Great. Good afternoon, everyone. Thanks for joining us. My name is Bernie McTernan. I'm one of the internet analysts here at Needham & Company. My pleasure to have Pattern. We have Dave Wright, CEO and founder of the company. Thanks so much for joining us.
You bet. You bet. My pleasure.
Pattern went public in September. I know some of the investors in the room still may be new to the story. Can you just start providing a brief intro on the company and maybe the origin story and then what the company does?
Yeah. At a high level, Pattern will take. We sit in between a brand. So you'll have a brand. Their goal is to sell some good to a consumer. They can do that in multiple ways. One, of course, 70%, 80% of the market goes brick and mortar. We don't do anything with that. We're all on the digital side, and specifically on the digital side, our tech platform enables brands on marketplaces. So think Amazon, Target Plus, Walmart, Tmall, JD, Coupang, South Korea. So we're connected to 61 marketplaces, and we're accelerating brands across every lever of an e-commerce ecosystem. The founding story was somewhat fun because I've had a very strong distaste for, over my life, advertising and marketing people in general and just that whole philosophy and whatnot, and my cousin started this little girl's headband brand. And she sold $2 million of it called Emma Jane.
She was like, "Hey, my return on my ad spend," and of course, I'm like, "Oh, return on my ad spend." She's like, "It's 5.6." And I was like, "Why don't you put in like $1 million then?" Because her margins were pretty good. She's like, "Something's wrong. It says I get 5.6, but I don't." And so then that started a pattern, really. I was like, "OK, this became a video game." And my background is all in the data, 100% just on the tech and data science side. So it started off with just some general correlation coefficient and so forth. And then pretty soon it became, "Hey, I think we could maybe crack this nut for brands better than what exists out there.
Right. Right. And maybe can you speak to some of the technology that's underlying Pattern?
I mean, it's vast, so it's all based on a formula, so revenue for a brand is pretty simple. It's traffic times conversion times the price of the good, and then is it available? So you can war room each of those segments, and then you think to yourself, "OK, do I have enough data?" So we have a system where we say, "OK," it's almost a stoplight system where it's red, yellow, green, and we say, "OK, do we have enough data to apply any principles, natural language processing, machine learning?" We have 29 patents either issued or pending. We had AI patents before AI was even a thing, or at least, well, machine learning is AI, but what everyone talks about, and we sort of shied away from the term for years.
But that's the basis of the entire tech stack is trying to figure out how do I optimize revenue for a brand.
Right. And so I think one of the useful things to me during just the process of researching the company was going on to Amazon and seeing, like, "OK, how do your keywords differ from maybe a 1P or someone or a different 3P operator?" And maybe how the labels differ. So can you just talk to some of the actual real-world things that Pattern's doing that a consumer would actually see?
Yeah. Yeah. Well, if you think about your behavior on a marketplace, it's almost all keyword, keyword phrase driven. You go there, you're like, "I want creatine before I work out," whatever. I am taking creatine. It's just this workout part that's hard. And so you type in some keyword or keyword phrase. We started thinking to ourselves, "OK, could we possibly get close to attaching a dollar amount to every keyword or keyword phrase?" The answer, yes. You can get close enough to make decisions. Then we started saying to ourselves, "What if I have a good product, like a good creatine? Shouldn't it win some of these keywords?" And we noticed a lot of the great products weren't winning.
So then you start saying, "OK, why would this inferior product that's higher priced be mathematically advantaged on some of these mechanisms?" So then we started saying, "OK, if we know the dollar amount that would for every slot at every keyword, then we use machine learning to say, let's take this product and all of its competing products, and let's decide where it might have destiny." So which keywords or keyword phrases? And so now we make 15 million-16 million bid changes a day on the platforms. And it's almost high-frequency trading for the finance guys, keywords, keyword phrases. And that's been quite successful. But then you have to start layering that in with other technology. So that's just one driver of traffic. Then you start looking at, "OK, when they land there, how do I get them to convert?" And that's a really fun problem.
You're like, "OK, what do all these other brands do? What are their conversion rates at a keyword level?" And then you start, you usually have limited real estate. So you say, "OK, how do I position this product to mathematically drive the most dollars at the highest conversion rate?" And then you just iterate. I mean, you just do that. You can literally do this forever. And then the marketplaces are always changing something. So you adapt. And then we got lucky in 2018 with the Transformer models, the agentic workflows that we're able to do now. So we're just going to be going for years on this.
Yeah. One of the things that really stuck out to me was when I think there was the example of creatine, and one of the keyword phrases was gluten-free. And that wins because of it. And you're like, "Well, every creatine is gluten-free. But if you actually say it, you sell more.
Yeah. Yeah. We had a brand. You guys may not know this. I didn't know this either. But I guess because we told a brand, one of the findings was, "Hey, maybe add gluten-free because 28% of the creatines said gluten-free, but 49% of the sales were going to products that said gluten-free. We don't care to decide if that's a good idea. We're just doing the math." And I said, "Hey, you should put that on your listing." They're like, "You can't make creatine with gluten." And we're not doing it because they're scientific purists, right? So they're like, "We're not going to say gluten-free on a product that never has gluten." And I'm like, "Well, I'm just saying." Over time, they're like, "Never mind. We'll put it on." So you get a few basis points. And yeah.
Yeah. Exactly.
That's the game.
All right. Maybe if we can talk about the industry a little bit. Obviously, e-commerce, the main thing that everyone's talking about right now is the evolution of e-commerce and agentic shopping. Just how are you thinking about that as it relates to Pattern?
You're referring to agentic shopping from a consumer perspective?
Exactly. 100%.
I don't know that anyone really knows. I mean, McKinsey has a few studies out. JPM has some studies out. I'm sure you probably have an opinion. The range is anywhere from, say, 10%-30% that I've seen by, say, 2030 that will originate somewhere in an LLM. Now, today, the numbers are about 27% of people's searches have something to do with a product. So you're still doing product research, even if you're not originating, even if that traffic isn't, say, flowing into a marketplace where the sale is executed. But in the agentic shopping world, you're probably familiar you guys are probably all familiar with, but there's two main camps here. So you have OpenAI, and they have what's called ACP, so Agentic Commerce Protocol, co-developed with Stripe.
The idea is, I can execute a purchase from the point of consumer all the way through without leaving the chat experience. So pretty cool. You can actually, I don't know if any of you have tried or done it. It's pretty nascent. And then on the Google side, the other big player really is UCP. Of course, they can't agree on, I think when OpenAI was like, "We're going to come out with a protocol, agentic commerce protocol," it wasn't called the OpenAI protocol. But Google was like, "No, we need our own." So UCP. But the concept there is the same. And I mean, they have some slight differences. But I don't know. Once we get to a world where if you're really leveraging an assistant, an AI assistant, and you say, "Hey, shop, I need my Lululemon pants. I know I want a 30, 31.
Go get them for me," I think that it will start changing the world, and there's other areas where if you're keeping track of, say, a shopping list, and you've got like, "You add this to the list. Ask Alexa, add that to the list. Add that to the list." You're like, "Hey, go buy my whole list and have it show up tomorrow." For people who love shopping, I think they will always enjoy that experience. Like my wife, she loves to sit and research, and for me, I'm like, "That is terrible. I don't want to do it," and so having somebody go do it for me, but we did some tests with the Atlas browser, which is OpenAI's agentic browser.
And you can basically send it on a 30-minute search to say, "Find me a North Face medium jacket, apply every coupon code you can find, and give me the best price. I want it delivered within a week." And you can watch it. It'll sit and churn for 30 minutes. And you can get some pretty good deals that way. I do think it will become a thing.
Right. Do you think there will be some sort of aha moment that gets us to actually shopping that way? I mean, because right now it's, what, 0.001%? We cover Etsy. They said, "Yeah, they're part of that ACP." But say, like, "Yeah, you have higher intent when you come in that way, but it's still just really small.
I think it's because none of us, including me, it's not really live. Right? How many of you have actually executed a purchase on an LLM? I haven't either, so I don't think there's a and I've done a bunch of work because it's supposedly live on Etsy, so I'm like, "Find me a handmade wallet." I've done the search, and not much comes up, but once it starts really getting, I know that they're getting quite a catalog, but it's really not live yet for all intents and purposes. Once it goes and you have that experience, I actually think people will move.
Right. Well, then so let's flip that back to Pattern then. So the world's moving that way. It becomes 20%-30% of industry shopping over time. What can you do to make sure that Pattern to make sure the agents are finding your products, not someone else's?
We have a data advantage. If you think about all the keyword, keyword phrase data, we have 46 trillion data points. We're collecting 120 billion data points a week. We can take things like let's go back to our gluten-free creatine example. Maybe you're like, "OK, we know that this product does well on gluten-free creatine." You can reverse engineer that into a question. Our scorecards for brands are, we'll just send to all the core LLMs, which creatines are the best if you're looking for gluten-free. It will respond. We do reach, rank, and sentiment. We start saying, "OK, does the brand show up? If it does, how does it rank against other brands? What's the sentiment?" You peel into that. You can start getting a sense of when an LLM views your product, what does it think?
So that's part of the equation. And then for Pattern, we don't really care which platform wins. If you take in Mexico, for example, right now it's Amazon going up against Mercado Libre. And we represent Pandora jewelry. And we don't really care if Mercado Libre wins or Amazon wins or OpenAI or Gemini. We're just trying to help Pandora. So in a way, we just are agnostic to where that sale happens as long as Pandora wins. So luckily, we're agnostic that way a bit.
Right, and I think it gets back to, and I know you've made this point before, but just agentic commerce entering commerce or entering e-commerce, that's making the whole transaction more difficult for brands to figure out.
Yeah. I guess as long as the more channels there are, the more complex it gets. I don't know if you guys have all done this, but if you sit down with a brand, I was shocked because I'm coming from a tech background. Most brands are run by like three to five e-com people, even the big ones. And you start saying, "OK, here's what" a lot of them don't even know what ACP and UCP is yet. So it's really hard for them to have that coverage. That's where Pattern comes in. And good assist.
I mean, the S-1, that was the most important data point. It's like the 80% or whatever it is of less than eight people. It's like you have hundreds of employees doing this versus eight.
Yeah. I mean, I got 400 people on just the tech alone. And our budgets there are increasing about 50% this year because we're trying to keep track of some of the agentic workflows. And then you have a poor brand who they got like a spreadsheet guy at best.
Right. Right. Maybe one last one on industry. But we just went through the holiday season. What's your pulse on the consumer right now?
I mean, it was pretty good. We get this question a lot. Like, how does Pattern see the consumer? Because we're doing quarter to date, roughly $2.3 billion in total revenue, which seems big. But when you look at consumer behavior, I almost prefer to go to like Visa, Mastercard. Because even Amazon, we benefit anyone who's like a Pattern, we're growing our brands more than, say, normal. So I'm always like, "Oh, everything looks healthy." But if you dive into like Visa Mastercard data, I think the numbers for the holiday season for e-com were about 7%-8% up, which is OK, 4% up or so in retail. Record year, but it's always a record year.
Yeah.
Hopefully. I guess if we ever stop having record holiday seasons, we've got another problem.
Yeah. All right. So let's focus on the Pattern business. So NRR, which I think you think is the most important metric for the company, grew from 115% in 1Q 2025, next quarter 118%. Then you just reported 122% last quarter. What's been driving this acceleration?
NRR is the most important number. It's how you know that our machine actually works, and it's how we know. For one, if a brand leaves us, NRR will go down. If I can't grow a brand, NRR won't grow, so it is somewhat of an if I can take one brand that is currently with us on in Mercado Libre and take them to Coupang South Korea, my NRR will go up, right? So any brand expansion I can do or growth I can do, NRR goes up, and we've had some pretty significant tech improvements this year, but scale is just a lucky thing in our world and is a game changer. I remember in the early days, there were lots of things I would love to do on the tech side.
But the way we pay for tech is if I implement this change, will it drive enough revenue to pay for the team to build that? That's how we sort of think about it. And back when we did $1 billion, a lot of the things were a no. And now that we're doing $2.3 billion, a lot of those things are becoming yeses. When we do $5 billion, even more become yeses. So the NRR should continue to perform well. But I mean, it was a record quarter.
Right. Can you walk through just how so a brand comes to Pattern, just how does that life cycle evolve? Meaning them giving you more products to sell typically or maybe starting on Amazon and moving to other markets, moving to other geographies, just trying to walk through the components of NRR?
Yeah. A lot of times you'll sit with a team or a brand, and they're like, "OK, maybe let's give them something we don't want to do." So maybe they have no presence in South Korea. And they're like, "Well, hell, I guess we're zero. So give them it. We require exclusivity. So for us to handle a channel, we need exclusivity to the channel." And then we do everything. And the way we monetize it is we actually buy the goods. So we'll buy the goods, we sell them, and we take our margin in the middle. And so oftentimes they'll give us sometimes it's a product line. Sometimes it's somewhere they don't care about that much. And then they're like, "Wow, that was better than I thought." And so then you just land and expand.
And that's where you see maybe I didn't explain that NRR is net revenue retention. So that is overall growth from a brand that passed their first 12 months. We keep track of how many dollars more than the first 12 months, and we continue to grow that.
Right. Thinking about customer acquisition, what's the pitch? How do you acquire them? How long is that sales cycle?
It's a roughly 90-day sales cycle. So it's not too bad. But I mean, and the pitch is getting easier just because we're getting more and more sophisticated, and the costs are going down to work with us. I'll give you one example, logistics. I mean, most of you probably, Amazon is a good way to talk about this because everyone's familiar. If you want to get a product into the Amazon network and you have 100 of them, Amazon has a regional node concept. So they'll say, "Hey, I already have 13 of these in New Jersey, and I need more in California." So at the point when you say, "Send it in," they'll tell you, "Send 12 over here, 12 over here." And so that's quite a complex for a brand. That becomes pretty expensive. Because what brands want to do is full pallets, full trucks.
If Amazon's like, "Hey, send 12 over here and 13 over here," the full truckload idea for most brands because it's between 13 locations and 17 locations per brand you now have to go to. So a brand can't do that. Pattern, I mean, we have a hard time. But right now we're at 95% full truckloads that leave the building. And if it's in a full truck, it costs me $0.11. If it's not, the cost goes up by four times for LTL, so a partial truck. And for a partial truck, if it's partial, you're talking a couple of dollars. That middle-mile logistics capability is a big cost savings for brands. So now we can go and I think we've reduced our what we call a Cost to Serve by three points since 2022.
So we're getting less expensive in an aggregate basis, and then we're getting more capable. So that flow is just getting healthier.
Right. And so I remember hearing from your head of go-to-market just how they can go to brands and say, "Hey, this is how much revenue you're doing now. And here's the four things we can tweak, and it's going to become 20% more revenue." Can you just talk about some of those tests that you're not tests, but projects that you're running to help the go-to-market?
Yeah. Well, it all comes down to that formula. So we'll identify, "OK, what are all the drivers of traffic?" One is, how does their product compare to their competitors in a lot of different dimensions: price, reviews, sentiment, customer reception, all those types of things. And then we start developing what we believe is destiny. Then we start measuring, "OK, how are they converting? Where are they winning? Where are they losing?" And we just map all that out. And we say, "OK." And some of it's geography. We're like, "OK, we believe" and we'll look for, say, branded search in Germany. If they have reasonable branded search in Germany, we know that there's probably an opportunity people are looking for that brand already in Germany doesn't exist. So that will be a lever.
But you start putting that all together, and we'll just map out for a brand. And we'll say, "OK, here's where you are today. Here's where we think we can take you." And that delta is the selling point.
Right. Right. OK. Competition. Who do you view as your primary competitor?
Remember our monetization models. We buy inventory. That has turned out to be great because all of the tech platforms that would actually compete with us and do well, they hate the idea of taking inventory because they want the SaaS model or whatever. Never mind the fact that we actually throw off free cash flow, EBITDA, and beat them 10 ways to Sunday, in my opinion, on all their financial numbers. They still don't want to do it because I don't know. Maybe they go to the Needham conference, and they're like, "Somebody said TMT is the best way to go." And so they stick to the SaaS model. I don't know. But so generally, we hardly ever compete with the folks that would truly be competitive to us.
And then there's the set of maybe what you might refer to as distribution folks who are better on the logistics side, like the middle-mile that might compete with us there. They rarely are good at tech. So for now, we're sort of alone. We thought once we went public that people would watch it. They'd be like, "Oh, wow. OK, so $2-$3 billion valuation. And maybe we should do this." So far, we're not seeing it.
So far, nothing. How about thinking about Amazon as a 1P competitor?
Yeah. Amazon, when you say competitor, well, maybe I'll give you an example. We were just in Canada yesterday, and Kellogg's asked us to cover Canada for them on cereal for e-com. So when you really start looking at it, you're like, "It's a $4 box of cereal. It's like this big, yeah." We could not do it, but 1P does it because they're willing to take a loss on that. I know that 1P business is shrinking over the last probably 15 years. I think Bezos' first letter to shareholders, they were at like 3% when they introduced the marketplace. Now it's 67%. Some people speculate Amazon's going to try to get out of the business of carrying inventory, especially with Jassy, a platform guy. I don't think they can because they want the everything store.
And there's a whole set of products that someone has to be willing to make loss leaders and to run that. I think that will always be a thing. And then the rest of the business, they don't really it seems that they're pretty open on what that model looks like. But the financial incentive, Amazon makes about $0.40 for every dollar that goes 3P. So I mean, I don't know that they have a big dog in that fight.
Right. Right. You touched on it earlier, but I just wanted to dive into fulfillment. Can you just talk about the fulfillment network and how has it changed over the years and maybe just reemphasize how big of a differentiator it is?
Yeah. Well, we underestimated in the beginning. I mean, it's like anyone else. In the beginning, it was Costco tables and people working. And we're just trying to figure it out. And I was never inclined on the ops side. My partner handled. She was our COO. And she did a pretty good job until one point she was like, "You know what? We need some help." We ended up going through a few people. But we hired a guy who was the youngest ever, to our knowledge, Level 8 at Amazon, which is managing about 5,000 people on their logistics. He was 27. And then he went and did a startup for four years, came back, and we managed to get him. For a while there, he was making more in terms of just base salary than me and CFO, everyone at the company.
And it was worth every penny because he has just transformed that world for us. He brought on the head of supply chain North America from Amazon, some great folks on the reverse logistics side. And now if you go into a Pattern warehouse, you're like, "Damn, this is a full operation." And it is really strategic in terms of cost.
Right. Yeah. No, I mean, he's very impressive. And yeah, also it looks like he's 12 years old too.
He does look like he's 12. Yeah.
How much more capacity could you handle in your just current footprint, just trying to think about the need or if there's a desire to build out more fulfillment centers?
We don't. Rob, who's the guy we're talking about, he always says, "We're not in the business of acquiring concrete." So we're strongly averse to just expansion that way, and over the years, I remember we had decided we have our initial building in Hebron, Kentucky. We had decided that we could process X amount of units in that space. I think it was 750,000 units a month, and two years later, we revamped some processes, built some tech, and Rob's like, "Actually, it's 2.3 million units now," and so it seems like we're just getting more and more efficient on the concrete, so I'm not sure where the limit is, but it's not that expensive if you lease a building in either, so you expand a little bit. Right now, we have three fulfillment centers. We're opening a fourth.
But that was more for coverage if we're not in a fulfillment network, like refrigerated products or some of these things are more difficult to finish. We will actually do end consumer fulfillment on some areas.
OK. Got it. Maybe let's transition to three, I would say, the three areas of investor concerns. Or when I'm first talking about the company, these are like the three things everyone always brings up first. So I'd love to get how I know how I answer it, but I'd love to hear how you answer it. First, just Amazon risk and Amazon concentration, over 90% of revenue. Just is that a risk in your view or not?
Well, I mean, there are some things that I think maybe investors are definitely overstated there. I read the forums. You shouldn't do this when you go public. You can't help yourself. People talk about you on X or whatever, and you're like, "It's just wrong." And you want to get out there and fix everything. And you're in quiet periods. You can't do it. But one of the things that gets brought up all the time is, "OK, because we run on about a 6% margin today." And people are saying, "Hey, well, if Amazon changed their fee structure and adds 6%, these guys are dead overnight," type of thing. And we, of course, have thought that through already.
All of our contracts have a pass-through language that if a fee or a commission or a fulfillment fee or whatever changes, we can change the cost of goods immediately. We stay completely whole on all of that. None of that is a risk. The area where if Amazon decides to go a different direction on the 3P marketplace, I mean, that would I mean, we still have our entire tech stack. At that point, you might be happy. We might just monetize the tech stack. We're already starting to do that for some brands. We call it 1P Accelerator if they're working with Amazon, like the Kellogg's example. We're starting to add solutions where it's just pure tech on that front. It would change our revenue model if Amazon decided, "I don't want to do 3P marketplace," or, "I want to shrink that," or whatever.
But that would be, I mean, that's just not the direction they've gone for 15 years. So, I mean, everything's a risk. It's in the S-1. We marked it as a risk. And then you just have to decide how material it is.
Right. While we're talking about marketplaces, so aside from Amazon, what marketplace are you seeing nice growth in right now?
As reported in our earnings for the third quarter, we saw 391% growth on TikTok. We have a great relationship with TikTok. We had one of their, the guy who actually started TikTok shops in China. He came to Pattern. He made a couple of comments that maybe you guys could think about. That changes your thinking. He was like, "Hey," because at first, I was like, "You have no chance of beating Amazon." He was like, "No." His view is like, "We're going to be bigger than Amazon." OK. I'll hear this out, I guess. He was like, "Well, in China, we beat JD. They have six-hour delivery across all of China." OK. Listening. He says, "If people spend time and I can entertain them, they will buy from me. I've proven it in China.
And I will prove it in the US." And he is so far doing a pretty good job, and the engagement isn't just for. I think his average engagement in the US is over an hour a day per person, which seems astounding. And then he's like, "If they spend time on my platform, I will sell to them." And he's proved that out so far. So that one's an interesting one. Coupang in South Korea is our fastest-growing marketplace we've ever onboarded, actually faster than Amazon. Now, of course, we're bigger now. But that one's going well. I think there's some other more niche marketplaces that an Alibaba. There's Mercado Libre, I think, going to be really relevant for many years to come. So I think outside the US, it's quite competitive. The team at JD and in China and so forth do well.
Right. And so you just added Coupang. How long does it take for it to become a more significant revenue contributor?
I mean, all those things take a while. But the main thing for us is we have what we call a level one through seven tech integration. And full integration with our stack is seven. And I can operate so much less expensively. If, say, I can only do a level three integration, a lot of it is the data they provide us and the APIs we have access to and whether they execute what we need in order to operate. Coupang and even TikTok's like, "Hey, what do we need to do to get to a level seven integration?" So now that we're getting bigger, we're getting more lean in that way, which is faster. But you're still talking six months, year, before you're like, "Does this even get material enough to matter?
Investor concern number one is Amazon concentration. Number two is just exposure to health and wellness. So 63% of your revenue in 2024 was in that category. How do you think about your reliance on that category, maybe diversifying it over time?
Yeah. I mean, one of the things that Jason and I have been doing, Jason's our CFO, is we're trying to get to the root of why people, investors, care about that because it's such a great category. And the way we got started in it is I had this idea. All entrepreneurs have some weird idea. And I didn't want to have a bunch of salespeople. So I thought we could hit a billion in revenue. It would be a cool badge to hit a billion in revenue with one sales guy. And the way we did it, I'm like, "If we build the right machine, we'll win." People look at their market share tools. And they're like, "Hey, what's going on with that brand? Let's sign up." So health and wellness was our first category. We won there.
And now we have, in all transparency, we have 60 supplement brands. Because we win one, they're like, "What's going on over there?" They call. Next one calls, call, call. And we didn't need a sales force because all we had to do was take inbound. The problem, until we got our Series A investors, and they're like, "Hey, this one salesperson idea, cool idea. But we want more." Right? So we diversified. Now we're getting a toehold in some of these other areas. No one called us from sports and outdoors because we won vitamins. So now we're getting a toehold everywhere. So I think that traction will just keep growing. I think the risk people think about in health and wellness is maybe that it's a fad because there are some brands.
But most of our brands carry every product you can imagine, from the vitamin Ds to the fish oils to the creatines, so I don't see it going away. I don't see it. It's not like, and it's not retail. I know that people are quite nervous, like, "Is this a vitamin shop or a GNC?" Because they've had bad experiences there. It's very different than that. So we're not really shying away from the category. But we're just investing more in others.
Right. Right. And maybe last one, we can call it inventory risk. But I guess, how do you think about buying inventory? Obviously, you said it's an asset. And you view it as an asset, not a liability. But just expand upon that.
Yeah. We've gotten very, very good at it. We didn't start out this way. So a few things you track. I remember when I hired Jason. Jason sat down with me, and this is early days. And he says, well, actually, in his interview, he goes, "So are you guys cash flow positive?" I said, "Yes." I didn't really know what that was. And he's like, "No, you're not." I was like, "OK, well, what's the difference between profit and cash flow?" He's like, "A lot." So we have since been now our days of inventory on hand, which was well over 100, is now down in the low 70s. And we're making great progress. We've had some months in the 60s. And our days payable, which we had this idea we're going to be very brand friendly. And we didn't want to go to the bank very much.
So Mel, our co-founder, she just paid all the brands every Friday. And they loved it, as you might imagine. But cash flow, Jason's like, "Hey, if you don't do it." Jason's like, "Hey, if you would have met me, you would have never had to raise money." Like, "That would have been nice." And so now that we've fixed some of those things, now days payable is in the 60s. DIO is in that mid-70 range. And then brands remit every within generally about two weeks. So now we actually generate cash as we grow. So now that whole machine, if you take a step back, big picture, by the time we went public, we actually had more money in the bank than we had ever raised. And we have no debt. And we maintain no debt. And now we have a $250 million revolver.
So in a way, we have a lot of weapons that way from a balance sheet perspective. And I don't know of any other pure tech company that has a balance sheet like ours. So it is a great asset if you do it right. And I didn't in the beginning. And in the last five years, we fixed that. And so now it's become quite a strategic asset. But it takes a lot of discipline to get that right.
Right. Right. Maybe moving back over to the tech side, just a general question. But what are some of the most exciting parts of your tech that's on the tech roadmap right now?
I presented at both Amazon's re:Invent and at Harvard on agentic workflows. You guys probably know this. The biggest game changer for us, if you are running a company or involved in a company where there is a lot of data, so data rich, and the data can trigger sensors, if you have a sensor and then you can actuate that sensor, and if that flow leads to a material change in dollars, then you are in a spot where an agentic workflow, like the typical workflow, would be, "OK, here's an input. Here's an output. Steps one, two, three, four in between." In an agentic world, it's the same model, except you have long-term, short-term memory. The tools themselves can consume data and make decisions on the path they go next.
And so we now will have, say, a market research agent that will research the breadth of digital shelf, all the competitors. And we've trained it that way. And then we could actually have a content creation agent that builds the content. But then we can also audit it with another agent. So we're like, "Hey, did that whole thing work?" And so that and our ability to execute it globally at a low price point via agentic workflows is staggering to me. I didn't imagine that it would come this far this fast.
Right. Right. Maybe we want to pause here and see if there's any questions from the audience as we're in the last minute or two. But if not, I'm certainly happy to keep going. Why don't I just keep going then? Margins. So obviously, you book revenue on a gross basis. So the margin's optically lower. But there's been nice margin expansion over the years. How should investors think about that margin going forward and maybe areas of investment that you have?
In 2022, our margin was 1.8%. In 2025, thus far, we're tracking about 6%. We have not gone to brands and raised any prices. We probably have some little places here and there. I don't want to. Maybe I'm overstating that. It's not something we go about doing. If we're off on if we thought a product's mix was whatever, I'm sure we've had to raise prices here and there. That is, if you talk to our brands, it's not what we do. We've gotten that expansion at just being better at executing the machine. The guidance that we've provided, which is more long-term, and by long-term, we're talking like long-term, is 10% in terms of margin. The path to get there, one is just gaining efficiency in the machine.
And then two is we have some other monetization models that are working pretty well. Like we talked about the middle-mile logistics. We now started. There was some company that was selling that middle-mile logistics called MyFBA Prep. We saw them show up on the private companies' fastest-growing list. I think it's the Inc. 5000. Anyway, they were 44th on the list, growing 8,000% CAGR over the last five years. I'm like, "What's this?" And I went out and looked at all their pricing. I'm like, "Wow, for us, it's about a third. We can be about a third of that cost." And so we've started monetizing that. Actually, we just crossed the 50% line where more than half of our trucks leave our building with non-Pattern-owned units than Pattern-owned, all high-margin business. So that's one business that we just started last year.
And then just on the tech side, monetizing just we have to build the tech anyway. And if we can expand that, then might as well.
Yeah. Well, Dave, let's leave it there. Thanks so much for joining us. Thanks for the audience and everyone on the webcast, and have a great rest of the day.
Thanks.
Nice job .