn't acknowledge that there's some impact because just the scale of which Temu and SHEIN, right, are spending dollars in the U.S. is substantial, right? I mean they ran all those Superbowl ads, so that they could teach us all to say Temu and not timu, right? And so for sure, it's having an impact. And I think like it will continue to have an impact. I think they benefit from some loopholes, right? We can talk about the de minimis tax rule. There's definitely -- I know Congress is looking into that. Just yesterday, actually, there was a bill that was introduced in the Senate about this exact issue called the Americas Act, which we were actually a part of the spearheading coalition of that. So we are a founding member of the American Circular Textiles Association, ACT. And ACT was the driving force behind this bill. And it has incredible meat in it. It's the first bill ever introduced in Congress that creates conditions to benefit resale, recycling, rental. I encourage you guys to look it up. It's really exciting because I think there's an acknowledgment that U.S. companies are not able to play by the same rules and that we all understand that we -- the government needs to be involved in sustainability. The same way government was involved in regulating cigarettes. The same way government was involved in inspiring solar adoption and curbing emissions in cars. I mean there's a Tesla truck parked out there, right? And so I do think -- I really believe that government is going to play an important role in making sure that the U.S. and resale can be more competitive in this environment. And so super exciting.

Okay. Great. Now let's just talk a little bit about OP&T. I guess how much of that is...

It's a nice transition.

Yes. Smooth segue. How much of that should we think about as fixed versus variable, given there's a lot within that bucket? And within the variable expenses, we've talked to a couple of areas to try to improve that, but how much do you think you can bring that down over the next few years?

I think it's probably 70%, 75% variable. I think what people probably don't understand is as we went from 2019 to where we are today, labor rates and logistic rates have gone through the roof. Yet we've been able to reduce our cost of produce -- or to move an item through the DC. That's impactful in a way that really has improved contribution margin. So what that means is we've really reduced a lot more than that because we covered up those big increases in logistics and labor rates. And we're able to continue to do that literally -- it's not forever, but I have a great confidence in the operations team that we're going to continue to make strides. And I know the things we're working on before we even get to next-gen AI, there are opportunities there that are continuing to improve that side of the business.

Well, what's was the number? I think -- was it -- we shipped 500,000 more orders? 500,000 more orders?

500,000 more orders.

500,000 more orders this -- or '23 over '22, same OPT, right? And so I think it speaks really to the leverage and the DC operations, and I think it's a powerful thing.

Okay. Well, 2 more for you. First one, in 4Q, you reduced the return rate by about 700 basis points year-on-year, really impressive improvement. I guess, what enabled the drop? How do Delivery Promise and Thrift Promise fit in? And then what do you see as the white space to continue pushing from that metric down?

Yes. I mean it's pretty low now, right? So I would not count on it happening -- driving another 700 bps, right? There are -- there's physics involved, right, around how the consumer sort of try stuff on and what ultimately fits. So I think the Delivery Promise part has been great around just helping customers get their orders faster. And I think people -- they're sort of like if you get it fast and you -- the experience is good, you're more likely to keep an item right versus if you wait. But on the Thrift Promise side, it's been -- we've used data science and sort of AI to figure out, okay, would we be better off letting that customer keep that item and -- versus sending it back to us. And so a lot of what Thrift Promise has been able to do is only encourage customers to return items that make sense for them to return, right? And so it -- and not only does it reduce the return rates, which are good for the P&L, it also -- there's -- you don't have -- you're not shipping boxes with 1 item around. And so it reduces emissions. We think it's ultimately like good for the planet. And so -- so I think there's more room on that, Nathan, but I think we're probably at the point where it's probably at the steady state of where returns will be.

Okay. Great. Now let's close it out with one a little bit more high level. So what do you believe are the one or two things that you think investors most either underappreciate or misunderstand about the story?

I think people -- I think that if you can sort of look at the face of the financials and not understand that sort of all the competitive advantage that we've built over time is continuing to compound. It's not -- we're not a fly-by-night commerce direct-to-consumer company, right? We have a huge defensibility in our operations, our data, the customer experience that I think is only going to compound with AI. And so I think we're a business that's continuing to build for the long term. We're really committed and sometimes it drives Sean nuts because I'm like, we're not going to cut that because we're committed to the long term. And so over the last year, sure, we could have done things that would have satiated the appetite of near-term investors. And we just never -- we're just never going to run the business like that. But I do believe everything that we're doing will compound over time to build a business that will generate a lot of incredible returns. So I think people, I think, don't get under the hood on us to really understand. So I think that's probably the first. And then I think -- I would say on the AI side, I think it's easy for investors to roll their eyes around the hype cycle that we're in. But it's a transformational opportunity for us. And I think if people really pay attention to that and play with it and think about like, "Oh, yes, how does this business benefit disproportionately compared to other ways that people shop online?" I think it -- people like have a moment like, oh, yes. These guys will really benefit from this relative -- and I also will tell you like, Lululemon, we all love Lululemon. They have 900 SKUs. They take lots of photos. It's not going to help them, right? Like maybe in the marketing side, but like it's not going to materially change their business. We take 200,000 photos a day. We've got 4.5 million SKUs. It will fundamentally change how the customer experience works for our business. And so I think that's the second piece I think people should pay attention to.

Okay. Great. James, Sean, really appreciate you both being here.

Thanks for having us.

Thank you.