Randy, Jim, welcome. So my name is Ron Josey, I cover the internet sector here at Citi, and look, it's always exciting to have, exciting to have Randy here with us, Randy Altschuler, the co-founder, CEO of Xometry, and Jim Rallo, Xometry CFO. You know, I'm sure most of you know what Xometry does, but it's a leading online marketplace for manufacturing. It's got 48,000 active buyers, 2,000 vetted suppliers, I believe, and you know, we'll get into exactly everything that Xometry can do, 'cause it's growing every day with galvanized steel and everything else. But with that said, Randy, Jim, welcome to the Citi Global Tech Conference.
Thank you.
Great to be here.
Great to have you. Great to have you today. What I want to do is sort of jump into maybe some, some of the questions that we often get are just... It might be very remedial, but just help us-
Of course
... here. You know, help us understand just the problem Xometry is solving, right? And the opportunities, and specifically... So that's high level, but specifically, we like to hear examples about, you know, we talk about mission-critical work- that's coming out. Give us examples here, and, and how Xometry's marketplace is helping to solve that.
Sure. So, yeah, we're addressing one of the largest verticals in the world, which is manufacturing, and within that multi-trillion dollar market, there's a sub-segment of custom manufacturing that is done by small to medium-sized manufacturers, and that's about a $2 trillion market. And that's filled with hundreds of thousands of small manufacturers across the world. Here in the United States, there are about 600,000 manufacturers. 75% of them have less than 25 employees. And so in this heavily fragmented market, it's very difficult for the customer to find the best partner for their particular order. So there has been, before Xometry, there's been no central, online place for them to go to find the best price and the best lead time. And pricing in this market is opaque and very inefficient.
So if that customer went to two different small manufacturers and asked them for a bid on a set of parts, the price differential could be 100%. The time difference could be 100%, and it would take that customer days, maybe even weeks, to get that price and that lead time. So for a customer, they have to source these important parts, and I'll give some examples, but doing it is very painful, and they're dependent on these small businesses that have minimal online presence. They don't have websites or very remedial ones, so that's very difficult for them. And for those small manufacturers, because this is a two-sided problem, the long tail of the internet hasn't touched them yet. So I'm a machine shop right outside of Detroit, and I historically have done all my work in automotive.
I have no visibility into what might be opportunities available in aerospace in California or medical device companies in Boston, and I'm selling something a little bit more difficult. There's no... I'm not a retail site. I'm not selling off-the-shelf parts. I'm not a service company where I can charge hourly. Instead, I'm selling my capacity, and I'm making custom parts. So Xometry said, "Hey, we have an opportunity here where customers are struggling to find the best deal, and suppliers are struggling to find work that not only fills their capacity but optimizes their profitability. Let's use a marketplace, which has been done so successfully in other industries, and unite and create that two-sided place.
And then we're going to use AI, artificial intelligence, to create pricing. Because, you know, as you see in many marketplaces, you can get an instant price, and you can transact in real time. In custom manufacturing, that wasn't possible before. Again, it would take days or even weeks to get pricing. So we use AI to create an instant price for both the buyer and ultimately the supplier, and to match the buyers and suppliers together. And just to the second part of your question, what kind of things happen from that? Well, super important things, like... So, for example, when Moderna during COVID was mixing, making the mixers for the vaccine, they used the Xometry marketplace to help source parts for those mixers.
If you have a BMW car and certain car, certain cars, and you have to go into your dealership for service parts for those cars... You know, those can be made in the Xometry marketplace. If you're in the airport and you see those Clear stands, Xometry makes a lot of the stuff for Clear, or in sports arenas. If you see satellites, you can't see them, but if there are satellites in space and things being launched, parts for those are being made by the Xometry marketplace. So mission-critical things are being done. Medical devices embedded in people's bodies are being made via the Xometry marketplace.
Clear medical devices, BMW, the list goes... Moderna. Why are they using Xometry versus their existing manufacturing?
So the bigger the customer, the larger the company, the more attractive the Xometry value proposition. Because if you're a large company, let's say a BMW or a large aerospace or defense company, you're spending $hundreds of millions a year, sometimes $billions a year, on outsourced manufacturing, and you kind of put that into two buckets. One bucket is where I'm doing millions of units, and I want to own that relationship directly with the supplier. So again, let's go back to the example like a BMW, where they're doing the steering wheel on every car, I want to own that relationship directly.
However, these other parts are really important, but I'm dependent on these small manufacturers, and so it's very expensive for me to manage hundreds of small manufacturers, and they're risky because they're small businesses. They're prone to financial difficulties, or somebody gets sick, or there's a weather problem, or maybe they're in Asia, and there's a geopolitical problem. So I need these parts badly. I need these assemblies badly, but boy, oh, boy, I'm trusting somebody very small to do that.... So the customer experience is bad. Like, there's no online service, there's no instant pricing, all those things that I'm used to getting in so many other parts of my business, that doesn't exist for me here.
So if Xometry's able to be that one throat to choke, the one place for them to go, instead of having to deal with 200 suppliers, I'm dealing with just Xometry Marketplace and one interface, that's super powerful. I mean, think about it, to a certain extent, like with Amazon, right? Where they've got, when you go to Amazon, and they've lots of stuff offered there. Amazon's not doing it all, but it's one interface. They make it so simple. In this case, these parts are even more important. It's not just about buying toothpaste and jeans. It's about, "Boy, I need these things to deliver my, my end product to my customer." This provides a lot of surety, that I, that I have to have.
Super helpful. That's great. So let's talk about to make sure that you have, or Xometry can support all of the demand that's happening. You know, we often think about just the marketplace menu, and, you know, as an internet analyst, I rarely get to use galvanized steel or stainless steel in, like, my reports or, like, how we think about it. But talk to us about how the menu has evolved and how the product has evolved, and what goes into that, like, core technology, so that if I'm an aerospace and defense contractor, I know I can use Xometry for an O-ring, for example-
Sure
... for the turn of the 1980s, right? Like-
Exactly. Don't put it in the water and everything.
Right, right, right. Exactly. Be careful.
So look, this is where technology's obviously so powerful. It, it has made our platform very extensible, and so manufacturing, particularly custom manufacturing, is broken out into hundreds of thousands of use cases. With our technology approach and with using machine learning algorithms, we're able to offer lots of different slices and lots of different things, and we're constantly adding to that menu, so we don't have to go out and buy a bunch of capital equipment. Instead, we can use these algorithms and our software platform to add all these things in a very cost-efficient and quick way to do that. And the good news is, from a network perspective, in terms of getting the manufacturers to do it, we are the leading two-sided marketplace.
There's not, there's not, like, other places for them to sell their open capacity like there is with Xometry. Not only do we have that technology that's very extensible, which we're leveraging, and we're getting smarter and smarter as we roll out into new geographies, as we offer new processes and new materials, we're getting smarter and smarter and faster and faster at that, but also, the suppliers that make the parts, they're... We're growing that very organically, and people are attracted to it because, you know, in our market, we're the leader, and it's not like a B2C marketplace where you've got lots of people there bang down the doors. In our case, we're the guys.
One of the things as you add newer tools, and you add these capabilities, and Q2, I think, the Instant Quoting Engine was updated even further.
Yes.
And so talk to us more about just how this helped to improve conversion rates, and maybe does the team know, or do the guys know, what Instant Quoting Engine is and why it's unique or differentiated to Xometry?
Yeah, so just to give people a sense of it, like, again, you're used to when you go online to a retail site, whether it's food or clothing, whatever it is, you're buying off the shelf, you're buying a SKU, right? And there's a price right there available for it. In this case, imagine you're a satellite company. You're building parts for your own proprietary satellite. There is no off-the-shelf part. You're building something that just works for your satellite or for your medical device, et cetera. That has been, as we sort of said in the beginning, it's taken forever for you to get pricing on that, and pricing's all over the place. So by using machine learning, or we use a technique called deep learning, we're able to actually give you a price right away, and that is a...
If you're in the manufacturing world, that's like a wow factor. Like, "Oh, my goodness, I can't believe I can get it right away. Nobody's ever tried this. It's risky. Like, my goodness!" And so that just creates... That enables us to offer an e-commerce-like experience that wasn't possible in the past in the custom manufacturing world. As we add more processes, that enables us to go deeper and deeper into our customers, particularly the larger ones, where when we talk about how they're forced to use many different small manufacturers because they have many different needs.
If I can offer all of that in my quoting engine and in our marketplace, then they only have one place to go, and that gives me the ability to go to a large company and say, "Hey, you are spending $10s of millions a year, maybe even more, in these 20 different categories. We've added them here. They're all here. Come just right here to Xometry and leverage our marketplace for all your different needs." So as we're expanding what we can quote, what we can offer, that ability to be the one-stop shop is constantly improving, and it's enabling us to go deeper in our... You know, we have a land and expand strategy, as many companies do. That enhances our land and expand strategy.
Super helpful. Let's switch topic a little bit. We went deeper a little bit in the... I know we can go deeper, but in product. I wanted to ask about the customer base, because we often get the question, like, "Who, What does the customer base look like?" Is it the BMWs and Modernas that we talked about? You know, anything along those lines? Because we've been pretty impressed with the active buyer growth, annual active buyer growth, so.
Yeah. So, you know, we've a lot of customers. So you mentioned in the beginning we have a record number of active buyers, 48,000 buyers, so that's folks who've bought within the last 12 months, and that's been growing very rapidly. Our year-over-year growth in active buyers was, you know, in the mid-40s.
Yep.
So the great news is we've got more and more buyers, so top of funnel's strong, lots of new people coming in, and we've got some really great cohort data that people, if they wanna look in our investor deck, but we also have some really big customers, too. So the 200 largest accounts or companies that order from Xometry in the U.S. made up, in 2022, about 50% of our marketplace revenue. So and those are big companies. If I told you who they were, you would know who 198 of those 200 were, and those are leading companies in everything from aerospace, defense, robotics, medical devices, industrial.
And again, these large companies are very attracted to the idea of being able to consolidate their spend and the surety and, you know, resilient supply chain that Xometry offers. The good news is, and we've got, again, a really good slide in our investor deck with some case studies, even though we've grown really rapidly with these existing accounts, you know, the CAGRs are off the charts, we're still just a fraction of their spend. So one of the reasons why we are so confident about our continued rapid growth is in these larger customers, we're still- so we have a long way to go. And we're making... You know, some of that's just gonna happen over time. Some of that's through enhanced sales efforts, or land and expand strategy.
Some of that's some technology tools that we're implementing to make that easier. So those large customers enable us to grow profitably and continue that rapid growth, alongside of adding lots of new customers as well.
And so as we think about, you know, I think, you know, focusing on the top 200 was a key point, like the five-point plan. We also just hired a new head of enterprise sales, I think, as well.
That's right.
So I'd be curious, as you think about how the business evolves from here, land and expand on the top 200, but also continue growing new business, how do you manage a sales force in that way? Or what? Yeah.
Yeah, so, you know, the top of funnel is largely digital marketing, so that's a very efficient tool to bring in new logos, new companies, as well as new individual users. We- you know, our Instant Quoting Engine is very accessible. We've actually- we've enabled it to be embedded in- within popular 3D CAD programs from companies like Autodesk or PTC, or Dassault Systèmes. You don't even have to come to our site, you have add-ins on those things. We're trying to make our quoting engine ubiquitous. That is a great lead gen opportunity because people don't usually quote parts, you know, on a whim. You know, and so, that-
Conversion, so if you're getting a quote, you, you know they're looking for something.
They probably are, yeah. I mean, not always, but it's not, you know, it's not like, you know, casual shopping in a retail store. So that gives you a really good way to convert buyers without having a salesperson touch them. And that's, you know, that's a really good CAC and et cetera for us. As you get further down, and as we develop, as those customers become bigger or they're large companies, that's really where we want to deploy our sales team.
We brought in this gentleman named Wes, who had been at Salesforce for a while, helping them stand up their manufacturing vertical, most recently was at ZoomInfo. He's helping us with our... with that enterprise sales team. Digital marketing, big funnel at the top, lots of conversions without anybody touching it. Majority of our customers don't talk to anybody at Xometry Convert, but for those larger customers, doing that land and expand strategy with sales folks who have that experience of selling enterprise. Again, that deal looks like talking with people about committed spend versus it just being transactional. We're also making technology enhancements to our platform to make it easier for groups of buyers to work together.
So in these larger companies, as we're working on larger projects or larger products, there's multiple parts involved, sometimes multiple assemblies, and there's a level of cooperation that has to occur. Because they need to know, if you're gonna successfully complete this full product, you need to know where all the different parts are arriving, when the assemblies are gonna be made, because they all come together. So we've been rolling out technology to enhance that collaboration within Xometry itself. We've been beta testing some of that software with four large customers of ours, large companies, and we expect to roll that out to our, to our top 200 customers in the months that follow. So alongside of a, an enhanced sales motion with, with Wes joining, we're also adding technology tools that will make it easier for these enterprise customers to commit more spend to the Xometry marketplace.
And that Teamcenter which I think is really interesting-
Yeah
... are in the four large companies, and I know we're gonna expand the beta or whatever but are you seeing that broadening of the buyer funnel?
Yeah. So one of the things that we've learned, which is just fabulous... So the way it works is, you're at a large aerospace company, and you're using this team software that we've got, and you invite people to join. You know, we're trying to make this, whatever it may be, this assembly, and you invite 25 engineers or procurement people to be part of this, to all collaborate together. What we've learned is that a very healthy percentage of those folks who are being invited are not today Xometry users, registered users. So it's actually turned into a great customer acquisition vehicle for us. And it's very powerful because similarly, the way, like, let's say, like Slack, you know, Slack grew so rapidly within companies because if... it became the communication, that's the way you would communicate, so you had to be in Slack, you had to start, you know, you had to do that. This is also enabling us: "Hey, if you're gonna be working on this particular assembly or production order, you've got to be part of this team that's now on, in the Xometry software." That's enabling us to add these users, and that, that's really good for us.
Something to look out for.
Yeah.
So, beta testing now, is this end of year 2024? How do you think about Teamspace going live?
We're pushing hard, and we're expanding the number of customers that we're doing the beta testing with. We're working in a very agile environment, so we're constantly rolling out enhancements, we're learning things. You know, we're talking. The great thing about doing beta testing is very active dialogue between you and the customers who are testing it. In return, you're making the changes they're asking for. So we're being responsive to what they're looking for, so it's gone really well.
... That's helpful. That's great, and we, you know, if you sort of back in active users and pricing, you can get order growth-
Yeah.
Maybe on average.
Yeah.
And so just, we're pretty, you're seeing pretty strong order growth, right? So we just talked about active buyers growing 44%. We talked, you know, I think orders growth is maybe 35%-40%, I don't know. But talk to us about the drivers of order growth is point number one.
Yep.
Question number one, and question number two, I'd be really curious, just the broader macro environment and what you're seeing there from a demand perspective?
Sure.
-and supply.
Yeah. So we don't disclose orders, but I think the way you're thinking about it in terms of active buyer growth is a healthy way to think about it. And so this year, our—you know, every quarter, our active buyer growth has been in the mid-40s, so that's been very strong year-over-year growth. I think it's even stronger than the growth we had last year. And that growth is coming in from a couple different ways. In terms of the active buyers joining us, that's coming as we expand internationally. So international has been an area we've been growing very rapidly. And just to put that in perspective, in 2020, which is our first calendar year that we were international, and that's really Europe, we did about, I think, $3 million in 2020.
Over the last 12 months, we've done $46 million internationally, so that's growing rapidly, adding active buyers. We're also seeing within, as you know, as we're doing our land and expand strategy, getting more active buyers from our existing accounts, as well as adding new logos. So it's coming from all different places, as well as also from our Thomas acquisition, which we can talk about at some point, too. That's all happening nicely. One of the dynamics. So the macro, and this is something that's hurt us this year, in terms of our top-line growth. So historically, if you look at 2019- 2022, and Xometry went public in 2021, but from 2019- 2022, our marketplace revenue grew 59% year-over-year.
You know, our CAGR was 59% for the last 3 years before this year. Our buyer and so that's really rapid growth. This year, that growth has been tempered, because even though we've had robust growth in number of active buyers and really robust growth in number of orders, what impact macro has been is things have gotten less expensive, and customers are trading down or looking for... They're trading price for lead time. So customers are saying, "We offer different options in Xometry. You can pay more to get it faster." And so customers, starting in sort of the end of the third quarter of last year, started saying, "Hey, I'm more cost-conscious right now. I want to pay. I'm gonna pay less and get a longer lead time.
I'm okay with that." So starting at the end of Q3, and really in Q4 of last year and through this year, we've had that trade-down, and so that's impacted our top line. The good news is that, we're about to lap that trade-down, so it's actually in September of last year that we saw customers start doing things for less. We're in September now, and things have sort of stabilized. So our growth in active buyers will converge with our, will much more converge with our growth in revenue. There's been a gap this year as even though our active buyer growth has been in the forties, our revenue growth for supplier, for marketplace has been lower because of that historic trade-down that happened, you know, at the end of last year.
We're lapping that as we come into the end of this quarter and into the fourth quarter.
And so since we brought it up, help us call it last year 4Q22 to current, and-
Yeah
... maybe you start here, and Jim, I'd love your thoughts as well. Just what were the changes? So it's sort of a macro environment if a buyer is willing to pay less and have a longer lead time. Like, what—just help us bridge Q4 or Q3 2022 to current, because I think there were changes in the platform, too.
There were. So, you know, again, the fundamental metrics are really healthy. We've had record, you know, we've been adding record number in Q2. Q1, we added a record number of active buyers. Active buyer growth year-over-year is really strong. Order growth is really strong. So we're taking market share, and this is obviously a huge market, so we're taking market share. And so that's very, very positive. When we did see that people were willing to trade things down, we're looking for longer lead times, we started offering less expensive options, you know, that were... You know, in the past, we'd offered options that were really cheap. We offered sort of like a mid-tier cheap option.
So saying, "Hey, some customers are price-sensitive. They don't necessarily want to buy the cheapest thing, the cheapest option, which is the longest lead time, but they don't necessarily need the fastest one." So we started offering more options, sort of modulate or cushion that interest in people looking for less expensive options.
Got it.
So we did that. So from a top-line perspective, that's helped us, you know, mitigate some of that risk of people trading down. I think also from a cost perspective, you know, we are focused on getting the profitability. We've talked about getting there in the fourth quarter of this year. So we've been taking out a bunch of fixed costs. We did two reductions in force, one of 6% and one of 4%. We've also, you know, closed some of our real estate and skinny down that footprint. We've been very careful to make sure that those changes don't impact our growth, but also to realign our cost structure, that if things are less expensive, we need to make sure our OpEx tracks along with that.
So that's been important strategy as well as we get to, you know, move, push to profitability.
Yes, I'd just double-click on that a little bit. So I think one thing I do want to level set something, too, like, we are a marketplace company, but we're unique in how we get paid. So there's no fees on our marketplace per se, right? If you think about a regular marketplace, it's a 5% fee, a 10% fee, whatever it is. That's not how we operate. We're a spread model. So as Randy discussed earlier, right, the buyer gets a price to make that part, and then a supplier is gonna get a fee from us to make that part, right? And then we make the difference in between. So why that's important is because it affects, that's our gross profit, right? That's our gross profit margin.
And if you look back, like, three years ago, like, our gross profit margins were in the sort of mid to high teens. Now we're, you know, we're over, you know, over the high 20s and into 30% in some cases. So this is all about our path to profitability. We've put a line in the sand here. It will be Adjusted EBITDA profitable in Q4. How we're gonna get there is, again, we've cut costs, as Randy mentioned earlier. We've gotten rid of all real estate, except of our main headquarters building. We've had two RIFs this year, which was about 10% of our workforce.
In addition to that, again, we've seen gross profit margin increases, and our revenue growth continues to be strong. So when you put all those together, we feel really confident about our ability to get to Adjusted EBITDA positive in Q4, and then obviously eventually positive on a net income basis by the end of next year.
That's super helpful. Yeah, we have a question in the audience.
It's very similar-
Please repeat the question since we don't-
The question is, are margins different for different, just to sort of cut to it, for different technologies, for different industries? Or sometimes we get asked by different sizes of orders as well, let's say production versus, let's say, a prototype order. Our margins are very similar across industries, across. Because, again, you have to go back to the fundamental inefficiency and opaqueness is across the board, and that's the problem that we're solving, whether it's aerospace or it's robotics, whether it's, if it's CNC machining or if it's injection molding, that dynamic is very similar. So it really doesn't get impacted. But we're as our algorithms get smarter and as we expand our networks, we're able to add that incremental value across our entire landscape.
Really, what's driving that increase in gross profit margin is we've got more and more suppliers coming into the marketplace. So as we have more suppliers, they're taking things earlier in the schedule. Let me explain what earlier in the schedule means, right? The AI is dictating what suppliers have the ability to manufacture that part, and it's gonna give it to the person or the person, the entity, that they think it's the best fit for. And then if that entity doesn't take it after a certain amount of time, it goes to the next one, then the next one, then the next one. But as that goes to the next one and the next one and the next one, the price increases. So the more suppliers we have that take it earlier in the schedule, the higher our gross profit margin.
And the more data, just to Jim's point, the more data that we have to know about, A, what's the price that the customer is willing to pay? What's the right price for this? But B, what's the right price to offer those suppliers? So Jim talks about that schedule. We give each of those suppliers a bespoke price, the price that we offer the suppliers based on their interactions with us. So think about like a recommendation engine. You know, you watch two things on Netflix, they recommend a third thing, or if your kids, you know, TikTok videos. It's the same thing with our suppliers.
So the more data we have on our suppliers, the more likely we are to find what is their ideal sweet spot, and that ideal sweet spot can result in a great price for the customer, really healthy margin for Xometry, and profitable order for the supplier. Those are just three things, three things we're trying to triangulate around, and data and the size of our networks are the critical way to get there. That's why our competitive moat is growing, and we're the leading two-sided marketplace. You know, we're the guys. Our competitive moat is growing as we're getting more and more data and growing those networks and as we're expanding internationally. It's a very powerful moat.
I wanna sort of dig into that a little bit more-
Yeah, please.
Because I think it's really fascinating. So how do you get the prices from suppliers? Is that based on the algorithm, you know what all the raw costs are? Or when a supplier comes onto the marketplace, they upload their sort of list pricing, and your algorithm figures it out?
It's a good question. So no, I mean, there, there is no list price, right?
Right.
So it's so-
These are custom-ordered parts.
These are custom-ordered parts. So, you know, when a supplier first joins us. It's a great question. You know, we're gonna assume that they're. We're gonna test things, so we're testing out different price points for them.
In general, we've got sort of where we think the norm is. You know, we're- when we give the price to the customer, we're coming- we call it market-based pricing. We kinda think, here's where the market is, almost like a, a market maker in, in stocks or things. So we kinda think the supplier is there. But as we see what the supplier does, rejecting offers, taking offers, we're able to calibrate in, in almost real time, you know, very quickly calibrate, "Hey, we thought this is where they are, but it turns out they're taking things, you know, at lower prices. It turns out they have, for these kinds of orders, a lower cost structure." And as we train our models, we're able to correct for that and say, "Okay, this guy is really good at that, but he's not good at that.
Super helpful.
Yeah.
And that's why-
And that's the AI-
That, that's the AI.
... the sauce.
That's why data, just more data is really powerful. Again, I go back to, like, it's like a poll. If you poll 10 people, you're gonna have a lot more margin of error than if you polled 100 people. If a supplier watches 2 TV programs, we're gonna have a lot less chance to know what the third should be than if they watch 10 TV programs. So we gotta—we, we want to keep them going into the marketplace.
So I've got-
Before, I just have... Hey, I'll hold your question for one second. We're getting there. Before I forget, Jim, I think you mentioned 2024 net income positive, or, or was that Adjusted EBITDA?
So by the end of this year, we'll be adjusted EBITDA positive, but by the end of the next year, we'll be net income positive.
Okay. Okay.
Yeah. Can you share any data on large customer cohorts in terms of their kind of, whether they're accelerating usage? And how... What does that look like in the context of maybe, you know, the economic cycles?
Yeah. So we don't break it out specifically for large companies, but there's a slide in our investor deck. We have a cohort slide, where we'll show you each of the cohorts by... So the company, I'm the co-founder of the company. We were co-founded in 2013, but we opened our doors from 2014. So we have. I think we show cohort data from 2016 on, because 2014 and 2015 were smaller years. And so we show you each year what that cohort produced in revenue from their initial year through 2022, through last year. We provide that for annual data. And so you can see, like, if you look at that 2016 cohort, and it's all good numbers, but it is, the spend in 2022 is pretty remarkable versus that 2016.
So what we've been successful at is taking those cohorts and growing their spend every year. The other good thing is that, and this just talks to you about the size of the market and how, you know, even though I'm really proud of our growth, we've got lots of growth to go. The cohorts that we've been adding, if you look at that chart, if you look at the first year of revenue from those cohorts, it has grown really strongly, too. So for example, the 2017 year one revenue, cohort year one revenue is much higher than the 2016, and every year that has marched up steadily. So we're not only expanding the existing cohorts, but the new cohorts we're bringing in are actually year one, bringing even more revenue in than the previous cohort. So, you know, this is...
We're trying to get both things going. We want to land and expand the existing cohorts, but we also want to keep that funnel going of bringing in new logos as well. I think we've done a nice job on both.
What about large enterprise, you know, anecdotally?
So there is also a slide deck. In the slide, there's a bunch of case studies, and you can see in there, we show you some of our larger customers. You can see what the growth has been. And it's been... I mean, the CAGRs are crazy, big, you know, good. You know, folks who started, you know, $50,000, they're doing, you know, $6 million. And again, the good news around those companies, and we don't give you, we can't give you the names of them, but you could probably guess at some of them who they are. These guys are spending they're spending every year more on custom manufacturing than probably our entire revenue. So we still have a long way to go with those large customers.
One metric, KPI that we do report out every quarter is the percentage of revenue from existing accounts. And that has been, since we've been public, anywhere from, I think, 94%-96%. So a lot of our, you know, our big customers are growing alongside of us, our accounts, and they're generating a lot of that growth in our revenue, which is, you know, something you want.
That's great. Any additional questions? Yes. Hey, it's good to see you.
Cohort, the cohort of large customers, you know?
Yes, sir.
Do you anticipate that expanding to 50-300?
Well-
That can be a growth in that number.
It is, but I think, you know, look, one mistake we made last year was. It was sort of an embarrassment of riches, and we had so many companies coming in that we diluted our effort of our sales team. And you know, you want to balance that, and you want to you want to focus on, you know, take those precious resources. So I think we just want to be careful. So yes, I can see the expansion, but I also want to make sure we don't fall in the trap where we've diluted our sales effort because concentrating those companies, which are so big and have so much spend, and we're still so early there, you know, already fresh ground there, more profitable for us to focus there. I would just want to be careful.
So we're wrapping up. We have 8 seconds left. I don't know if there's any other... One last one.
So you're giving a quote. Do you always have to honor that quote? I mean, what if you can't find a... What if a manufacturer does, won't produce it for what you're quoting? I mean, do you eat that or-
We would eat it.
How often does that happen?
We don't give that number, but one of the reasons why our gross margins have gone up so much is because we eat it less and less.
Maybe, Randy, you can talk about gross margin guidance for 2024.
Yeah. So, so just to... So Jim alluded to this. In 2019, our gross margin was 18%. In 2022, in Q2 of this year, our gross margin for Marketplace, which is what we've been talking about, overall gross margin was 38%-39%, but our Marketplace gross margin was 31.7%. So from 2019, we went from 18% to last quarter, 31.7%. We have guided that our overall gross margin will be 40%-45%. Today, we're in the high 30s, but in Marketplace in particular, we think that'll be 35%-40%, and we've said we'll get there by the end of next year. So we've been able to minimize those losses. And just five...
You know, we, we've got some baseball fans, I'm a baseball fan, so we track our batting average, which is basically your winners versus your losers. You don't want to be at 300. This is a league where you have to bat, and you're perfect, but that, those are the kind of things, we're measuring those very, very carefully. We also measure our slugging percentage. We've got lots of metrics around, making sure not only that we're making money, but also seeing the magnitude of, of, of big wins versus big losses, too.
What happens if you do a good job, and I find a great manufacturer, and he does a great job at a great price, and now I know him, so for my next job, I'm gonna go to Red?
You mean being disintermediated?
Yeah. So I, I go back to, you know, there's examples in the retail world, whether it's Amazon, if you're buying 3P, and, you know, even they solicit you directly, or you've got an Uber, and the Uber driver says to you, "Hey, on your return trip, because..." The customer doesn't, isn't coming to Xometry because they want another supplier. The customer is coming because they don't want to deal with this, this journey anymore. They don't want to have to depend on these small manufacturers, the risk, the lack of customer service, the... All these things that are inherent. They're never sure they're gonna get the best price. With Xometry, they can lock that in and get that benefit. So that's what's most powerful for the customer. It's not just about finding yet another supplier.
With that, I think we're in overtime. Randy, Jim, thank you very much for the time today.
Thank you for having us.
Thank you for having us.
It's been insightful.
Thanks, everyone. Thank you.
Thank you.