Morning, everybody. Thank you for joining us today at Carvana's first-ever Analyst Day. We're very excited to have all of you here today with us, and we're looking forward to sharing our vision with you. Please note that today's event will be simultaneously webcast on the IR section of our company's website. The slides from today's event will be posted on the IR site following the conclusion of today's event. We have a lot of information to cover today, so we ask you kindly reserve all of your questions until after the presentation is concluded. Joining me today are Ernie Garcia, Chief Executive Officer, Mark Jenkins, Chief Financial Officer, and Dan Gill, Chief Product Officer.
Before we start, I'd like to remind you that the following discussion contains forward-looking statements within the meaning of the federal securities laws, including, but not limited to, Carvana's market opportunities and future financial results that involve risks and uncertainties that may cause actual events to differ materially from those disclosed here today. A detailed discussion of those material factors that could cause actual results to differ from forward-looking statements can be found in the Risk Factors section of Carvana's most recent 10-K and Form 10-Q. The forward-looking statements and risks in this event are based on current expectations as of today, and Carvana assumes no obligation to update or revise them, whether as a result of new developments or otherwise. Now with that said, I'd like to bring up Ernie Garcia. Ernie.
Thank you. Everyone can hear me? We're all excited because we heard the safe harbor, so that's a good place to start. We've got a lot of great information to share today. I'm very excited to get in front of you. We've got an action-packed day, so it's going to be wire-to-wire. I'm going to go ahead and dive right in. Today is all about the future. I think oftentimes the best way to know where you're going is to take a look at where you've been. We're going to start with that. This is all of our numbers when we were going out in our IPO about two years ago. The story we were telling, the company was the exact same. This is where we're headed. This is our trajectory. We're very proud of those numbers.
We've now had seven reporting quarters since then, so let's take a look at what we've been able to accomplish since then. These are our numbers today. I think we're even prouder of these numbers. One thing I'll point out, clearly there's a little bit of scale and variance in this. I think as I click through that, you can barely even tell that the numbers are different, you just see the scale change. That's because we've been continuing to grow at very fast rates. We've been continuing to push up GPU. We've been continuing to push up EBITDA margin in very consistent ways. I think when you look at the company at this level, you ask all the right questions and you arrive at all the right conclusions. The conclusions are, this company must be in an enormous market. They're growing very fast.
They must have a great customer offering. They must have a scalable business. They must be executing. I think the questions are, can they maintain that customer offering? Are other people going to catch up? How big can they be? Where's gross profit per unit going to go over time? At some point, that's likely to slow down. What's EBITDA margin really going to look like? We think those are the right questions to ask about the business, and that's how we run the business. We think in five years from now, when we look back, we're going to be looking at numbers that we don't know exactly what they're going to look like, but we're going to be analyzing the company at this same level.
If you were along for the ride and you look at kind of all those dark blue bars since we went public, you'll probably remember a lot of the questions that we were answering during that period. They weren't always those big-picture questions. We're talking about pinch points and response to a hurricane and an August bump and turn times and residual values and interest rates. I think all those things matter a little bit, but I think sometimes because of just the quarterly cadence of reporting, you can get pulled into the smaller picture, we think this is the scale that really matters. In five years, we're going to look back on the company and these are the types of questions that are going to matter.
Today is about us trying to provide you with our framework for how we think about the answers to those questions. What we're going to do today is first Dan Gill's going to come up. Dan Gill was one of my college roommates, so I know him better than anyone should probably know anybody. He was a world-class gymnast. He was on the U.S. National Team, so you know he's a very talented person. You know he has enormous biceps. That's important when it comes up here. He's great. He knows every detail of technology. He's able to zoom back out and understand all the importance of strategy. He's an incredible manager. He's able to manage a great product pipeline.
Really excited to get him in front of you, he's going to talk about what we're building over the next several years from a technology perspective, how we're going to maintain those great customer experiences and make them very scalable. Mark's going to come up. Mark's our CFO, for those of you who don't know him. Mark and I met about 10 years ago. He was doing research for his PhD thesis on the automotive lending market, I was at DriveTime at the time, within about an hour of meeting him, I realized that he was probably the smartest human being I've ever encountered. I know the questions in a lot of your minds out there. You're thinking to yourself, "Ernie's encountered me." I want to make sure I clarify that Mark is definitely smarter than each of you.
That's also important to me to get that out. He's going to come up and talk about our long-term financial model. He's going to talk about where we think we can take our total gross profit. He's going to talk about where we think we can take our SG&A, and ultimately, where EBITDA margin goes over time. I'm going to come back up, and I'm going to talk about how big we think we can be and why we think we can be that big. Before we start, I want to tell you what I at least hope the takeaways are from today. First, we always start with the customer. We get this question all the time, "Why'd you decide to sell cars online?" The answer is, we didn't decide to sell cars online.
What we did is we said, "Why are customers not getting experiences they're excited about in automotive retail? What do customers need?" We think the customers need the car they're looking for. They need a fair price. They need a great and simple experience. They need confidence that the car they get is going to be a high-quality car. We said, "How do we build that solution for customers using the latest technology?" That led us to many parts of our business being online and all of our business being supported by technology. We started with the customer, and that's what we do with every problem, big and small. We start with the customer and what they need. Number 2, we have great people.
I don't think you're going to find a lot of companies that take an alternative view there, I'm not sure how much information you're supposed to take from me making this claim, I do think that there's at least some objective evidence this is true. I think the charts that we looked at a couple of pages ago are very unique charts. I'm not aware of any other company at our scale growing at the same rate. I'm not sure there's even that many historical examples. I think we've been able to achieve a lot in a short period of time, and that's only possible with great people. I really do believe, people talk about network effects as kind of a buzzword these days. I think the biggest network effect there is is great people. Very talented, curious, interested people that are motivated every single day.
They want to come work with people that are similar. When you get a critical mass of that, it feeds back, it's an exciting place to be, your productivity goes way up, and you can move faster than those companies could ever hope to. Next, we're an ambitious company. Everybody has to decide when they wake up every morning, how hard they're going to push it that day. A lifetime of 80% compounds to a solid place, maybe not a great place. We're a company that wakes up every day with an itch that we can't scratch, excited about what we can accomplish and also disappointed in what we did the day before, we're going to continue to be that way, that's pushed us further faster than I think we otherwise would've been pushed.
Lastly, we get the opportunity to build an enormous and very profitable business. We're going to dive into that in all the other presentations, I won't build that up too much. With that, I will hand it off to presenter extraordinaire with enormous biceps and a big vocabulary, Dan Gill.
Thanks. All right. Thank you all so much for making the trek. We don't get a whole lot of Wall Street in Phoenix, so we appreciate you all being here and are really, really excited to get to share a little bit of a view sort of behind the scenes. We've been working incredibly hard. You're going to get to see a lot more depth around the technology that we have in place today, as well as some views into the future that we're really excited to share. I'm not going to stay high level for too long, but I do want to start at sort of an extremely high level because I think the philosophy that we use strategically to make the technology decisions that we make is enormously important.
I want to start there for just a couple of minutes, and then we're going to get into sort of examples and demonstrations and things like that. I want to use the broad frame of how challenging it is to stand out in a competitive retail environment. Right? To oversimplify, there are really three major vectors that you can try and stand out. You could be an experience brand, and that calls to mind things like Nordstrom. Right? Luxury retail. Generally, being an experience brand comes with an enormous amount of expense, but you're trying to craft an experience that's differentiated from your competitors, and a lot of brands have been very successful competing in that way.
Competing on selection was sort of the domain of department stores and big box retail for a long time, and obviously became much more the domain of e-commerce over the years when you're not limited by physical shelf space, but instead you get virtual shelf space that can be sort of, for all intents and purposes, infinite. Then value. If you want to be able to sell the same product or a substitutable product at a lower price, you have to have some sort of structural advantage, right? One of the favorites that this brings to mind for me would be Trader Joe's. Trader Joe's went straight to suppliers, white label those brands, limited on selection, limited on marketing, distribution of customers through word of mouth, and therefore can offer a great product at a lower price because they've innovated throughout their entire structure to do that.
Competing in these vectors, in automotive retail is really, really difficult historically. The fragmentation of our market makes it hard to achieve scale in so many ways, but it's also that most of the dealers are getting their vehicles in the exact same way. Most of the dealers are merchandising and fulfilling those sales in the exact same way. The labor and the cost structure of the salesperson, the sales manager, the F&I manager, all of the headcount required to do the work of actually bringing a deal to fruition is largely consistent. Same third-party software and systems. There's not a lot of custom development in an individual dealership, but there are some enormous software providers that provide very consistent software experiences to all of the dealers.
You're reliant, generally speaking, on third-party lending, which not only has sort of customer experience costs, but limits the ability to participate in the gross profit created through the financing. If you're a franchise dealer, you have to comply with all the OEM hurdles. Whether that's the marble that you need in your lobby, or any number of other regulations and challenges, it imposes a lot of cost on the dealer body. This generally plays out with very similar experience, selection, and value in the dealership world. At Carvana, we try to build in a large number of very structural differentiators inside of our model. As many of you, I hope, understand well, that starts with centralization of our inventory all connected through nationwide logistics, where we can get a much broader selection available to our customer base.
That has an enormous differentiation relative to merchandising dealership to dealership. We're real estate light, particularly in terms of our retail footprint. From an experience perspective, massively differentiated in that we're bringing the car to you, free delivery to your driveway, or coming and picking up the car that we're buying from you. Integrated lending. We are a lender, and have woven that into every aspect of our customer experience, with huge benefit from both an experience and a gross profit perspective. Self-service. We're open 24 hours a day, seven days a week. We have an enormous amount of technology that we've developed in-house that we're going to get a little bit more insight into today. Lastly, there is so much data that we get to collect by having a digital experience.
Every page view, every dwell time, every click, every bit of everything that happens, we get to ingest that and use that to feed an enormous number of proprietary machine learning models to inform every aspect of our business. All of that built on top of this foundation of a complete and total customer obsession. Everything that we do is focused on, again, as Ernie mentioned, it's not about how can we create e-commerce for cars, it's about saying, what if automotive retail was done with the best interest of the consumer in mind at every turn? What if we just did what people want? Right? That's woven into every decision that we make throughout.
We think it's actually possible for Carvana to have the best experience, the best selection, and the best value simultaneously, which is not only extraordinarily rare in automotive retail, but is rare in retail, period. I get the amazing, fun, challenging, stress-inducing job of trying to unlock and scale all of those capabilities through technology. Let's start digging into some ways that we're doing that. I ran across this quotation recently and really, really enjoyed it. Clayton Christensen coined the term disruption and sustainable innovation, those kinds of things, and has a new book that just came out, and he says, "Creating the right experiences and then integrating around them to solve a job is critical for competitive advantage.
That's because while competitors may attempt to copy products, it's difficult for them to copy experiences that are well integrated into your company's processes." A paraphrase of that that my teams are exhausted of hearing from me is that a chain is only as strong as its weakest link, and we have a lot of links in our chain, right? You're stringing together vehicle selection and financing and trade-in and e-contracting and delivery, all of those pieces, and you only have to get one part wrong for the experience to really fall short of expectation, right? If you want to offer home delivery, but you're FedExing contracts back and forth, the customer experience is, "Eh, it was okay," right? That's not good enough to fuel word-of-mouth growth and the kind of adoption that we're seeing.
We think that this is a really important concept. How do we weave that into some of our technology? For us, the best experience in our user interfaces means intuitive, transparent, simple, fast, and self-service. Let's roll that video. What we're seeing here is our soft credit pull flow. This is a customer coming to the site, entering a very limited amount of information, no hit to their credit, and immediately getting to-the-penny-accurate financing for every single vehicle in our inventory in under 30 seconds. Right? What we think here is that customers don't want financing. No one goes out to get auto financing. They go out to buy a car, but you have to get financing in order to get that car, right?
What we think is also very true is that no one ever wanted an auto loan from Carvana, right? No one ever sought automotive financing from Carvana. What we do see from even our earliest days is north of 60% of people choosing to finance through Carvana, and now more than 70% of people choosing to finance through Carvana because it's so seamlessly integrated into the entirety of the customer experience. We think that when you make those experiences seamless and transparent and self-service, you get much, much higher adoption than you would otherwise. Let's kick back to the slides. It's not just about being utilitarian. We also want to make sure that we have some surprise and delight, right? We try and bring some whimsy into our brand. We have an amazing creative team.
One of our co-founders, Ryan Keeton, and our amazing head of creative, Paul Keister, really bring a lot of flavor to the Carvana brand. If we can roll this forward, this is another example where now, obviously, we want to make our trade-in experience seamless, self-service, transparent, simple, as articulated. We're asking for very limited information, no photography, no physical inspection, get an instantaneous value using machine learning, all from the comfort of your home. As we go through this very, very limited number of questions, you see we're introducing the character of the Cardian Angel, right? The Cardian Angel is your guide to walk you through the whole concept and the trade-in experience. Yes, we're giving you all the data. We're giving you comfort that the value is real.
What's happening behind the scenes is we're also in real-time rendering a video frame by frame about your car. This video that just got created is about the 2014 Nissan Murano. In order to do that, we have to create every single frame of this video in real-time. It's some of these innovations where we like to bring technology to bear to try and craft these experiences that just give you a sense of wonder and real pleasure in going through the experience with Carvana. Let's keep going. Back to the slides. There we go. Another piece is enabling exceptional service. We have a phrase that we use a lot internally, which is that we sell cars, but we're not car salesmen. With that, we're scaling a lot, right? 82 markets.
There are 82 places around the country right now where people are waking up and going to work for Carvana. We need to give them that sense of customer obsession and confidence to give the kinds of customer experiences that have fueled our growth to date. What we don't want to weigh them down with is the nuanced rules and regulations state by state of what documents are needed and what disclosures, and where do I need a wet signature, and all of that is different for every car-customer combination, depending on which IC they bought from, whether we have a dealer license, all of these different things.
What we needed to do, for us, was we decided we needed to build a completely custom tablet application so that when our market operations advocates go out to perform a delivery, they have total confidence and can focus only on providing exceptional customer experience. It's a beautiful application woven in with navigation of where they go. Simultaneously, the GPS behind the mapping is fueling our mobile app so that the customer is getting an experience of watching the hauler come to their door. Once they're there, we know exactly what documents, what photos they need to collect. This not only creates a really modern experience from a customer perspective but allows us to hire in a very different way. We get to not hire from automotive, we get to hire from luxury retail, we get to hire from hospitality.
We get to hire from these areas where these people get to focus on customer experience and nothing else. It's our responsibility as a technology organization to ensure that it's easy for them and that they're confident in giving the customer experiences that we want them to be able to deliver. Lastly in experience, I think we'd be remiss if we didn't touch on our vehicle experience. This is, I think, what we've been known for from the beginning. We pioneered sort of the 360-degree photography. We can roll that. This is the experience that you would see on the website today, and it's been pretty well unchanged for the last five years. We were sort of the first to do it, and we're really proud of it. What we're trying to do here is remove the need for a test drive.
If you think about what a test drive is, a test drive is an information exchange where I'm trying to get enough confidence to pull the trigger on the purchase. How can we make that information exchange happen online? How can we give you enough confidence in a self-service way that doesn't require a salesperson and a trip to the dealership, et cetera? We have to be able to pass forward a lot of information to you. This spinner started to show its age, and we love it very dearly. I'm very excited. I'm bordering a little giddy because I get to click next, and I know what's going to happen and you don't. We get to show today what's coming. We made an acquisition of Car360, and that is an incredibly talented team.
We recognize the skills that that team had, and how much vision and talent and the underlying software architecture that they had in place. That's unlocking a brand-new vehicle spinner experience that you're going to start to see rolling out from Carvana here in the coming weeks and months. Click play on the video here. Now we get this beautifully immersive experience where we've had to rethink the lighting, the cameras. You're getting more frames, higher resolution. We're getting a three-dimensional experience around the car and a spatial awareness around the car. You can see as we're touring the features, we're rotating the vehicle and showing you exactly what you're looking at in the context of the broader vehicle. It's just an enormous step forward from a user experience perspective, and we think certainly best in class and something we're really excited to roll out.
Because this did require completely rebuilding aspects of our photo studios, it's requiring a complete rebuild of our photo processing architecture, putting in more computer vision, machine learning, and so many of the other image processing technologies that are enabled for us because we control that full stack. You're going to start to see this roll out over time because we have to change over the photo studios. What's going to come first is sort of a retrofitted experience that takes the 16 images that we've collected for all 14,000 cars in our inventory and creates an experience that's along these lines. We're going to start collecting much larger volumes of images, this whole new experience, and you'll start to see that roll out through the beginning of 2019. We couldn't be more excited about it. It's really beautiful.
I'm so proud of the work from the creative team and from the technology teams behind this. We can't wait to put it out. I'm going to transition now from experience to selection, and how we can leverage technology to really enable the benefits of our expanded selection. Sort of e-commerce 101, and thinking about us in contrast to sort of a physical dealership. If you're a physical dealership, you're marketing in your local area, and if you have lots of physical dealerships, you really need to keep sort of in stock the most popular items in a very redundant way. You're going to need to repeat your inventory over and over again in each geography.
Whereas if you can centralize that inventory, whereas for us, all of our inventory is currently in 4 locations, we can carry a much more diverse selection, and all that's linked together through logistics. That's been a huge benefit for us, but we're not even close to the potential. I want to illustrate that really quickly. This was a survey that we did where people who created accounts on carvana.com and didn't buy. We went and asked them, "Why didn't you buy?" These are all people that went ahead and purchased a car at a dealership afterwards. We said, "What was the reason you purchased a car at a dealership and not at Carvana?" That giant bar sticking out is, "I did not find the car I was looking for." Not, "I didn't understand what Carvana does.
I wasn't happy with my financing terms. I wasn't happy with my trade-in value. I'm not comfortable buying a car online." It's simply, "I wanted a Subaru Forester XT Touring, and you didn't have one." The important thing here is that's a really big bar, and if we think about the percentage of people who do buy from us, the size of that bar is much larger than the percentage of people who do buy from us. What that suggests is that as we expand our inventory, we're going to get a lot more conversion, even with the demand that we're already generating today. Let's get a little bit more into how do we unlock all of that by having a diverse selection.
14,000+ cars to choose from, that's great, but having, again, a huge selection doesn't really matter if it's not available to your customers in a timely and predictable way. I love this little illustration. That's a tiny little element on our new vehicle detail page that you'll see rolling out. That little clock, finish your order within one hour, 20 minutes and 24 seconds, and get your car delivered as soon as the day after tomorrow. In order for that to be true and for that to be an accurate statement, what's actually happening is we're making an API call into the entirety of our logistics network. If you're not logged in, we use your IP address to know where you are. We know where you are. We know where the car is. We know all the transport legs that exist between those two points.
We know the capacity of those legs. We know the hierarchy of importance of every car on every leg to know that car doesn't actually need to be there for five days, if we bump that, we can put this one on. All of that is fully automated, all of that decisioning is happening in real-time. We have visibility into our underwriting flows and all the work necessary to make that deal all the way to fruition. Then we have visibility into the last mile logistics of what labor do we have available, what delivery slots do we have available in order to enable the customer to get that car. Done some same-day deliveries here in Phoenix when you're so close to the Tucson IC. Third-party logistics is out there. There is no time certainty in speed in third-party logistics.
Not only do we see that as a cost advantage for ourselves, but it's the only way to offer the customer experience is to build it in-house and operate it ourselves. The other part is that if you want to get the advantages of that diverse selection is you need to make it low risk for the customer, right? That's always the number one objection. How can I buy a car without test driving it first? Well, the reality is we're going to come and get it from you for any reason at all. You're protected by a seven-day test own, we'll come pick that car up, no penalties, no restocking fees, anything of the sort. In order for that to be true, we have to build in the systems and processes where we can know, okay, well, there's a delivery going out. It doesn't have a trade-in.
That hauler's going to come back empty. We can link that with a return, go pick up that car, bring it back. Does that car need a light refresh, or does it need a full recon? Where does it need to go? How do we make this low cost and operationally efficient for us to take that inventory back and get it efficiently back on the website? All of that is unlocked through the technology that we put under the hood and all the automated decisioning that powers that. Last, the challenge of a really big inventory, 14,331 cars on the site when I did this maybe yesterday, is that we want to help you find the right car. I'm not going to go super deep here, but I'll tell you that helping you find the right car quickly is something that starts with good data.
Vehicle data, we've talked about this a little bit before, means being able to take a used car VIN and just take me at my word for the moment, we'll certainly go deeper another time. A VIN to all the details of a vehicle is not a straight line path. Being able to accurately describe a vehicle all the way down to the packages and the options and all the things that you might add to that car is really, really difficult, then helping consumers to navigate that is even harder. I'm going to give two illustrative examples, certainly not exhaustive of the way we think about this problem, but here's part of our search experience. We'll use one example first, which is Chris Coleman, the head of our pre-purchase experience, calls this the spear fisherman. We'll click into the video here.
What's being typed here is black Mercedes C-Class, this is the person who knows exactly what they want. They know that the multimedia pack is the package that has all the cool technology. You'll notice that we didn't have to type black Mercedes-Benz C-Class multimedia package. You'll notice we're able to extract out black as a filter, Mercedes, turn that into Mercedes-Benz as a filter, C-Class, turn that into a filter, search for multimedia pack, immediately pull up 12 or 11 black Mercedes C-Classes with a multimedia pack, kick those into a side-by-side compare. Now I'm able to find out what are the exact differences? What are the trade-offs I'm getting for the $4,000 difference between those cars, right?
We want to help you make those decisions really quickly. I think you'll find if you try complex queries like that on other sites, it's not supported. There's a lot of magic that's happening behind the scenes here. Can we go to the next one? Another example, if we go to the next video, is what we call the other use case. Maybe, Gordon, can we start that one again for one second? We'll take another use case here, which is more of a discovery use case. This is me. We always joke a lot, but it's just true. I am not a car guy, right? No experience in the industry, don't know much about cars. When I had my second child, what mattered to me was a bigger car, right?
In this experience, it's helping me click and find SUV, third-row seat, leather interior, got to have the remote start in Phoenix. Those summers are pretty savage. I'm going to go look. Oh, blind spot sensors. I'm going to want those for sure. What's actually happening here is that every click I make is being processed in real-time. We use our machine learning models to start suggesting the next filters that are going to help you narrow your result set the fastest and give me confidence that I'm narrowing to the exact car that I want to find.
This is not comprehensive of the capabilities by any stretch of the imagination, the point we're trying to make here is that when you have incredibly granular and well-structured vehicle data, real-time processing of every click, you have very powerful capabilities in terms of unlocking really, really cool user experiences for our customers. Let's go back to the slides. Next, we're going to cover value. I'm not going to go through this line by line. This will be posted, you can take a look at it yourself. The concept here is a little bit like we mentioned earlier with you can't offer the same product at a lower price without some sort of structural innovation, right? From our perspective, we think there are two primary facets to that.
You'll see at the top from a vehicle journey perspective of sort of us versus a traditional dealership, we've centralized an enormous amount of our labor and empowered all of that through automation and technology. That has huge implications in terms of our long-term variable cost structure. On the customer journey, when you look at the amount of work and the number of people involved in consummating a transaction versus how much of ours is self-service and technology-enabled, that's another enormous structural advantage that we have that we think will unlock a significantly lower variable cost structure as we continue to scale. The other piece is that when you fully integrate financing and ancillary products, there are significant gross profit opportunities that come with that. Now going into sort of the examples of how we're using technology to scale best value.
This part I'm also very, very excited about. I get to welcome up Tom Taira. One of the reasons I'm so excited about Tom, if you haven't met Tom or don't know Tom was a co-founder and served as the Chief Product Officer for TrueCar for a decade. Tom is also probably the biggest man crush that Ernie and I shared for a long time. There were several years ago when we were in a meeting with TrueCar, and definitely walked out and said, "Holy crap, is there any way we ever get" Brings so much experience to bear and an incredible passion and knowledge base in our industry, and now as part of our executive leadership. Didn't think it was fair four weeks onto the job to make Tom speak about the whole thing, although he could very capably.
One of the ways that we brought Tom into the fold was through the acquisition of a company that he was working on, Propel. I want to give Tom an opportunity to provide some context there, and we'll get into what that is.
Yeah, thanks. Just as a start, first time I met these guys was five years ago. I think it was June or July of 2013. I remember walking away from that, and I think there was a lot of crazy vision going on. They were in one market, Atlanta. I think Ernie may have said they were selling dozens of cars. I think it may have been like threes to fours to sixes at that point in time. I know I walked away with one thing, which was, Jesus, these guys are approaching this problem the right way, which is how do you actually change the experience, the experience that I've spent literally the last two decades trying to figure out. How do you actually solve it? It's at the transaction level.
What they were doing was not just trying to create a new facet, like a delivery service or a new way to present a vehicle, but it was this holistic view on how you could actually change the experience. It really started with a platform, a transaction platform that enabled you to do the entire process online. That's what got me excited. Literally fast-forward, you look at what Dan has been being presented today or presenting today, and you see a platform that enables a multitude of new car buying and selling experiences, which we're not going to get into, but I really believe this is a platform that allows people like entrepreneurs and innovators and inventors to be able to come in with ideas and be able to build something that even exceeds what we have today.
I'll talk now a little bit about Propel. We started off trying to solve a problem, and the problem was when you're actually going out as a consumer trying to buy a complex product like a car, you have a ton of questions that aren't just simple questions like, "Is it available?" "Are you open?" and so forth. You want to ask that retailer questions around does it have a specific thing on it, right? What is the monthly payment? How do I do my trade-in? What's my financing? There's a series of questions that you need to get answered in order for you to be comfortable. What Propel did is we built an AI platform that creates automated or enables automated conversations between the retailer replacing the salesperson and the actual consumer itself. Are we running this?
Let's run this.
Yep.
What you'll see is we have the ability to start answering questions that are not just simply the basic ones, but things that are around the actual cars itself, the process itself. What you'll notice over time is that we're going to start with this idea of the introduction of these AI-powered machine or bot or virtual assistant, whatever you want to call it. We're going to start with nurturing initial customers. We're actually addressing the questions up front. Then over time, we will be building this to be able to answer and address questions across the entire life cycle of purchasing a vehicle all the way to ownership. What that does is it allows consumers who have a need for immediate, accurate, and simple answers to very complex questions. It allows them to get in self-service and get those questions answered immediately.
What that also does, it alleviates some of the burden and stress that you have on advocates that have to figure out how to answer these questions really fast when a machine can do it much faster, and we can enable them to free up time, so when the advocates get on the phone with the consumer, they are having much more fruitful conversations with the idea of being able to scale them to another level going forward.
The approach that we are taking with Propel is not, hey, how do we get our variable cost down and reduce the number of advocates that we need and all those kinds of things. It is people want instant gratification. They want 24/7 service, and we want to be able to provide that. We are incredibly excited about Tom and the team. For those of you here, you are going to get to meet Mikhail, one of the co-founders of Propel, and see a lot more detail on Propel. That will be exciting as well. I am going to skip through this one. We have talked a lot about vertically integrated financing and ancillary products. The video that I would show is someone going from start purchase to signing contracts with their finger in 60 seconds. It is just completely different, right?
We still see ourselves as very early in terms of how simple and intuitive and powerful that we can make that experience. I really want to get Mark and Ernie up here, so I am going to skip through that. One other part of, again, structural innovation that allows us to have the best value in our industry is also scalable customer acquisition. We spoke about our incredible creative team and I think the brand that we are building, but we also have the ability to provide online-to-online capabilities that are not possible for other dealers, right? Here is an example of I am on LendingTree, I go through their enormous flow to find out when banks compete, you win, and I want to get my terms from LendingTree.
What you see here is we have made a call that takes a matter of two, three seconds into our API to get my loan term, my APR, and my monthly payment to the penny accurate. Below that, and you see on the Carvana line item there, the call to action is Continue. Below that is another lender. The call to action is Email Lender. It is not very satisfying. The asynchronous nature of, okay, well, we think you can get this, but then go find the car and then come back and tell us what car, and then that is going to change the loan terms and the steps. In ours, you click the button. They are passing that through a secure API, and the only field of information we need from you is a password. That is it.
Everything else has been collected on the LendingTree side. Now I'm shopping on Carvana with to-the-penny-accurate terms for 14,000 cars instantaneously. We do this with Capital One when you get pre-approved. We do this with Bank of America now. We're enabling this with what we call Pre-Fi, right? There's refinance, we have pre-finance, where you can get pre-approved, pass those terms through secure APIs, and apply those terms to our inventory on Carvana. This online-to-online handoff, so many brands are spending all their time and energy getting people to download their apps and get online customer acquisition because it's attributable and it's scalable, yet the handoff is then print out the thing and go to the dealer, right? Or we've injected it to the DMS and you can resume maybe at the dealer, but there's going to be a lot of questions and back and forth.
For us, complete online-to-online seamless handoff. We think that technology enablement's going to be really important. The sum is that we're able to share some of our structural advantages with customers in the form of lower prices. We're going to get to why we think we can still reach our profit targets at the same time. We touched on really a small fraction of what we're actually doing here day to day. We're very fortunate about five months ago to have a new VP of Engineering join us from eBay, an incredibly experienced engineering executive, Ram Akella.
He came in and said, "The mission of our organization is to craft the platform of choice for car buying and selling." I think there's a really important term in there, which is platform. We're building out all of these capabilities as services that enable us to iterate faster and faster and faster. The philosophy that we take is that I see my job as building a machine that can parallel process all the needs of the business. We talked about a handful of items today, but there are actually 32 technology teams in place today that are parallel processing. We have a team that's fully dedicated to ingesting vehicle data and normalizing data on every vehicle available in the wholesale markets every day. Pricing optimization team.
We have teams that are focused on custom software for our inspection centers and the visibility across our entire supply chain. We have teams focused on software tools and custom systems for customer relationship management. We have purpose-built teams that have defined mission, defined systems ownership, isolated KPIs, and the resources that they need to execute autonomously against those missions. We're going to keep adding those so that we can parallel process all the needs that we have going forward. Inside of that organization, we have 11 former founder CEOs. The last thing I'll leave you with is that the truth of the matter is we've got some really good stuff that we're really proud of. We're also pretty much all embarrassed by the state of things. We see warts and bugs and problems and weaknesses everywhere.
We think we're pretty good in a number of ways, and we also see nothing but opportunity, in terms of the long-term roadmaps that we have. Anybody who's ever spent an hour with Ernie knows that he can craft three more years of work in an hour. There is no shortage of things for us to do in the future. Thank you so much for the time. We're going to get the opportunity to meet some of those leaders later, go a little bit deeper, and see some of the technology that we've been building and we're excited to share. I want to get you over to our CFO, Mark Jenkins.
All right. Great. Okay. Thanks, everyone, for coming today. I'm really excited to talk to you about our long-term financial goals. I am going to cover an awful lot of material in this section, and I'm going to do it very quickly. I just want everybody to know the slides are going to be posted online. You don't have to take notes that frantically as we're working through everything. I wanted to start this section with a question that investors sometimes ask us. That question is: Does Carvana want to be the largest automotive retailer or the most profitable? Our answer to that question is both. Why is that our answer? We deeply believe that e-commerce enables new opportunities in automotive retail.
In particular, the scalability of the online sales model, we believe will lead to larger industry players than we've historically seen in automotive retail and more profitable players than we've historically seen due to a long-term lower cost structure. I'm going to spend this section talking through our long-term financial goals as it relates to profitability. Ernie will then come up and talk about our long-term goals as it relates to growth. All right. Like I said, this is a section where we're going to cover a lot of material. I'm going to start by talking about our long-term goals for total GPU, and in doing so, I'm going to do a couple of deep dives into some of our larger opportunities. That is buying more cars from customers and monetizing our finance platform.
Also briefly mention other GPU opportunities and sum up with some total GPU long-term goals. I'll next talk about our long-term cost structure goals, which is where we believe we have a significant advantage versus the traditional brick-and-mortar model, and I'll summarize all that with our long-term EBITDA margin goals. Okay. I'm going to start by doing a little bit of a deep dive into buying cars from customers. This is a place where we think we have a significantly differentiated experience relative to the dealership model, and an area where we see a lot of opportunity for growth over time. What's differentiated around our experience? We make it fast, simple, and convenient for customers to sell their cars to Carvana. How does that work?
You can go, if you want to sell your car, to our website, fill out a short credit application form, or not credit application, but a trade-in form. We then kick off our algorithms to value your car, present that offer to you, communicate that offer to you, in creative ways like Dan mentioned. We'll come to your house using our existing last mile delivery network to pick up the car. It's incredibly convenient and it's a profitable business for us today. There's two ways to give a little bit of background on how this process of buying cars from customers works. We basically have two acquisition channels and two disposition channels. The two acquisition channels are, first, we buy cars from customers as trade-ins when they're buying a retail car from us.
Secondly, we have our sell to Carvana channel, which includes customers that want to sell us a car but don't necessarily want to buy a car from us at that particular time. Both of those channels then flow into our two disposition channels, which is selling cars on our website, retail, and selling cars in wholesale auction markets, which generates wholesale sales. Both of those are distinct total GPU opportunities. In the retail case, there's a GPU opportunity because cars sourced from customers are more profitable than cars sourced at auctions. In the wholesale case, there's a profit opportunity because we make a spread on the cars that we buy from customers and sell in the wholesale market. I'm going to start with our goals for the retail disposition profit opportunity.
This is an area that we've gotten really excited about recently about what our business can bring to bear as it relates to buying cars from customers. We've talked a lot about the traction that we've seen in the last couple of quarters. Of the cars we sold in Q3 2018, 16% of them were originally sourced from customers. That's up from 11% in the prior quarter and 6% approximately in Q1. That still leaves us significantly behind the used car industry leader in terms of customer sourcing. We've set our long-term goals on reaching industry leadership levels. I mentioned earlier that cars that you source from customers are more profitable than cars that you source at auction. The typical incremental profitability out there in the industry is $500-$1,000 of incremental profit. We've set our long-term goals at that incremental profit as well.
When you kind of multiply those through in terms of the total contribution to GPU, it's a $200-$500 total opportunity for GPU contribution relative to 100% auction sourcing by buying more cars from customers, and this is what we will target over time. I also want to touch upon the opportunity in selling more cars wholesale that we take in from customers. Here we've been relatively stable between 15%-17% of units sold wholesale relative to units that we've sold retail. That also places us significantly below the current industry leader who wholesales approximately 55%-65% as many cars as they retail. We believe in the opportunity that we have in front of us and are setting our long-term goals at matching this industry leader.
Similar to the incremental profit generated by retailing cars that you source from customers, selling cars in the wholesale market that you source from customers is a source of profitability. In the industry, there's an opportunity to make, we believe, $500-$1,000 per unit wholesaled and are setting our long-term goals at this target range. When you multiply those through, we believe there's a $250-$650 total gross profit opportunity in buying more cars from customers and selling them in the wholesale market. I've talked a lot about our opportunity to increase the size of our business of buying cars from customers and selling them either retail or wholesale. Where do we see the growth levers in driving growth in that future business? A really big opportunity that we see is increasing customer awareness. This business for us is a nascent business.
We're in an incredibly early stage. As I mentioned, we've just started really focusing on it in the last couple quarters and are seeing significant traction. I think a great way to quantify how early we are in this business is company life to date, we've spent around $1 million advertising VFP or sell to Carvana or trade-in related advertising messaging. We've done nothing. We're at the very early stage. In addition to a big opportunity to raise customer awareness of our great experience of buying cars from customers, we have a number of internal process and technology developments that we think will also drive growth, including optimizing our fulfillment, vehicle data and bidding, as well as website testing and our disposition processes. That's our opportunity in buying cars from customers.
I next want to spend a little bit of time talking about our long-term financing monetization opportunities. This is an area that I think we are incredibly excited about and view as a significant differentiator relative to what is out there in the industry today. I'm going to spend a little extra time actually talking about this opportunity because of how big we believe it is. In doing that, I'm going to start by taking a little bit of a step back and talking about the structure of the auto finance market, then that will provide some context for what we're doing differently and the opportunities that lie therein. Taking a step back, there are really three primary players in the automotive finance market. There are dealers, there are lenders, and there are investors. Dealers and lenders do a lot of work.
The dealer acquire the customer and help arrange the loan. Lenders do the most work. They underwrite the loan, they credit score, they price, they verify, in many cases, service the loan, et cetera. As a result of doing the most work, lenders also earn the most profit on a given loan. Investors don't do much work but they do hold credit risk, so will earn a risk-adjusted return for holding that credit risk, much like they'll earn in other fixed income markets. These three primary players interact in a number of different ways in the industry. By far, the most common way for them to interact is through what's called indirect lending model, whereby a dealer partners with multiple lenders who compete for the customer's loans that are being originated by that dealer. These are all outside parties that the dealer partners with.
These lenders then, in turn, will partner with investors who hold the ultimate credit risk. As I mentioned before, the dealer earns some profits, the lenders earn some profits, typically the largest share, and the investors earn a risk-adjusted fixed income return. A hybrid model, which is utilized by CarMax, who many of you are familiar with, replaces some of the outside lenders with an in-house lending segment. Basically everything else that I described plays out similarly. For some customers, there's an in-house lender who then pairs with investors. For other customers, there are outside lenders who then pair with investors. Our model is different. Our model is a fully vertically integrated retail and financing platform, which we believe has significant advantages over the other models that I've described, in particular over the indirect lending model. What are the nature of those advantages?
The first, and really debatably the most important, is it enables a seamless customer experience, okay? By controlling both the retail and the financing platform and being able to seamlessly integrate that, we can give the customers an incredible experience that involves filling out a short credit application form, receiving credit terms, a wide range of credit terms nearly instantaneously, and receiving those terms for all cars on the website. 15,000 sets of financing terms nearly instantaneously. That's almost impossible to do in a model where you have multiple lenders dealing with small local dealers. In addition to customer experience benefits, there's also a number of financial benefits that are significant in executing a paired retail and financing model. What are those benefits? Well, first, by vertically integrating, you cut some costs out of the system.
In particular, relative to indirect lenders, you cut out dealer relationship or management costs because you're vertically integrated under one roof rather than having to partner with many lenders to many dealers. Second, that fact that you're vertically integrated rather than partnering many to many allows you to automate more of the process, which also decreases cost in the long run. In addition to lower costs, the vertically integrated model, we believe, produces better loans. The better loans come from a number of sources. One, we can certify vehicle quality. We know exactly what cars the customer is buying. They've gone through our 150-point inspection process, our 100-day warranty. That ensures a very high-quality car. Moreover, we operate in a controlled environment, so that reduces adverse selection that can sometimes come from lenders bidding against each other in the traditional indirect model.
The vertically integrated model typically allows for longer-term commitments than dealers are able to achieve in the indirect model, where loans are approved one by one with no long-term commitment underlying it. In addition to benefits of vertical integration, there's also some significant benefits of the online sales model relative to the brick-and-mortar model. Those come from our lower long-term cost structure, which allows us to provide lower retail sticker prices to the customer, which in turn provides positive feedback into the lending program through two different channels. First, lower retail sticker prices leads to lower monthly payments, everything else equal. Second, lower retail prices leads to lower loan-to-value ratios, everything else being equal. Both of those facts flow into better loan performance. Now, that is exactly what we are seeing today in our data.
This next slide is very complicated, but it's also incredibly powerful. What this slide shows is the relationship between weighted average FICO, which is a typical industry standard measure of credit quality, and expected cumulative net loss, which is a complicated term that basically is a measurement of loan performance. What the black line in this chart shows is a number of securitization market transactions, so basically sales of pools of loans with their weighted average FICO and their expected cumulative net loss. You can see a clear pattern there. Loan pools that are sold in the securitization market with higher weighted average FICOs tend to have better performance, which is reflected by lower position in this chart. The blue line is the Carvana model, which benefits from many of the things that I just described.
What the chart shows is that the loans generated in our model significantly outperform the loans that are sold in the securitization market, which enables greater profits opportunities over time. There's two key ways, or at least two key ways, to generate strong finance gross profit over time. One is to have great performance. The second is to have a low cost of funds. We're doing well on great performance. Today, we see a significant opportunity in low cost of funds. What this chart shows is the current cost of funds that we face when we sell our loans to our finance partners today. This chart here shows in year-to-date 2018. These are measured, by the way, as a spread over benchmarks, which is typically like a two to three-year swap rate.
The bar for Carvana here shows our cost of funds borne in year-to-date 2018, then the other two bars are the cost of funds borne in 2018 by a large retail and finance company and a large indirect finance company that sell pools of loans in the securitization market. I want to note on those, the cost of funds borne by both this large retailer lender and this large indirect finance company include both their cost of debt and a market cost of equity for any component of the loans that they continue to hold after selling them into the securitization market. We see a significant opportunity here, and our long-term goal is to reduce our cost of funds by 175 to 250 basis points over time.
Now, how does that 175 to 250 basis points translate into GPU? It's a little bit of a complicated calculation, but has a very similar feel to understanding the impact of benchmark rates in the fixed income market. Here's an example that gives you an illustrative sense of how changes in cost of funds translate into total GPU. Basically what this example shows is an example loan of $18,000 that's outstanding on average for a weighted average period of 2.3 years. Multiplying those by a 100 basis point change in cost of funds changes loan value by about $400. If cost of funds coming down, that's a $414 improvement in loan value.
Multiplying that by a finance attach rate of 70%-75% leads to a total GPU improvement of approximately $300 per retail unit sold. This is a significant opportunity, as I mentioned, to increase our finance profitability. We also believe the customer experience as well as the economics of our finance platform create, as I mentioned, a significant differentiator and a long-term barrier to entry in this market. We've covered buying cars from customers, and we've covered monetizing finance. I'm going to spend a moment on other GPU opportunities, which we've covered at length in our other materials, and then we'll talk about our total GPU improvement potential. As I mentioned, we've covered these other GPU opportunities at length, have benefited from them greatly over the past couple of years and expect further gains in them going forward.
Those opportunities are reducing average days sale, a place where we've had tremendous success over the past couple of years, improving utilization of our IRC logistics network, increasing attachment of our existing products, including VSC and GAP, and adding new ancillary products to our checkout flow, an area where we've only started to scratch the surface of the opportunity by adding GAP last year. We think there's a significant opportunity to add many more products that customers desire related to their vehicle purchase to that checkout flow over time. Summing all of this up, this slide presents what we see as the potential opportunity to increase total GPU. We see an opportunity to increase GPU by $1,250-$2,550 in the long term. One point that I would make here is this is all without changing our pricing strategy.
This is all from other sources that I've walked through. We see a tremendous opportunity. We've obviously executed very well in marching toward this opportunity at this point, up in Q3, about $1,300 from where we were in 2016. Next up is long-term cost structure goals. I made this point at the beginning, but I think this is a really powerful point. That is, we believe in the long term, the online model has a significant cost structure advantage relative to the brick-and-mortar model for a number of reasons. Dan touched on a lot of those reasons in his segment, but they include customer self-service, automation, centralization, vertical integration, real estate efficiency, and last and importantly, scale.
I'm going to talk through now for each component of our long-term cost structure, which includes compensation and benefits, advertising, logistics and market occupancy, and other SG&A, what our goals are in the long term for those components, all of which are going to be impacted by one or more of these key drivers. I'm going to start with compensation and benefits. In compensation and benefits we've made a lot of progress over the last two years. The chart on the left shows that we've reduced compensation and benefits as a percent of revenue from 10.2% to 6.5% in the last two years, about 370 basis points of improvement. Our long-term goal is 2.25%-2.75% of revenue for compensation and benefits.
That's lower than the typical dealer, who can range from 5% to 5.5% of comp and benefits as a percent of revenue. Where does this advantage come from? I mentioned on the previous slide, how do we see this evolving and playing out over time? In the long run, we expect scale to provide tremendous leverage and tremendous benefits for reducing compensation and benefits. We expect scale to impact two different components of compensation and benefits. The first is fulfillment and customer service, which we expect to be about four-fifths of comp and benefits as a percent of revenue in the long term. Scale benefits fulfillment and customer service comp and benefits because it increases utilization of the logistics network, increases utilization out of the markets, and then there's also an opportunity to significantly automate many of the manual tasks that we're still performing today.
Scale is the key driver of reduction in technology and corporate comp and benefits per unit. I think that one's fairly intuitive. These employees are relatively fixed relative to revenue growth, we expect them to fall as a percent of revenue over time. The next component of SG&A that I'm going to discuss is advertising. I'm not going to spend a lot of time on this one because we covered it at length. We see advertising leverage at the market level. We've shown significant reductions in advertising expense per unit and as a percent of revenue, market by market over time. That's as accumulated awareness grows and word of mouth gets going. We've also seen our newer markets start at a much lower advertising expense per unit or percent of revenue early on.
At the company level, our advertising expense as a percent of revenue is driven by a strong mix effect of adding new markets that have higher ad expense per unit and seeing existing markets lever. Over time, we expect our mix of markets to be much more seasoned than our current mix of markets, which drives leverage along with these other dynamics that I described. Okay. In logistics and market occupancy, we also see a significant opportunity to reduce expense as a percent of unit over time. I think the big takeaway here is there are significant benefits over time from adding more inspection and reconditioning centers. We are currently serving nearly the entire country, coast to coast, from Boston to San Diego, out of four inspection and reconditioning centers. That is not the long-term state for our logistics network.
As we add more inspection and reconditioning centers over time, we expect the average distance that we're driving to our customers to come down, and average logistics expense as a percent of revenue to come down as well. Ernie will spend some time talking about this as well, so I'm not going to talk. We do a lot of modeling and simulation to try to get a good understanding of these average distances over time and the impact on logistics costs per unit, I'm not going to dig into that here. Ernie will spend a moment talking about that. Finally, there's some scale benefits in the logistics and market occupancy network. That just comes from higher utilization in logistics and leverage, or more throughput through a somewhat slightly fixed infrastructure and market occupancy. Okay.
Lastly, I'm going to hit other SG&A, where our long-term goal is 1.75%-2.25% of revenue. This is another place where we've shown significant improvements in the last few years, reducing other SG&A as a percent of revenue from 9.4% in 2016 to 6.4% today. The gains in other SG&A come first and foremost from leveraging the technology and corporate components of SG&A, which we expect to be about half of this bucket in the long term. We also see some benefits in a second component of other SG&A, which are more transaction-related and customer benefits. We see some opportunity there from insourcing and automation that we expect to lever that over time, and in the long term, we expect it to be about half of other SG&A. Okay.
With that all said, I want to combine everything that we've talked about today, and we've talked about a lot, and it's gone fast. I want to combine all of that into our long-term financial goals, in particular, as it relates to EBITDA margin. You can see here our long-term gross margin goals, our long-term SG&A goals, all summarized up. Those all sum up to a long-term EBITDA margin goal of 8%-13.5%, which we're very excited about and believe is enabled by the significant benefits in both things like our finance platform and in our lower long-term cost structure. Okay. On this slide, I've covered a lot, but there's a couple of questions that I expect, or we all expect investors to be asking about this slide.
The first is: How, if at all, have your expectations for long-term financial goals of the business changed since you IPOed? The answer there is that our expectations, based on all of the operating we've done over the last two years, have actually increased since our IPO. In particular, we see bigger opportunities in buying more cars from customers and in monetizing our finance platform than we did a few years ago based on the traction and the success that we've seen over the last couple of years. Moreover, I would say the last two years of operating have bolstered our confidence in the long-term cost structure based on the leverage that we've experienced over the last two years and all of the multitude of opportunities that we see to drive gains from scale and other technology improvements as we look forward at SG&A. That's question one.
The second question we expect investors to ask about this long-term model slide is, when is long term? We don't plan to provide a specific timeline on that. However, we do plan to continue to make significant progress toward achieving these long-term goals, as we have done over the last several years. Okay. In 2019, we expect significant gains in GPU, significant gains in EBITDA margin, and a decline in EBITDA losses relative to 2018. We think of 2019 as basically the next step toward these long-term goals, which we've been marching to now for several years. I think that basically wraps up what I wanted to share with you. Just to reiterate, I think we're incredibly proud of what we've been able to accomplish in building this business over the last five years, and also in our roughly two years since our IPO.
That's a testament to the team, as Dan alluded to. We're very proud of it, we're very excited about where we're headed. With that, I'll hand it off to Ernie.
My mic's up. Everyone's excited about EBITDA. I noticed you didn't clap when I got off stage the first time you did for Dan and Mark. I'm not going to worry about it too much, but I'm keeping score, just so we're clear. What I'm going to cover is where we're headed. I think it's still very early days for Carvana, we're growing fast. We're a big company, but relative to the size of this market, we are still incredibly small. Trying to figure out where we can go is a difficult task. We're going to take two swings at it. One, we're going to start from the top down. We're going to look at the broader market, see where our limitations may be, and see where we think we can go. We're going to look from the bottom up.
We're going to say, what have we learned over the last six years? What are the sensitivities that we feel like we understand, and what does that suggest about where we can go? We'll put that together for our long-term expectations. Let's start with market size. We're going to start top-down. The market is 40 million units annually. It's often broken into three sub-markets. There's a franchise dealer market, independent dealer market, and a private party market. Our view is if you start with the customer, those kind of separations of the market is emergent from dealers not necessarily serving all of the customer's needs. Customers that go to a franchise dealer, an independent dealer, they're choosing to go probably spend a little bit more money, but be able to handle their trade-in and their financing in one place, have an institution standing behind the car.
Those that choose to go to the private party market, they're saying, "I want to save a little bit of money. I'm willing to do more work myself." Those that are deciding between a franchise and independent dealer, probably today, the biggest driver of that is which location has the car that I'm looking for. We believe that our model with great pricing, a very simple experience, bring the car to you, great selection, we believe that we can address all of those markets, we don't think we're limited in this way. Next, we're going to head over customer demographics. Who are we selling to, and what does that tell us about how we may be limited in the future? First, what does our age distribution look like?
If we look at the distribution of customers that we are selling to, it looks a lot like the distribution of car buyers in the market more broadly. The dark blue line is car buyers in the market more broadly. The Carvana blue line is the customers that we are selling to. I think there are 2 major takeaways from this graph. One, we are selling to customers of all ages, we do not believe that we are limited. Two, we are skewed younger, which we think is positive. That is the portion of the market that is growing over time, that excites us. Next, are we limited in credit in some way? Are we attracting customers of a certain type of credit and not attracting customers of a different type? Clicker is not working. There we go. Here is our credit distribution. Again, dark blue is the market distribution.
Light blue, Carvana blue, is our distribution. For real world data, that is about as close as distributions get, the difference between those distributions can be completely explained by the difference in customer age distribution to the left, we are not limited in credit either. Going forward, we are going to talk about cars. That is kind of the demand side. What about the supply side? Are we limited in any way in the types of cars that we can sell on our platform? First, let us look at the types of cars that we are selling today. The Carvana blue line is total retail sales that we have sold in 2018 year to date. You can see there is a pretty heavy skew to two and three years. That is actually very normal in the market.
There is a lot of cars coming off lease, a lot of cars coming off rental, that skew exists in the market more broadly. We are selling significant numbers of cars out to about nine years. If you look at the market generally, you look at registration data, about two-thirds of all car sales are from zero to nine years, we do not necessarily believe that that is a fundamental limitation to the age we can go out to. Interestingly, this is a place where customer source cars really helps. You can see the cars that we are buying from customers, then selling to other customers is shifted to the right pretty strongly. There is another distribution there that we are not showing, which is cars that we buy from customers that we do not sell to other customers because we are not electing to sell those cars yet today.
That is shifted further out to the right. It makes sense that basically every car in the market is owned by a customer prior to being sold to another customer, if we are buying from customers, we have access to the sum of all supply. We definitely believe that we are not limited here, we believe we can continue to move to older cars and expand the market that we are serving. I think the next question, though, is, okay, what does that do to your brand? Can you do that if you are selling cars sight unseen, and if customers are worried about the quality of the car that ultimately gets delivered to them? Are you putting your brand at risk by doing that? We can look at some data to try to figure that out as well.
What this graph is the average rating that we get from customers. Every time we sell a car to a customer, we ask them to give us a review of how we did. Our average review across the site has been 4.7 stars for a very long time. It remains 4.7 stars today. If we look at that by age of car, we can see there's virtually no correlation there. There is a very subtle correlation. It's in the direction that you'd expect, but it is extremely slight. A little bit of noise that shows up in the tail is largely due to smaller counts. I think what we're learning from this is that customers have rational expectations. When they're buying a nine-year-old car, a 10-year-old car, they expect it to have a couple more dents and dings, and that's not impacting us, right?
A review is really relative to a customer's expectations for how we're going to perform, and so we're seeing that we're doing a good job there. That's exciting to us because, again, we feel like we're not limited in this enormous market. Another question is, we started this whole conversation with 40 million, and I think where people want to start is they want to say, "Okay, there's 40 million cars sold annually. How are we going to limit this down to figure out what is addressable for you?" I think that that's a fair place to start, but it's also important to think about where 40 million units comes from. In most markets, the markets are defined by production and consumption. A very simple way to think about it is groceries.
Groceries have to be produced, and then somebody buys them, and they eat them, and they're gone. That's how the market works. It can only grow up to the amount that people are consuming, and it is constrained by the amount of production that occurs. In cars are used products. Customers aren't consuming a car. They're consuming miles and time. There's not really a fundamental limitation how many transactions there's going to be. The transactions is simply a function of how often customers choose to trade those cars out. What can we learn from 40 million units? If we start with the total car park in the U.S., about 270 million cars total, and then we say, what is implied by the fact that we're seeing 40 million units per year? What's implied is that customers are choosing to switch cars, right?
They're trading out which car they're consuming miles and time in roughly every 6.75 years, and that is what's defining the market. That's really the fundamental that drives the 40 million unit number. Why are customers choosing to switch cars every 6.75 years? Why don't they do it faster than that? We think there's three primary reasons. One is there's cost frictions. It's expensive to go buy a car. If you think about two customers simultaneously going to the same dealership and trading in their car and then buying the other customer's car, you can think of the costs of transacting are effectively that dealer's retail gross margin. That's effectively what it costs to switch out your car. That's expensive. You only want to do that every so often. The next one is experience frictions.
There are many dealers that treat customers great, but the industry reputation isn't the greatest, and most customers aren't excited about going to the dealership for 4 hours. I think you have to really decide, I want to do this, right? I need to really want to switch to a new car before you decide to do it. Another one that I think is very interesting is negative equity. A customer also has, aside from just cost, which is kind of like, that's maybe financeable, what are the cash requirements for me to get into my next car? If I bought a car and I'm in a loan and the loan's amortized and the car's depreciating, if the loan balance is higher than the value of the car at any point in time, I need to come out of pocket with cash to pay that off.
Because we're selling cars at a discount, let's use a simple number. If we're selling cars for $1,000 less, that basically means that early in the loan's life, our customers have $1,000 less they have to come out of pocket in order to choose to switch another car. As you make it all the way through to the end of the term, that collapses to zero, but because of the shape of amortization, it stays very high for a long time and dramatically reduces the cash cost of choosing to switch. What would happen if this market kind of evolved in a different way, and if customers chose to switch cars a little bit more often?
If customers chose to switch cars every six years instead of 6.75 years, the market as a whole would kind of immediately morph into something that's 12.5% bigger, and there'd be 45 million transactions per year. If people chose to switch cars every five years, it would switch to a market that's 35% bigger, and it would be a 54 million unit per year market. If you go to four and three, you get to 68%, 125% larger market. That's a very powerful dynamic when we're thinking about the size of this market. I think when we think about it from our perspective, it's unlikely the entire market is going to evolve in these particular ways. What can happen is we can go out to our customers who face different frictions to transacting, and we can get them to switch more often.
Let's say our customers switch every five years instead of 6.75 years. We may have X% of customers, car-buying customers, and 1.35X% of transactions because our customers are choosing to transact more often. We think that's an incredibly powerful opportunity, and we're very well positioned to handle that because the other requirement to choose to switch your car is you have to kind of have the moment where you decide, I want a new car now. We're very well positioned to provoke that moment, to speak to our customers and say, "We know what cars you were looking at on our website. We know that you decided to trade down from that $25,000 car to the $20,000 car at the last second, and we know what you really wanted, and now you've paid down your loan. We know what your loan balance is.
We know what your credit is. We know how to value your trade-in sight unseen, and we can shoot you an email and say, 'Hey, do you want to spend an extra $40 a month and get into a brand-new car?'" We're positioned to do that very easily and to conjure up that moment. I think that's a very exciting opportunity for us. It's also interesting to kind of try to learn from other industries. First, let's size our industry. This is slightly different data we've presented in the past. This is Bureau of Labor Statistics data, and they define markets slightly differently. What you're seeing on the left is automobile dealers. Basically, they're adding up all revenue from all automobile dealers. It's new and used and service, but it's not private party. It's a slightly different definition.
It ends up adding up to roughly the same size of the market. Look at the size of that market compared to these other markets. We've got grocery stores at $640 billion. We've got health and personal care at $333 billion. It's interesting to see how large our market is and how much of the retail market in general it makes up. It's also interesting to now dive in and say, how are the structure of these markets a little bit different? What this slide does is this looks at the market share of the largest player in each of these different retail verticals. There's one incredibly clear outlier here that begs a lot of questions. Why is automotive retail so fragmented. It's hard to really define or explain that, I think, in a way that is satisfying.
We've taken a lot of different attempts at it. The best and most objective kind of answer that we can come up with is just that historically, automotive retail has had very, very few economies of scale. Generally, the way the business operates is it operates as sort of a single unit. The dealership has all of the operations, all of the management contained inside of that unit, and it's printed one at a time. If you want to accumulate many of those units, you actually need to create a hierarchy that sits on top of those units that has mass. Because the way that the business is run, it's very hard to make up for that mass across all these different dealerships. As groups get bigger, they tend to fragment over time.
I don't believe that's a completely satisfying answer, but it's the best answer we can come up with. There may be a lot of inspiration here for all these other businesses, though. Are we supposed to look at our industry to try to figure out how big we can be? Are we supposed to look at these other verticals that look a lot more similar to one another? If we take that approach and we say, how are these other businesses doing, and what can we learn from the average of all these different businesses? What is implied by the size of players that there could be in automotive retail? We arrive at this. We look at the average of the five largest retail verticals, and then we look at the four largest players in each of those verticals and look at their market shares.
The largest player generally has a 31% market share. The number two player generally has a 17% market share, then 7% and 4%. This market was defined slightly differently than the way we define our market, which is 40 million used units. We're now going to apply those market shares back over our market of 40 million used units. If you take that 30%, multiply it by 40 million, that's not a really hard equation. You get to 12 million. That's an enormous player. That's unlike anything that's ever been seen in automotive retail. No one's approaching even closely a million units today. That suggests that there's an absolutely incredible opportunity. From where we are, that's north of 100x. We're a long way from there, but this suggests that there could be room for something absolutely enormous.
Let's kind of switch and let's take a bottoms-up approach and look at this a different way. When we think about growth, we've got a very simple growth model. There's generally three drivers. The first is we open new markets. Everyone's become very comfortable with that. The second is those new markets catch up to old markets. We're learning a lot every day about how the old markets are performing, and we expect all of our new markets to perform similarly over time. The next is older markets continue to grow. We keep getting additional market penetration. This is the hardest part to try to figure out of this equation, at least in our minds, because we don't have a model to look at. There's nothing crystal clear to model it off of. How do we think about it?
We think that there's basically two underlying drivers of growth. You can decompose total growth into two things. One is underlying demand. For purposes of this, we're defining demand as someone who comes to your site and is likely to transact based on any given service level, let's say a perfect service level. If you have every car and every car can be delivered in minutes, how many sales would you get? As you get more people coming to your site, that number goes up. As you have less people, that number goes down. Those movements in the number of people that want to transact conditional on a service level, that's demand. There's also this question of conversion, right? As we're growing, we're getting more cars in our inventory. Our delivery times are changing. We're learning things about that.
If we look at each of those and try to say, what do we think we can grow by in each of these different areas? We start with demand, and we've got a framework for this. There's many ways to think about our demand funnel, but this is how we think about it. The first thing has to happen is customers have to become aware that we exist at all. That's the same in virtually every business. Where we start to get a little bit different is they need to get an understanding of what we do. We're asking customers for a behavior change. That requires education and takes a little bit of time. That's different than many other industries where you're just growing into a well-defined industry. The next thing, they have to have an open mind. Not all customers want to be online car buyers.
There are some people that view themselves as car enthusiasts, and they want to go kick the tires, and they really feel like they understand every part of the car, and they want to lift the hood up and check everything out. They're not open-minded to what we do. The next is credibility. For this one, I like to think about my mom. One of our values is our next customer could be your mom. My mom is someone who, five years ago, if she wasn't related to me, would have thought buying a car online was completely crazy. I think if she then spoke to one of her friends or neighbors or family members and they said, "Hey, you go to this site, they've got a great selection. It's really easy.
The prices are very fair, and they show up, and if you want to return it, they don't make it a pain at all," she would have switched. I think that's probably our biggest area to be able to grow over time. If we just deliver great experiences, we're going to be able to continue to get people to view us as more credible. All of these are channels for potential growth. Awareness is generally driven by marketing or word of mouth. Word of mouth is the only thing that scales exponentially, so we work really hard to deliver great experiences so we scale that way. It's interesting to think about marketing as well, though. Our company's total advertising spend life to date, we're a less than six-year-old company, but life to date, the total dollars that we've spent are roughly $175 million.
If you put that in the context of others in this industry or any of the other large players or businesses that are building awareness, that's not a huge number. We think there's a lot of room for us to continue to grow awareness. Understanding, that's driven by marketing, word of mouth, website communication, advocates. There's a lot of things that we're working on to constantly try to drive up understanding, and there's many levers there. Open mind, this is something, at least in the way that I define it, that we can't really control, right? This is something that's just going to happen over time. Different people are going to have different preferences. The good news is we do believe that customer preference migration is at our back. On average, customers are getting more comfortable all the time buying more and more things online.
They're defining themselves less by their car, which makes alternative transaction methods easier. Credibility. Credibility, most importantly, by a long way, is driven by great customer experiences. The most important thing we can do is give customers a great experience. They tell their friends. That's a much cleaner channel for communication than anything we can ever do ourselves. We're constantly working on that. A customer to buy a car from us has to make it through all of these different gates. They have to be aware, they have to understand, they have to have an open mind, they have to have credibility, and we're working very hard to get growth in all those different areas, and those are great levers for us.
It's hard to say, how do we take all of these kind of amorphous concepts and try to figure out how much growth we think is possible? We've taken a swing at it that is far from perfect. I don't think traffic is in any way, shape, or form a perfect kind of proxy for demand, but it's pretty good. Traffic definitely varies in quality and the way that we define demand for this, you'd like to ideally control for the quality of any given traffic. We don't think our web traffic has varied that much over time. We also don't think our web traffic is that different from the traffic that other players see. How has our traffic in Atlanta, for example, grown over time? This was 2013. 2014, we saw a lot of growth.
That growth converted over to a lot of growth in sales. 2015, we actually saw less growth in underlying traffic. We saw enormous growth in sales. That year, that growth was largely driven by conversions. Our inventory grew quite a bit. 2016, we saw another great growth year. 2017, continued to grow traffic. We've been growing traffic very quickly. It's hard to take that line and say, where does that go? We at least know there's a fair amount of momentum in that line that suggests we can go upward. Another approach we can take is we can say, let's look at other players in the industry. Let's look at their website traffic. Of the biggest automotive websites in our industry, what does their traffic look like compared to ours, and how might we be able to perform compared to them?
If we look at the biggest players in Atlanta alone, we think that there's 3x-6x as much traffic on these other automotive websites as we've got. That's not a perfect proxy. As I said before, traffic may convert differently on our site than it does on their site. We think it's pretty good. We also think that it's not necessarily a logical ceiling. Given the fragmentation that we showed you earlier, it's not clear that those largest players are the ceiling of where the largest player can be in the future. That suggests a lot of possible growth. We think it's very exciting. Now let's move over to conversion. We moved off demand. Demand is going to grow at some speed. We're going to try to predict how much growth there is there. Now the conversion side.
The two number one drivers of conversion are, one, increasing inventory size, and then two, decreasing delivery times. We have internal estimates for how these two things impact customer selection or conversion. We're not going to share those. It's very granular. The estimates are very meaningful. We think these are the two biggest drivers. Let's do a little bit of extrapolation. I think many businesses can grow through execution. Generally, the form that takes is you open more stores, right? I think every business faces a lot of different risks. Probably the best risk that you could take is execution risk, because at least that's contained as a management team. That means the ball's in our hand. We've got control over it. We think we can grow through execution in two ways.
One is going to be opening markets, and we'll start with that. If we take the Atlanta market share from Q4 2017, which was the last time it was disclosed, we say, let's say we get to 90% population coverage over time, which roughly ties this mental model of the 200 largest MSAs. What does that suggest? We've got 1.54% market share. That number is defined against the 40 million annual U.S. sales times 90% population coverage. This suggests 540,000 sales. That's what we think we could do by basically opening markets and then letting them just catch up to Atlanta. As we're doing that, we're getting more and more sales, right? That's feeding back into this system that's got benefits from conversion. What does 540,000 sales mean? It means more inventory on the site, and we can calculate how much.
If you have 540,000 sales and then you've got a 45-day turn time, we're picking a number here. It may be a little higher, but a little bit lower over time. There's a lot of opportunities there. That roughly suggests you'll have about 30 days on site. If you've got a car on site for 30 days, that means that of your total sales, you divide by 12 to get how many cars are on site at any point in time. That suggests there'd be 45,000 cars on carvana.com. We can also calculate what it means for IRCs. If you're selling 540,000 cars per year, you need more inspection centers. We can produce about 50,000 cars per inspection center. That suggests we'd have about 11 IRCs. The benefit of that is that means cars are closer to customers, and therefore you have faster delivery times.
Now we're just going to kind of pose a question that is incredibly poorly formatted, but you're going to get it. First, what happens when inventory goes from 8,000 units on the site, which is what was on the site on average in Q4 of 2017, where that 1.54% market share came from, goes to 45,000 units on the site. That's going to be 5.5 times increase in the inventory that's on our site. What happens if we go from 3 inspection centers to 11, meaning that cars are closer to customers, they can get it faster? What does that do? Again, we're not going to quantify it today, but we think it's actually very meaningful. Then excitingly, it feeds back on itself. That's not an equation that you do once. You run that equation back over and over again. We get more sales.
That means we have more inspection centers. It means cars are closer to customers. We have shorter delivery times. That leads to more sales. You've also got the feedback loop of inventory. We just did that equation. We grow our inventory. Now our market share goes up because we have higher conversion. Now we got more sales again across the entire system. We got to recalculate what our inventory is, and we keep cycling back through. We really do believe there are very meaningful fundamental network effects in our business, and we think that suggests something that's extremely exciting. We take those two perspectives. On the left, we've kind of got our top-down perspective, and on the right, we've got our bottoms-up perspective. You get to really big numbers. If you've been doing mental math in your head, you've gotten to very exciting places.
Both of those approaches are kind of a demand-side approach. We also have to think about the supply side, right? We have to buy cars, recondition cars, build these inspections that are shipped into customers, run all the operations associated with running our business. We need to also think about what we think we can operationally accomplish. When we put that all together, we think that demand is seemingly not going to be our constraint. We're much more focused on the supply side of the equation, and we believe that over time, we're going to be able to sell Ooh, man, this is going to be a very dramatic reveal because there's a white slide that I didn't realize was going to be there. Hold on a second. Here it comes. We think we can sell 2 million-plus units per year.
That's a number that is a significant multiple of anything that's ever been seen in automotive retail. We feel confident we can do it. We feel like you can extrapolate there very comfortably from what we've already seen. As I said, we think the risk that stands, or the primary risk that stands between us and achieving that goal is our ability to execute year after year after year. It doesn't take a lot of imagination anymore. We can see it in the data, this kind of opportunity exists, and we're an ambitious group that's going to go out and make this happen. This is our goal. With that, I think we're going to switch over to some video.
Something that's very interesting about any business with network effects is aside from trying to decide where you're going to end up, you can look at it another way, because if scale is driving scale, you can also say, "Let me go out to that scale point and let me look at what the business would look like there. Let me think about what that business looks like, and let me think about whether or not the attributes of that business actually make sense. Am I being under or overambitious?" Ooh, we had a little. Let's go back. Did I go forward accidentally? Did that happen? Are we on a slide, or is this the video?
It ran.
Oh, it ran. Oh, God. It was running while I was talking. My bad. I'm sorry. I called that wrong. You just saw this, now I'm going to explain it. What we've got here on the left is total inventory that we've got at carvana.com. We've got in our x-axis is basically miles that those cars are away from a customer. What this green line represents is an interesting line. This is if we look at our 200 theoretical markets that get us to this 2 million units per year, what is the average population of that market?
If we assume that there's 40 million cars sold per year, we extrapolate that down to how many cars are sold per person, and we look at how many cars exist inside of each market, that's the line for the total inventory in all of these markets. What you just saw there is basically from when we launched to today, right? We've constantly grown our inventory. You can get a sense of how far away our inventory is from customers. We've already crossed over the total number of cars that most customers in these 200 markets actually have in the sum of their entire market. That's really, really exciting. There's a lot more because of all the feedback we just talked about.
Now what we're going to do is we're going to move to the next one, which will take us from current to 2 million units and say, what does this world look like? I'm going to watch it this time so I can speak to it better. As we're growing, if you get to 2 million units per year, and we do the same math we did before, we say we're going to have 30 days on-site, what does that mean your ultimate inventory is? It's 2 million divided by 12. You've got about 165,000 cars on your site. That's an enormous number. That means that on average, we would have on the order of 10x as many cars on carvana.com as exists in the entirety from private party sellers, franchise dealers, independent dealers of the market our average customer is in.
That's a huge, huge, huge advantage. Conversion, we hope, would go way up as a result of that. Now, interestingly, you can also look back to the left on this slide. We'll also have many, many cars extremely close to customers. We're going to have a multiple of the number of cars that are in any given market within 500 miles of customers. We're going to have significantly more than that within 1,000 miles, within 1,500 miles. Obviously, this is a simulation. The actual data will play out in different ways, but this is an objective approach at trying to build out an understanding of what this is. What happens to conversion at that point? We think it could be very, very exciting.
The next one we're going to go to is we're going to think about the inspection center kind of drivers and how that unfolds over time. If we're going to ultimately sell 2 million cars per year, that means we're going to have 40 inspection centers. As we build out all those inspection centers, we're going to move closer to customers and delivery dynamics are going to change quite a bit. Let's look at that. This is basically our history to date, and there's some interesting dynamics going on. We've got two lines here. The blue line is basically the distance from our average customer to our average car. It's equivalent to the distance from our average customer to our average inspection center.
The red line is the distance from our average customer to the nearest inspection center, so to the cars that are nearest to them, and those are two important lines when we're thinking about how this unfolds over time. At first, we only have one inspection center outside of Atlanta, so that's why there's a single line. When we open our second inspection center outside of Dallas, you see the average line move way up because we're moving away from a lot of our sales in the Southeast. The next interesting dynamic you see is we open up the Florida markets. Both the nearest inspection center line and the average inspection center line move up because we opened up markets that are far away from customers.
The next is we open up the Delanco IC, the inspection center outside of Philadelphia, and you see things kind of move up again. You see this interesting kind of immediate up and down move in the red line. What's happening there, we opened up the market of L.A. L.A. is a huge market, and we opened that up. We moved much further away on average from our customers. We then opened up the Phoenix IC, and so that pushed it back down. These are numbers that have been moving up pretty quickly over time and pretty significantly over time, and that's driven our pinch points, and it's driven some of what you see flowing through our financials in terms of our average logistics expense, et cetera. Now the question is, what happens from here, right?
As we continue to build out more and more of these inspection centers and get more and more demand, what happens to the shapes of these curves? Let's go ahead and play the next video. Again, this is simulation. Real life won't look exactly like this, but we're trying to do our best to figure out what it might look like. What we've just done is taken that endpoint, moved it all the way to the left, and we've changed the scale a little bit. The first great news that I want to bring up here is you can see that the average distance, the top line, the blue line, is kind of barely moving from here. We've basically expanded across the country now. We've got a footprint that enables us to deliver cars to customers everywhere.
Now as we get more customers, open more markets, put down more inspection centers, the average distance away from a customer is only going to move very, very moderately. We are also approaching this point where the distance to the nearest customer is about to shrink very, very quickly. This is a different scale version of the same graph that Mark presented, if you're paying attention to the numbers there, what is suggested by this is that the car that is closest to our customers is going to go from around 300 miles away to around 75 miles away, call it. We'd move to about one-fourth of the distance that we are today. That's really big. Now we also need to think about what's happening to the average customer, right? Customers are going to choose cars in a certain way.
Early in our life, we have a relatively small inventory. If you think about the customer's kind of flow, I go to the website, I look for the car that I'm most interested in, I pick that car, I see a delivery time, I buy it. If I'm randomly selecting because I'm running into all these different cars and there's not a lot of great substitutes, I'm going to shift into a car that's nearby, I'm going to be at the top line. If you go all the way in your mind to an infinite inventory, if you have every car everywhere, now what's going to happen is I look for my car, I have perfect substitutes, I'm going to go to the red line.
What should happen over time is you should start closer to the blue line, you should approach the red line, which means kind of the slope should be even steeper than the slope of the red line, that is the slope that drives our expense and our average delivery time. We think there's a lot of gains to be had there that's really, really exciting. Now, something else that's implied by this, if you kind of combine these two animations, what this means is every one of our inspection centers, we've got 165,000 cars on our site, we have 40 inspection centers. We have about 4,000 cars at each inspection center. This means that we're going to have about 4,000 cars within 75 miles of our average customer. That's an hour-and-a-half drive away.
You can imagine a world that'll take a lot of work and a lot of ops and a lot of creativity and great technology, you can imagine a world where we can be delivering thousands and thousands of cars to customers faster than they can go to a dealership and buy it. In that world, we think things could be very, very exciting. That I think is a fun exercise as well. We'll now click over. Perfect, let's go to the next slide. Oh, I got to do it. We're going to close on this. Our mission is to change the way people buy cars. When we say that, we've said it since the very beginning, we mean it. We think we can sell 2 million-plus cars per year. We think we can be the most profitable automotive retailer that's ever been.
We think we've got the business model to do it. We think we've got the customer response to do it. We think we've got all the levers, the team, the scalability, the ambition. That is our intention. With that, we'll close the presentation portion of this. We very much appreciate you guys coming, and I again am looking for a clap, so thanks.