All right, to go ahead and get started. Good morning. Number two, Citi TMT Conference . Very pleased today to have the CFO of Carvana, Mark Jenkins, in town. Mark, nice to see you. Thanks for being here.
Hey, thanks a lot. It's great to be here. It's good to be back.
Yeah, awesome. We obviously had a great discussion last night at dinner, which was great. But yeah, I wanted to go through a few things. I'll get through sort of a few, sort of a list of questions, but obviously would welcome, encourage one or two from the audience if you have any. So, things have changed a little bit in the last six to 12 to 18 months. I think the question number one is like, okay, so you guys are clearly in this moment of acceleration, right? And I think you're aware of this. One of the questions we need a lot is just how and why has this occurred? Like what are the events that have led to this at this point?
We'll dig into it more, obviously, but at a high level, like we've been to the 30%, 40%, maybe 50% growth and endorsing those numbers, but third party led us here.
Sure. Yeah. That's a question that we get a lot of. One of the things you've all seen over the past couple of quarters is we've been growing at a mid-30% rate on a year-over-year basis on retail units sold when we guided to accelerated growth relative to the second and third quarters, and that's happening in an industry environment where industry sales are approximately flat, so we're significantly outpacing the industry. We're gaining significant market share, and then a natural follow-up question is, well, what's driving that? And I think we see many drivers. We see many of the things that we view as long-term growth drivers that are actually playing out today, so I can list off a few of those, so we're always making the product offering better.
I think we've been really on a roll over the past couple of years. Our focus on efficiency has also led to better customer experiences, and so that's doing things like improving the customer shopping experience on the website, improving the customer transaction experience, trying to make that as seamless as possible, improving the customer delivery experience, including faster delivery times, making the post-sale experiences as sort of robust and efficient as possible on things like title and registration. We're also trying to improve the experience of customers that are selling cars to us, and so there's all kinds of things that we're doing to improve the customer experience that we see. It's been an outcome of some initiatives, but also something that we view as a long-term growth driver. In addition to that, we've been raising awareness of our brand, and there's a few different ways that that happens.
One, we think we're the beneficiaries of a long-term secular tailwind of customers buying more and more things online. That's our growth over our first 10 years of existence. To go back, prior to our focus on profitability, we were one of the four fastest companies to join the Fortune 500, along with Google, Amazon, Meta, and Carvana. And so we think we're benefiting strong secular tailwinds toward customers buying more and more things online. We think that that's helping us now as it has been over our life. We're also continuing to build awareness of our brand. Even though our advertising spend has been relatively flat over the past couple of years, that spend still layers on top of each other. We're continuing to build awareness. We're getting gains from word of mouth, repeat customers, things like that. So I think that's also a driver.
And then we're also only in the beginning phases of really restarting the positive feedback in the model. And that comes from things like adding more inventory pools, adding more selection, further speeding delivery times through the addition of incremental inventory pools, all of those sorts of drivers, only in the very beginning stages of taking advantage of those so far this year.
Got it. Okay. Then I think at times in the last, call it 12 months or so, as you were kind of working on the operational rigor and efficiency of the company, I think you were pretty clear like you were throttling, full-on throttling inventory from a supply perspective. So obviously that was going to impact demand. Exactly what were you doing during that period that was actually going on? Then secondarily, how much is that change? Is that a part of this strategy? You mentioned inventory pools in place. So you kind of explain what that means.
Sure. Yeah, so I think there's a lot of things to discuss on that question. What are the important components of scaling supply for us? It's really, I think, first and foremost, it's scaling production capacity. This network of inspection and reconditioning centers through the third quarter was sort of on the order of one-third utilized from the standpoint of utilizing the actual infrastructure that we have. And by infrastructure, I mean space and buildings. The key then to scaling production capacity is to add staffing to the existing infrastructure. And so a lot of what we think of when we think about scaling the supply side, it's all about scaling staffing. First in the reconditioning centers, but the business, you have to scale staffing across. We have four major operational areas: inspection and reconditioning, multi-car transport, last-mile delivery, and then customer care and transaction processing.
Those four big operational groups need to scale, and staffing is the key. Given we have so much infrastructure capacity that we built in sort of the 2021 to 2022 timeframe, as well as through the acquisition of ADESA.
Got it. Okay. And then on that third staffing or whatever, is that literally like production lines that will start to scale up? Is that what you mean by that?
Yeah, that's exactly right. So we have very significant advantage for being able to produce cars with the land and the buildings that we need. The key is adding production lines by adding staffing at these centers. Now, the component of that, some of that is staffing up traditional cars. A second component of that is integrating Carvana-style reconditioning into ADESA locations. So as background, we acquired ADESA, a nationwide physical auction in 2022, 56 locations. We've started to integrate Carvana-style reconditioning into some of those locations so far. And we think that's a benefit because it adds more inventory pools that makes inbound and outbound transport more efficient and puts more cars closer to customers. But that has been a component of growing production is integrating these ADESA physical auction locations with Carvana-style reconditioning. What does that mean exactly?
I think there's two key steps to integrating these ADESA locations into the Carvana reconditioning infrastructure. One is adding our proprietary reconditioning software, CARLI, our internal platform for reconditioning cars, adding those into the ADESA locations, and then the second is process management, so basically, these ADESA sites start as physical auctions, and they have their own set of processes for managing wholesale marketplace customers. We need to go in and adjust those processes so that the sites can handle traditional Carvana retail reconditioning while also continuing to serve the sellers and buyers of the traditional wholesale auction, and so that's a process management integration.
No, that's clear. So I guess the question is like, okay, you're running this, sorry, first thing focused on this one-third utilization, essentially. So you start to staff that up tomorrow, right? Does demand just flow in from there? Is there anything else you need to do to drive demand once you start to do that? And then secondarily, what will ultimately instruct you to start that process of filling in those lines in a more heavily utilized way?
Sure. Yeah. So I think there's a couple of components to that. So I think driving production, which drives selection, is a long-term driver of sales. So I think that's exactly correct. I think we see staffing the IRCs as the key to driving selection. We think we have great access to inventory, and so we're really focused on staffing. The other things that you need to get right as you scale is you need to staff not only production, but you need to staff those other three major operational groups that I listed earlier: multi-car transport, last-mile delivery, customer care, and transaction processing. Those all have to be moving in lockstep so that you can have the full experience all moving at the same speed to provide the customer experiences that we're looking for. So that's some of the key to scaling.
I think the way we're thinking about approaching that is in 2024, we're taking a very measured approach to steadily ramping production capacity. As I mentioned, one component of that is through we've done six ADESA integrations where we're ramping production at those locations. We're also ramping production in our traditional Carvana IRC locations. We're taking a measured approach to that, I think, and feel really good about the plan that we have and feel like we're executing well.
Got it. And so as we've seen kind of the volume rise and we'll get into margins in a sec, I mean, is it fair to say that at each level you guys have this new playbook in terms of processes and software and those integrations? At each level, are you seeing the marginal contribution match what's already in place? Because you guys, I feel like 18 months ago, you kind of ran stuff. You shrunk things, right, to run it in a test tube, so to speak, and learn and figure out these processes knowing you would eventually scale. As this process has started happening, what are the challenges you've run into? Have there been any challenges or you're just really, you think you've got the playbook nailed now?
Yeah. I mean, I think we feel very good about the playbook. Obviously, on a year-over-year basis, we grew significantly while also increasing margins significantly. We have a heavy fixed component or heavy overhead component of our cost structure. So I think that that leads to a complementarity between growth and margins. In addition, I think if you zoom in a little bit, in Q3, we grew, but we also improved our costs. So I think the goal for us is to, as we grow, to actually improve our costs. We think we can do that because of some of the fundamental gain initiatives that we have that all of our teams are focused on, as well as we have the opportunity to take advantage of overhead leverage.
Yeah. Yeah. Got it. And then, yeah, so leading to GPU on that question, I think when I go back, God, this is five, six years ago, right? And we look at the targets you laid out, at least on an absolute dollar basis for GPU, you're running, my math, not yours necessarily. I think you're running something on the order of 50% higher than what we might have originally thought. I don't want to ask why your modeling was so off because it's a good thing, but why are we so far ahead than what we used to think for an absolute dollar perspective on GPU?
Yeah. Yeah. So I think there's a lot of things to hit on that question. So the first is there's been significant inflation in the industry since 2018, which is when I think the targets that you're setting, that you're referring to, were set out, and we do think that that obviously has an impact. It has an impact on industry costs, and as a result, it also flows through to having an impact on GPUs since dealers need to cover their costs, and so I think the fixed dollar targets obviously are not evergreen because there is inflation, so I think that's actually an important point when thinking about something like 2018 dollar figures. There has been material inflation, and that's a big driver. Now, moving on from that specific point, I think there's also been meaningful business changes. So for example, we acquired ADESA.
We're generating a meaningful amount of non-GAAP gross profit from ADESA, a third-party marketplace business, and so that's something that wasn't baked into initial gross profit numbers, for example. I think we've also had success in other areas. For example, we have a first-party nationwide logistics network that can serve customers very fast and very cost-effectively. And we've been able to generate some revenue from that logistics network. That's something that wasn't necessarily part of our 2018 GPU targets. I think there's been, I think, first and foremost, there's been real inflation in the industry that has raised everyone's costs and that also flows through to GPU. And I think that's really important to keep in mind, but then also there's been some fundamental business changes that we think we're benefiting from that have had meaningful and sustainable positive impacts.
Got it. Okay. So yeah, let's dig into that a little bit. So again, with the inventory coming online, maybe with ADESA, right? Obviously, that integration is clearly occurring. Talk about the mix of where you're getting your cars from and has ADESA sort of changed any of that? Yeah, start there if you could.
Sure. Yeah. So the significant majority of our cars we purchase from consumers. We think we provide a great experience there where you can get a value for your car online in just minutes, and then we'll come to your house and pick up the car. It's an incredible experience. Customers are absolutely. It's resonating with customers. They're absolutely adopting it. We think there's a lot more room to continue to grow awareness of that particular form of customer sourcing from a brand perspective and also improve the customer experience, just like we believe there's opportunity to improve the customer experience everywhere. Outside of that, we source cars in the wholesale market, for example, from auction. The acquisition of ADESA hasn't meaningfully changed our approach to sourcing. We always acquired cars through a variety of auctions prior to the acquisition of ADESA.
That's continued to be the case for those cars that we do acquire in the wholesale market. We continue to acquire them from a variety of sources. There hasn't been really a meaningful change to our approach there before or after the ADESA acquisition.
Okay. Okay. And then just on the retail margin portion of GPU, I mean, I think you talk a lot about sort of kind of subcomponents of, I think, non-vehicle retail or cost of revenue, I should say, that have improved. And yet, I think every quarter, you and Ernie get on the call and talk about more opportunities. Where are those opportunities you can still squeeze sort of more juice out of?
Sure. Yeah. So the two big, other than acquiring the car, the two big retail costs of sales are reconditioning the car. That's the labor, parts, and any third-party vendor services that are required to get the car up to Carvana-certified standards. That's the largest component. There's also a much smaller component that is the inbound transport of getting the car from the customer or an auction into the nearest inspection and reconditioning center. So we have made significant gains there. We think those gains look even bigger on an inflation-adjusted basis where we really think we've outperformed the rest of the industry in fighting off inflation through all of our efficiency initiatives. But having said that, we do think there's more room for gain. I think there's still enhancements that we could make to our software.
I think our teams are constantly working to make improvements in parts procurement to get the best, the most cost-effective parts that we can. I think there's leverage opportunities in continuing to scale through the inspection and reconditioning centers, and I think there's also logistics network opportunities on inbound transport as well, so it's really our gains over the past two years have been broad-based, but we do see more gains and would expect those to be fairly broad-based as well.
And so given that, and like I said, it seems like every quarter, right, we hear that message of like, "Hey, there's more to go, there's more to go." Is that instructing or affecting having you guys throttle any of your demand at this point? Meaning, do you want to get it to a level where you feel like, "Hey, we're finally really optimized, maximizing from a GPU efficiency standpoint, let's go on demand"? Or do you kind of view it more as like, at this point, you've got enough in place in terms of the operations now, operational efficiency, and you're kind of letting both things go at the same time? Which of those would you say it is?
Sure. Yeah. So we do view ourselves in a transition phase as it relates to efficiency focus versus growth focus. I think one way to think about that is just where are our teams spending time and what projects are they prioritizing? So in 2022 and 2023, we're highly efficiency-focused. Every team was focused on making fundamental gains, reducing costs, increasing GPU. And that was almost a singular focus for a large portion of that period. Now I'd say we're in a transition phase where many of our projects and much of our time is still allocated toward efficiency initiatives and driving further fundamental gains, but some of our time has shifted over toward initiatives that are more growth-oriented. But it's only a portion. We were still allocating a large portion of time to fundamental gains. Why are we doing that?
You kind of pointed to it in your question, but we think that improving efficiency has a lot of value for future growth and future scalability that comes along a couple of dimensions. One, making fundamental gains in unit economics just gives you more unit economics to either flow through to margin or pass back to customers as part of your longer-term growth. Also, most efficiency initiatives also make your business more scalable because you need fewer people to execute a transaction. You've automated more. You're using technology more, things like that, which make you more scalable for the long run as well.
Got it. Okay. And then lastly on GPU, is there anything going on under the hood right now, whether it's in the financing component or any parts of GPU really that are maybe a little bit of an anomaly or benefiting from something unique at the moment in the cycle that we're in? Anything going on there that's maybe a little bit better than you might see sort of under normal circumstances?
So not that we're aware of. I think we've had high GPUs very steadily now, quarter after quarter. If you look out across the industry, I don't think we're not hearing people talking about it's a good environment. That might be something that we heard more in, call it 2020 or 2021. That's not the case today. Industry sales are down relative to 2019 levels. Interest rates are higher than at least their pre-COVID short-term interest rates, which are what matters for auto loan payments. Things like that are higher than their pre-COVID levels. So there's not a lot of talk out there about it being a particularly good environment, and we're not seeing anything either that says the environment is particularly favorable.
Yeah. Yeah. Got it. Okay, and then just, again, several years ago, we used to talk about, we used to dig into the cohort curves, right, and we kind of knew the markets you'd rolled out in years one through seven or eight or whatever of the company, and it was just, it was an incredibly predictable set of lines, right? We don't talk about those anymore. Obviously, COVID sort of changed everything, but as we think about the cohort activity you've gotten back to seeing, maybe just talk about, has that predictability generally returned, and then secondarily, you used to talk about spending sort of market by market, right? The brand was new. You had to sort of grow it up in each market. Eventually, you would get to some more national campaigns. We'll set Pickleball aside for a sec.
Yeah, sure.
What can you do with advertising here as you look to maybe stoke demand here going forward at some point?
Yeah, sure. Yeah, so I think individual markets, I would say a lot of the general patterns that we have described in the past still hold. Older markets tend to have higher market shares than newer markets. I think there's a few different reasons for that. One is they've had more time to accumulate awareness, accumulate word of mouth, repeat customers. The stacking effect or long-term layering effect of advertising has had more time to raise awareness of our brand, so I think that's one driver. Older markets also tend to be closer to inventory pools and have more close-by selection available to customers, and so that's historically been a benefit. I think integrating more desk locations, having production capacity and inventory pools in more locations will bring more inventory closer to customers in more and more markets, so we think that's a benefit that I've obviously pointed to before.
And then in terms of advertising, we have a very broad base of advertising. I think a lot of it is national today. A lot of it is brand today. We're big believers in brand advertising. We think some of the long-term market share gains that we've seen in individual markets and also our steady reductions in advertising expense per car over time are aided by brand advertising, raising awareness of Carvana's offering, and just letting more and more customers know what we're bringing to the table. I do think we're still in the relatively early stages of really making customers aware of all of the benefits that come from shopping on Carvana, buying a car on Carvana relative to the traditional model.
And one follow-up there, this is random. When you think about the markets, Atlanta is the easy one, but Phoenix, I don't know, Nashville, Charlotte, places you were sort of early on. When you over-index in those markets, market share-wise, right? Because that, again, like you guys have shown that to be true at times over the years. What's the characteristic of that market as you start to way over-index? So going from, say, like low single digits to approaching mid-single digits, for example. Is it just dealers going out of business or you just stealing volume? What's the characteristic there that's enabling the share gains or reflective of the share gains?
Yeah. So we don't think it's the competitive environment. Obviously, the market's highly fragmented, and I don't think we see a major impact of competitive dynamics. I think the two big things that I would point to is time that has passed for awareness to accumulate and us to have the word of mouth and repeat customer engines as well as just accumulated brand awareness accumulation. That's one driver. The second driver is proximity to larger inventory pools, and I think our older markets tend to be closer to more inventory pools and larger inventory pools based on the way we sort of built the business out starting in Atlanta and then moving out from there over time.
Got it. Okay. And then just lastly, a cash flow question. Obviously, the EBITDA has exploded here recently. Obviously, as you ramp stuff up, you're going to have to make some investments here and there. What are sort of the abnormalities we might see in terms of the cash flow statement, operating cash flows, etc., as you ramp the business up? I think everybody can understand that, okay, the EBITDA margins are expanding and you guys have been doing that really measurably and mindfully. Anything that may come up quarter to quarter from a cash outflow that we should be thinking about that's just a function of sort of ramping things up?
Sure. Yeah.
Why wouldn't the cash flow sort of match EBITDA to some degree?
Yeah. Sure. Yeah. So I think I would start by saying I think one thing that excites me about our business from a financial perspective is our Adjusted EBITDA is very high quality. The flow-through of Adjusted EBITDA to, in particular, GAAP Operating Income is very strong. And we also think there's significant leverage in that over time. Even our flow-through that we're seeing today of Adjusted EBITDA to GAAP Operating Income, we expect that to go up over time because a very large portion of our non-cash expenses are heavily overhead or heavily fixed. So I think that's a very strong aspect of our business model. I also think it's starting to lead to, for example, Operating ROA. That looks pretty strong, right?
Over last two quarters, we generated a run rate of about $1.2 billion of operating income if you just sum the last two quarters and double it. And that's on a base of between $6.5 billion and $7 billion of total assets. So those aren't quite the very, very best, but we're doing that while also having significant infrastructure to scale into. So we have more assets than we need to generate this level of operating income. That's a little bit of a tangent, but that's at least one thing that I get excited about when looking at our financial model and how it could scale over time. On the cash flow front then, I think the most notable thing to keep in mind is that inventory growth shows up in cash flows from operating activity.
As we grow, we do invest in inventory, and that shows up in cash flows from operating activities. Now, I would say that's also true for all auto dealers, and particularly used auto dealers. There's also efficient sources of financing for that. As we grow inventory, we can either fund it with cash flows from operations or we have access to floor plan financing, which is the way the rest of the industry fuels inventory. It actually significantly limits your cash outlay.
Yeah. Got it. Didn't leave a lot of time. Any questions from the audience?
Yeah.
Oh, Greg. I don't remember that stat .
Oh, you got a mic.
Yeah.
Sorry, you talked about awareness. I think I remember a stat you guys had shared back in the day with unaided brand awareness was somewhere in like the teens. I could be wrong, but sort of vaguely remember that. Has that grown over time? And then maybe talk about the last year or two if that growth has kind of contributed to market cohort growth as well.
Sure. Yeah. So I do think we have increased our awareness over time, but there's still a lot of opportunity to add to awareness from where we stand today. And one example of that, so we do see in survey data, one of the reasons why customers don't buy a car from Carvana is, "Oh, I didn't even know that was an option." It may feel surprising to everyone in this room that not everybody knows about this, but not everyone follows the financial press. Not everyone is necessarily in the weeds of the Carvana story. And so there's actually a lot of opportunity to continue to raise awareness of our brand. So we do see that as a meaningful opportunity over time. And we think brand advertising plays a role there. Earned media plays a role there. Continuing to drive word of mouth plays a role there.
But we do see a lot of opportunity in that area. As it relates back to cohorts, I just echo what I said to Brad earlier. Generally speaking, the overall patterns that we've seen in cohorts persist. And then for the most part, cohort shares have followed along with the company more or less. And so we're still continuing to see the same patterns.
So unfortunately, we are out of time, Mark.
I think that's it.
Thank you very much for being here. That was great.
Thank you, Brad.