Ernie Garcia - Chief Executive Officer Mark Jenkins - Chief Financial Officer Meg Kehan - Investor Relations

Chris Pierce - Needham Michael Baker - D.A. Davidson Seth Basham - Wedbush Securities Ron Josey - Citi Rajat Gupta - J.P. Morgan Chris Bottiglieri - BNP Paribas Michael Montani - Evercore ISI Daniel Imbro - Stephens

Good day, and welcome to the Carvana, Third Quarter 2023 Earnings Conference Call. All participants will be in listen-only mode. [Operator Instructions]. Please note, this event is being recorded. I would now like to turn the conference over to Meg Kehan, Investor Relations. Please go ahead.

Thank you, Varsh. Good afternoon, ladies and gentlemen, and thank you for joining us on Carvana's third quarter 2023 earnings conference call. Please note that this call will be simultaneously webcast on the Investor Relations section of the company's corporate website at investors.carvana.com. The third quarter shareholder letter is also posted to the IR website. Additionally, we posted two sets of supplemental financial tables for Q3, which can be found on the Events and Presentations page of our IR website. Joining me on the call today are Ernie Garcia, Chief Executive Officer; and Mark Jenkins, Chief Financial Officer. Before we start, I would 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 results to differ materially from those discussed here. A detailed discussion of the material factors that cause actual results to differ from forward-looking statements can be found in the Risk Factors section of Carvana's most recent Form 10-K and Form 10-Q. The forward-looking statements and risks in this conference call 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. Our commentary today will include non-GAAP financial metrics. Unless otherwise specified, all references to GPU and SG&A will be to the non-GAAP metrics and all references to EBITDA will be to adjusted EBITDA. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our shareholder letter issued today, a copy of which can be found on our IR website. With that said, I'd like to turn the call over to Ernie Garcia. Ernie?

Thanks, Meg, and thanks everyone for joining the call. The third quarter was another great quarter for Carvana that continued the clear momentum we've had since we rolled out our plan 18 months ago. Over that period of time we have cut $1.2 billion in annualized SG&A expenses out of the business, which has dramatically improved our efficiency, while also reducing annualized inbound transport and reconditioning COGs expenses by about $250 million annualized, which has driven up our GPU. These efforts have combined to powerful effect. Over the last two quarters, we have generated $300 million of adjusted EBITDA, with about $200 million of that excluding non-recurring items. These are important numbers. They send a clear message that the plan our team laid out 18 months ago was the right one. They send a clear message that our team is executing at a very high level. And most importantly, they send a clear message about the long-term power of our business model. Since we went public in 2017, we have been clear about our goals to sell millions of cars and to become the largest and most profitable automotive retailer. Two years ago, there was a general belief in our ability to become the largest, but the questions remained about our ability to be the most profitable. While those two questions won't be definitively answered until we actually cross both milestones, the last two quarters provide a lot of support for our belief that we will ultimately be the most profitable automotive retailer as well. In today's shareholder letter, we are providing clear evidence of this with our breakouts of our overhead and operational expenses over time. The takeaways from the data are the following: One, in Step 1, we successfully right-sized the operational components of the business and began making significant progress in reducing our operational expenses per unit. Two, our efficiency gains are continuing in Step 2, with our unit economics rapidly becoming a significant competitive advantage. And three, our current fixed-cost utilization is low, providing a very clear path to both significant scale and significant cost leverage when we turn to Step 3 of our plan. Our plan is working, and we are going to see it through. While it is unquestionably tempting to turn our attention back to growth, we believe it is a better long-term decision to remain disciplined and continue harvesting the efficiency gains we see in front of us before shifting our organizational focus and moving on to Step 3. In the fourth quarter, we expect to continue making fundamental gains through various projects that comprise Step 2. We also expect volumes to reduce relative to Q3 due to normal seasonality and potential industry softening observed over the last four weeks. Despite this, we continue to operate in our target unit range for Step 2. The sum of these dynamics leads us to expect to produce another great quarter, but as in past Q4s, we expect lower industry-wide volume, coupled with higher industry-wide depreciation rates to cause lower adjusted EBITDA than we have achieved in the past two quarters. Looking forward, we are excited. The difficulty of 2022 wasn't how we imagined our path in the beginning, but that doesn't mean it wasn't good for us. Through it, we have come together to do our best work. We have learned the power of focus. We have gained a new appreciation for discipline. We have learned an even greater sense of urgency. And now, as a result, we have competitively differentiated unit economics. 2022 has added to our arsenal. It has made us better. And while that is a change, there is a lot that remains exactly the same. We are still an ambitious group of people who care. We are still willing to roll up our sleeves and do the hard work to achieve our goals. We are still delivering new experiences to people buying and selling cars that they love. We still have a business model that is incredibly difficult to replicate and that mechanically gets better as it gets bigger. We are still just a small player in the largest retail market in the world, and we are still on the path to selling millions of cars and becoming the largest and most profitable automotive retailer. The march continues. Mark.

Thank you, Ernie, and thank you all for joining us today. Our third quarter highlighted the significant and sustainable progress we have made on our path to profitability. We set third quarter company records for total GPU and adjusted EBITDA, both with and without nonrecurring benefits. As part of our earnings materials, we provide a detailed look into our Q3 results, our three-step plan, and the components of our cost structure. I'll start by summarizing three key takeaways. First, in the last two quarters, we have clearly demonstrated the profitability of our online business model. In Q2 and Q3, we generated more than $300 million of adjusted EBITDA, which includes approximately $110 million of non-recurring items. We did this despite carrying costs of an infrastructure that supports 3x retail unit growth, and despite a difficult used vehicle industry backdrop. Second, our focus on driving fundamental operating efficiency in Step 2 is generating significant unit economics gains. Of particular note, we have reduced non-vehicle retail cost of sales and operations expenses by $1,000 per retail unit in just the last two quarters, and we see opportunities for further gains from here. Third, we have significant excess capacity in our existing infrastructure to support multiples of profitable growth. We expect this growth to be paired with significant operating leverage as we leverage our underutilized overhead costs. Moving to our third quarter results. In the third quarter, retail units sold totaled 80,987, a decrease of 21% year-over-year and an increase of 6% sequentially. Total revenue was $2.773 billion, a decrease of 18% year-over-year and a decrease of 7% sequentially. As we've previously discussed, our long-term financial goal is to generate significant GAAP net income and free cash flow. In service of this goal, in the near term, our management team remains focused on driving progress on a set of key non-GAAP financial metrics that are inputs into this long-term goal, including non-GAAP gross profit, non-GAAP SG&A expense, and adjusted EBITDA. Due to the dynamic nature of the current environment, we will focus our remaining remarks on sequential changes in these metrics. In the third quarter, non-GAAP total GPU was $6,396, a sequential decrease of $634 driven primarily by smaller benefits from non-recurring items. Total GPU in Q3 was positively impacted by approximately $500 of non-recurring benefits, which we describe in more detail below, compared to approximately $900 of benefits in Q2. Non-GAAP retail GPU was $2,877 versus $2,862 in Q2. Sequential changes in retail GPU were positively impacted by a $250 reduction in non-vehicle retail cost of sales and a 25 day reduction in retail average days of sale, both at the favorable end of our previous outlook range, as well as wider spreads between retail and wholesale market prices, partially offset by higher retail depreciation rates and a smaller inventory allowance adjustment benefit compared to Q2. Notably, our strengthened retail GPU continues to be driven by several fundamental improvements in our business compared to our previous high watermark of FY 2021, including lower reconditioning and inbound transport costs, a higher customer sourcing rate, and higher revenue from additional services. Non-GAAP wholesale GPU was $951, in line with our outlook, versus $1,228 in Q2. Sequential changes in wholesale GPU were primarily driven by higher wholesale market depreciation rates in Q3 compared to Q2, which negatively impacted wholesale vehicle volume and gross profit for wholesale units sold, as well as seasonal changes in wholesale marketplace volume. Non-GAAP other GPU was $2,568 versus $2,940 in Q2. We estimate that a higher than normalized volume of loans held and sold in Q3 increased other GPU by approximately $400, other things being equal, compared to a $650 benefit in Q2. In addition, other GPU in Q3 was primarily impacted by lower origination rates relative to benchmark interest rates, partially offset by lower credit spreads. In Q3, we continue to make progress lowering SG&A expenses, reducing non-GAAP SG&A expense by $13 million sequentially. Notably, we reduced non-GAAP SG&A expense per retail unit by more than $400 sequentially, while growing retail units sold by 6%. In our shareholder letter and accompanying materials, we provide additional details on the components of our SG&A expense, including a breakdown of operations expenses, which are more variable in nature, and overhead expenses, which are more fixed in nature. Our significant sequential operating leverage in Q3 was driven by both continued improvement in operational expenses, as well as leverage in the fixed component of our cost structure. Adjusted EBITDA in Q3 was positive $148 million or 5.3% of revenue. The aggregate impact to adjusted EBITDA of the previously described nonrecurring items was approximately $40 million. On September 1, 2023, we closed the previously announced corporate debt exchange offer with 96.4% of note holders agreeing to exchange $5.52 billion of our unsecured notes for cash and new secured notes, reducing total debt by over $1.325 billion, extending maturities and decreasing required cash interest payments by more than $455 million per year for the next two years. On September 30, we had approximately $3.2 billion in total liquidity resources, including $1.6 billion in cash and availability under revolving facilities and $1.6 billion in secured debt capacity and unpledged beneficial interests. Turning now to our fourth quarter outlook. While the macroeconomic and industry environment continues to be uncertain, looking toward the fourth quarter of 2023, we expect the following as long as the environment remains stable. A sequential decline in retail units sold, driven primarily by industry and seasonal patterns, non-GAAP total GPU above $5,000 for the third consecutive quarter, and positive adjusted EBITDA for the third consecutive quarter. Looking toward 2024, we expect to drive significant total GPU and adjusted EBITDA for the second consecutive year. We are excited about the path we are on, and we look forward to making continued progress toward our goal of becoming the largest and most profitable auto retailer. Thank you for your attention. We'll now take questions.

We will now begin the question-and-answer session. [Operator Instructions]. Our first question comes from Chris Pierce with Needham. Please go ahead.

Hey, how you doing? Non-GAAP – hold on a sec. Non-GAAP SG&A per unit. Can it continue to go down in Q4 on the lower number of units you're guiding for? And if it can't, would that be a signal internally to turn on growth?

Sure. I would say you can see that we've pushed down our dollar spend pretty consistently over the last several quarters. I think we're extremely happy with the dollar reduction this last quarter that we saw. It came in the form of basically a $200 operational expense saving per unit, and kind of that being offset by roughly 5,000 additional units quarter-over-quarter. So we're clearly making pretty significant gains there. I think that we absolutely believe that through Step 2 we have additional gains to be made. We expect those gains to be across the board, both in our corporate expenses and in operational expenses. But I think as far as kind of that detail of disclosure, we're going to stick with the guidance that we've provided.

Okay. And then just broadly, what would turning on growth mean? Because you have to get the cars to sell the cars. So is that something – or would turning on growth be increasing advertising to sell more cars in different geographies at a higher rate or would it be sourcing a greater number of cars from consumers? I'm just kind of curious how you would go about it to the extent you've had those conversations.

Sure. So let me start with – I think roughly a year and half since we kicked of with this plan. We internally called that Project Catapult. And basically everyone across the business had a series of goals in every operational group that they were aiming to hit, many of them with monthly targets and most of them with weekly targets that we were marching to basically through Q2 of this year. We also then kicked off what we call Operation 100, which is the next series of our goals across every single one of our operational groups. That also has very clear targets, some of them monthly and most of them weekly, through Q2 of next year. I think that has clearly led to tremendous gains. That level of focus the discipline that we've had, of making sure that we are maintaining operational excellence and that across all locations in the company we're performing at the best level of any locations across the company. I think it's clearly led to tremendous benefits and I think that focus has been very, very valuable. So I think the most important thing that we will do is in time to turn to growth and Step 3 is basically to just reset the entire company focus on growth goals. So I think instead of everyone being focused on, how am I going to drive additional efficiency and all those targets, being focused on how are we going to get more efficient as a business, they'll instead be focused on how do we make customer experiences even better? How do we drive growth? How do we position the company to grow? And so I think that is the number one thing that we will switch when we move to Step 3 from Step 2. I think the gains that we're seeing are still very significant and so we believe it makes sense to stick in Step 2 for a bit longer. I think when it is time to grow, I think advertising, inventory are absolutely two very big parts of that equation. I think we grew from 2018 Q1 through ‘21 Q1 at approximately a 70% CAGR. During that entire time, we were clearly growing inventory pretty quickly. That was leading to increased conversion. We are growing advertising pretty quickly. I think that was partially because those customers converted even more efficiently as a result of that extra inventory and our brand was building. And so we benefited from positive feedback throughout that entire period. Over the last 18 months, the opposite has been true. Over the last 12 months, we've shrunk inventory by 50% and we've shrunk advertising by 50% and that's clearly pushed a bit in the opposite direction. But when it's time to grow, we expect to again grow inventory. We expect to again grow advertising. I think there are other elements of positive feedback that we expect to kick off as well, and we think that that will drive a big part of that growth.

Okay, thank you.

The next question comes from Michael Baker with D.A. Davidson. Please go ahead.

Hi, thanks. So you said a lower cost structure is a competitive advantage. How does that show up? To me, that means you can get aggressive in price to win back shares. Is that the right way to think about it or is there another way we should interpret a lower cost structure being a competitive advantage? And maybe just as a second question, I'd like to call it a follow-up, but it's unrelated. I also heard you say the last four weeks have been weaker. If you could just talk about the trends through the quarter and early into the fourth quarter. Thanks.

Sure, so let me start with that. I think unit economics we view as a competitive advantage. So I think I would extend that not just to cost structure, which we'll hit in a moment, but also to GPU. I think a huge advantage of our business model across all of time has been our vertical integration. We're able to give customers a very high quality, very simple experience, and we're able to monetize much more of that experience than many others in the industry as a result of that vertical integration. And I think those benefits are showing up very clearly in our GPU. We just had record Q3 GPU on the order of $6,000. We had record Q2 GPU as well. So I think unit economics start there and we're able to achieve those GPUs despite giving our customers significant discounts versus many of the largest players in this industry. So I think that's the starting point, I think, of the power of our model. I think now if we kind of fold that into expenses, you can see the gains that we've made in operational expenses. We're extremely proud of that. That's obviously not a super easy thing to do, especially while holding volumes flat. We've dropped our operational expenses by $200 from Q2 to Q3. We dropped them by $200 from Q1 to Q2 before that. And we're now at a place where those operational expenses are actually lower than where they were in Q1, ‘21, and lower by about $100, give or take, versus Q1, ‘21. That's good, but it doesn't sound like it's incredible progress, but I think it's much better when you take into account the fact that there's clearly been quite a bit of inflation over that time. I think based on Bureau of Labor Statistics data, labor on average has been up by approximately 11% over that period. When we do the best job that we can to compare our progress to that of other automotive retailers, generally speaking, we're seeing several hundred dollars of increases in SG&A per unit over that time period. And so I think that the gains that we've made are pretty significant. And then the sum of those things, I think, ultimately adds up to our unit economics. And then we do think that puts us in an incredibly powerful position. That gives us a lot of options in the future. I'm not sure that we are exactly sure the ways in which we will express those options, but it undoubtedly gives us a lot of options. On your second question, I would just say, the last several years have been very unique. I think there's been a ton of distortions in the market that have made things a little bit harder to kind of precisely forecast. Last year we saw rates going up quite a bit. In ‘21 we saw in the fourth quarter in ‘21 we saw prices going up quite a bit. So, I think it's a little bit hard to know exactly what to expect out of seasonality. I do think over the last four weeks, given all the data sources that we have, we can see kind of industry-wide data. It looks like things have probably been a bit softer, not to a degree that I think is unique or merits a ton of attention, but definitely that's been the direction. And so we'll see where that goes. Like I said, I think it's a little bit hard to even figure out exactly what seasonality should be right now given all the changes we've seen over the last couple of years.

Great. Thank you. Fair enough. I appreciate the color.

Our next question comes from Seth Basham with Wedbush Securities. Please go ahead.

Thanks a lot, and good afternoon. So my question is, when you pivot to growth for the first couple quarters that you do that, do you think that your leverage – your excess capacity will more than offset the cost to support that growth in terms of advertising, staffing, vehicle acquisition and operational inefficiencies that could result?

So let me start with just pointing to history. Well, let me start with our expectations. Our expectations are that we expect our operational expenses to go down from here over time and we think they can go down despite growing over time and we expect our corporate expenses per unit to absolutely go down as we grow. So those are our expectations. In the data we provided, we provided a chart that kind of breaks out our operational expenses per unit going back in time all the way to Q1, 2018. And I do think looking at the period from Q1, 2018 through Q1, 2021 is somewhat useful. During that period, we were supporting 70% annualized growth continually. We were adding many, many markets. We were adding many, many inspection centers. We were building the machine of Carvana as we were going and we did not have excess capacity to grow into and we were able to hold those expenses roughly flat. I think if you look at them, you'll see variation to the tune of a couple hundred dollars. There's kind of an outlier data point as COVID hit. But in general, they were relatively flat. And then you'll also see that there were even some gains as we headed into the better seasonal parts of the year as well. So I think that's pretty clear evidence that we've done this before. We've supported very high rates of growth and we've grown in a way that is more difficult, because we weren't growing into existing infrastructure. We were building infrastructure as we went. So we don't think that it's obvious that there need to be meaningful expenses or expense increases in these line items as we turn to growth. And overall, we absolutely expect there to be significant leverage. I think that's very clear in our GPUs and very clear in our operational expenses. When you look at the gap between those two things, there's a lot of room for growth to be extremely beneficial financially.

If that's the case, then I respect the decision to continue focusing on operational efficiency, but you're leaving a lot of opportunity on the table. So why not pivot to growth sooner rather than later?

I think financially in the near term, that would likely be the right choice. I think the scale of the unit economic benefits that would come from growth, I think that likely would be the right choice over the next couple of quarters, however kind of long we're headed into that. I think when you look at it over a longer period of time, there is a fixed cost to reorienting the entire business. I walked you through Project Catapult, which we kicked off 18 months ago, and Operation 100, which we kicked off approximately seven months ago. Those are long-term projects with many underlying goals where the entire company is focused on driving different metrics in the right direction. And that focus has been tremendously beneficial, and we believe there are gains yet to be had. And so I think that points in this direction of discipline suggests let's keep getting those gains. I don't mean to be defensive on this one, but I think if we go back in time even nine months ago, I think the average expectation consensus was on the order of $500 million of negative EBITDA this year. Well, it was kind of the expectation that people had for us, and over the last two quarters, it's been positive $300 million. That's obviously a tremendous difference, and I think that demonstrates the kind of value of that discipline, of setting clear goals and of marching toward them. But again, I think basically on math, I think the answer is clear. We should turn to growth. We just think that on math over a longer period of time, we should harvest the gains that we're in position to get, and then we should turn to growth. And when we do, we'll be even more efficient than we would be today with even lower costs.

Thank you.

Our next question comes from Ron Josey with Citi. Please go ahead.

Thanks for taking the question, guys. Always good to hear from you. I wanted to ask too, please, Mark, in the presentation of the investor letter, the per unit insights were super helpful. As we think about, staying in Step 2 or when do we move to Step 3. I want to ask just about specifically I think we saw inbound transport costs declined $900 over the last 12 months on a per unit basis. So we'd love to hear what drove those gains and how far away do you think we are to getting to that, like peak efficiency, maybe not just with just inbound transport costs, but just overall. I assume there's more ways to go. But we'd love your thoughts there. And then, Ernie, as we think about, preparing ourselves for whenever the next step is, same day delivery is something that's super interesting. We'd love to hear just early results for now and I think two or three states and cities. We'd love to hear any insights on how that might be driving greater conversion rates and really just how the network is positioned to support same day delivery. Thank you.

Sure. Ron, well, let me start with your first question. So I think what you're referring to is, in the last 12 months we have reduced our reconditioning plus inbound transport costs by about $900. I think that's a very significant win and one of the progress points that we're very excited about when we're thinking about the progress that we're making in Step 2. I think what were some of the drivers of that? So I think if we start focusing on reconditioning, in-sourcing was probably the single biggest driver. So basically taking services that had previously been provided by third parties in the inspection and reconditioning centers and taking them in-house and doing them ourselves, and that obviously saves costs and allows to have better control of the process. So I think that's been a big win. In addition, I think we've gotten better at standardizing processes, making sure we're properly normalized at our staffing levels. We've made some big gains in proprietary inspection and reconditioning center software, including fully in-sourcing our inventory management system. I think that's a big win and a big improvement that allows us much better control of the process and allows us to really have, again, clear standards and processes that are executed on every single car that's rolling through the facilities. We've also made gains automating parts procurement. It used to be a highly manual process and a highly more variable process than it is today. So I think we've made big wins there in the proprietary software development department. So those are a few of the places where we've made real gains in reconditioning. I would note that we do see opportunities for further gains in those areas that I just pointed to in recon. We don't think we're done. We do see opportunities for further gains in reconditioning costs per unit. Moving to inbound transport, I think a couple of the sources of gains there are, getting better at logistics network utilization, as well as having lower inbound miles. And so both of those things I think have also led to declines in inbound transport costs. Now, one thing I did want to call out is there's actually two $900 improvements that are of note. I just talked about, and I think your question was focused on the $900 gains in retail cost of sales. However, we also have an additional $900 of significant gains in our operational selling general and administrative expenses. And that's, we've broken out some detail on that in our shareholder letter as well as the accompanying materials. And so in total, that's $1,800 of total gains in those more variable cost components. If I just say a few words about the gains. That second $900 of cost gains, I think, that has also been very broad based. So that would include cost improvements, for example, in our customer care centers. It would include cost improvements in our market operations, which is our last mile delivery network. It would include cost gains in our logistics and transportation departments, which house the logistics network that moves cars around for us. And we've made very significant gains in some of those operational areas as well. Those have been driven by process automation, improved staffing and scheduling, better proprietary logistics route management, and activity pairing software. We've made some real improvements, both in terms of processes and technology, that have driven that second $900 of unit economics improvement, and that one being since Q4 2022. So I think obviously we're very excited about all that. $1,800 in reductions on a per unit basis in the more variable components of the cost structure, I think is something we feel really great about and is part of the reason that we're so excited about what we're doing here in Step 2, before we transition to take advantage of our excess capacity and return to profitable growth.

Yeah, and I'm going to jump on that for a second as well before heading to same-day delivery. What I would also say is, those gains are very high-quality gains. Those aren't cutting corner gains. During this period we've seen warranty claims, for example, go down. So we've seen frequency of warranty claims go down, while we've been driving down reconditioning expenses. A lot of the cost savings that Mark is talking about are actually making the customer experience better. The scheduling benefits in logistics means that we're delivering cars faster, we have fewer pushes. When customers are calling advocates, calls per sale are down approximately a third year-over-year. We've seen calls per advocate – sorry, sales per advocate have doubled year-over-year. We're seeing benefits across the board in title and registration. We've seen very significant changes as well. 75% increase in efficiency per hour of work in title registration. Year-over-year, 15% increase quarter-over-quarter. So a lot of these things are also just making the customer experience better. And during that period, we've seen NPS go up as a result. So I think these are deep, fundamental, real gains. They are not tradeoff gains and so we're extremely excited about them. Heading to same-day delivery, I think same-day delivery is a very good example of one of these projects internally that we've been working on, that is sort of a little fuzzy in terms of what is the ultimate goal. Is it Step two or Step three? And clearly it can play a role in both. The way that we've been executing that to-date, first of all, as you said it's in a small subset of markets. So I don't want to get everyone too excited about this one just yet. But the way we've been executing it to-date, has more been aimed at efficiency gains. There are times when customers change their delivery date or they cancel a spot that they've previously booked and were not able to book that for another customer. By adding same-day delivery, we can kind of fill in those empty slots. And so we can offer faster delivery to our customers, and we also can do it more efficiently, because we fill in time that would have otherwise been wasted. And so that's generally how we've rolled that out. We're also just beginning to test same-day purchasing from customers as well, where they can go to our website, they can get a value for their car, they can bring the car to us. We've actually had customers complete that process on the order of one hour from kind of being on the website to getting a check from us as they dropped off their car at one of our vending machines. So we are working on those things. I think those are exciting things. Today they're aimed at efficiency. Tomorrow they certainly could be aimed at growth. In order to do that, I think it would just have some implications for our staffing models. I think whenever you're trying to offer very rapid delivery times, the fundamental tradeoff there is always just between sort of efficiency and kind of excess capacity and time. But we don't expect those costs to be super high, but they will require focus, and we'll decide again how to use those fundamental gains as we head into step three.

Thank you, Mark. Thank you, Ernie.

Thank you.

The next question comes from Rajat Gupta with J.P. Morgan. Please go ahead.

Great. Thanks for taking the questions. The third quarter units saw a bigger pickup than I guess what you had guided to, what you were expecting. It seems it was a little better than the normal seasonality the industry experiences. So I'm curious, like, what drove the pickup there? Was it just aided by the industry or anything you were experimenting internally? As you think about moving from Step 2 to Step 3, and why is that a reversing here in the fourth quarter? Is it, again, like a refocus on the operations or is it just more purely just industry pressures that you're facing? And I have a quick follow-up. Thanks.

I think the simplest way to think about Step 2 is that during Step 2 we're effectively setting up the company to be kind of neutral to industry volume changes, whether that's seasonality or otherwise. And so I think, Q3 was great from a unit perspective. I think that was largely driven by seasonality. I don't think there was a ton of specific stuff that we were doing that was aimed at growth. In the quarter there may have been some benefits from some of the other projects we had, but there weren't gains that were aimed at growth. I think heading into Q4, we again expect to kind of roughly move with industry, because we are remaining in Step 2 and we're keeping the settings of the business fairly neutral. So I think normally there's seasonal decrease in Q4. We expect that to occur. Again, I think most importantly, kind of the unit range that we expect is very much in line with what we've seen over the last several quarters. And we think that that's a reasonable unit range for us to continue to make gains as we power through step two and get ready for step three.

Understood. That's helpful. And just a quick one on just the financing and the ABS markets. We have seen spreads widen a bit, benchmark rates have gone up. How does that change your thinking around monetization in the fourth quarter? I believe you still have a bit of a backlog, from earlier this year that needs to be cleared. You talked about holding on to more residuals going forward. I mean, is that something, we should expect to start seeing this quarter already? And just maybe like just broader thoughts on how we should think about the finance or other GPU from Q3 to Q4. Thanks.

Sure. I think there are – certainly all the changes you pointed to are reasonable. There's been some kind of choppiness, I think across all financial markets over the last many months, and I think that that's shown up in benchmarks, its shown up in spreads. As we have in the past, we would expect others to be passing that on to the consumers, and we would expect to do the same. So generally speaking, I don't think that we have a tremendously different expectation there. I think in terms of our general monetization plan, I think that first order, our plans remain the same. The plan that we've had in place has been very good for a very long period of time. Even in this last quarter we had four securitizations that we were pretty happy with. I think we still have room to improve our effective cost of funds, those securitizations, and I think if we were kind of achieving industry best cost of funds there, there would be another kind of area where we would have significant improvements in our unit economics, and over time we absolutely expect that to be the case. I think we also are clearly in a different capital position than we've been in quite a while. Most importantly, the improvement in that capital position is led by the gains that we're seeing across the business that led to $300 million of EBIT dollars over the last two quarters. And then we're in a good liquidity position with the business operating the way that it is. So I think that we are now communicating that we will be more open to maintaining residuals in the future. We'll see exactly what forms that takes over time. I think there are lots of forms that can take, but we view that as an opportunity as well. In general, they are very high-quality assets, these auto loans we generate, and the residuals offer, pretty significant returns that are very robust to increases in expected losses. And I think it's a tried and true strategy for many in the industry to maintain those residuals and get paid pretty handsomely for it. So we think that's an interesting opportunity. We'll start to explore. We would like to maintain the right to do that flexibly over time, but that may be a change.

Understood. Great. Thanks for the call, and good luck.

Thank you.

Our next question comes from Chris Bottiglieri with BNP Paribas. Please go ahead.

Yes, thanks for taking the question. So I just want to – I kind of want to clarify the new disclosure on the auto overhead expense. So your supplemental deck shows its $141 million, excluding DNA its $45 million. Is this largely the same bucket expenses as the $172 of other expenses in the core P&L? And I guess the way you're framing it, it sounds like the vast majority of this $141 million plus the DNA are fixed, and that you need growth to scale. Is that the correct – Is that the correct way to frame it? Like how much of that is actually fixed versus variable? Because when I look at the way you define other expenses in the queue and just commentary given the past, it's IT, it's corporate occupancy, its professional services, it's insurance, its limited warranty, its credit losses, customer swag. Like some of those stuff sounds variable, and some of it sounds like you could cut it if you were structured. So kind of tell me the sense of how you're thinking about that bucket from a fixed versus variable perspective going forward.

Sure. I'll take that one. So I think to hit the question very simply, the overhead expenses of $141 million are not the same as other SG&A expenses that we've historically reported. That new slide on overhead expenses per unit and also in dollar terms is intended to capture the more fixed components of both, compensation and benefits, as well as the more fixed components of other SG&A expenses. And so just to give some sense of what that is in compensation and benefits, that would be your corporate teams, accounting, legal, some of the more general and administrative-oriented corporate teams. It would also be your technology teams. So I think there, product, engineering, data science, those compensation benefits expenses would be included in that new overhead slide that we provided. When you move to the other SG&A component, some of the expenses that underlie our historical other SG&A expense that would then be included in this new overhead expense measure would include facilities expenses. It would include some technology expenses, and then it would also include some non-payroll, other general and administrative expenses. And so that's just to give a sense. And most importantly, with respect to the setup of your question, clarify that that is not – the new overhead expense measure is not a parallel or proxy for that other SG&A. Now, in terms of the question, what component of overhead expenses are fixed and what component are variable? So those expenses we view as primarily fixed. There are some semi-variable components. For example, a few corporate departments may vary somewhat with units. Certain technology expenses may vary somewhat with units. But for the most part, those expenses are heavily fixed, and we think we have the ability to lever those expenses very significantly over time. And just to give a little bit more color on that last point, one of the reasons why we are really excited about the leverage opportunity that we see in front of us is our existing capacity utilization. And so as you probably saw in the shareholder letter or the accompanying materials, we have built and currently own significantly more capacity, whether it's in our inspection and reconditioning centers or our logistics network, even our corporate office or not even corporate, but more sort of customer care center office space. We own significantly more capacity than we are utilizing today. We think that gives us an opportunity to increase retail unit volume significantly, while levering those overhead expenses meaningfully in the process.

I got you. Okay. So it sounds like its very much growth dependent. I mean, kind of leads to like a related follow-up. But I know you're not willing to give guidance at this point for growth, but from our shoes, how do we think about growth? I mean, the market's a little bit depressed, call it 10, 12 points. I think there's probably some supply scarcity that's caused you to lose market share that gets better over the next two, three, four years. But beyond that, like what are the – you have your tangible growth drivers like prime now like delivery, but how do you get to this 40%, 50% to 200% unit growth that you need to lever these fixed costs? Like what gives you comfort that you can get there? And from our shoes, like how do we model that? Like what does that mean for an investor?

So, let me throw you that one. Let me start with this. I mean, I think until we really start to grow, these are all conceptual arguments and everyone has to decide kind of what they believe and what their expectation is, but I'll start with ours. We expect to sell millions of cars. That has been our goal since the beginning. That is our absolute expectation, and I think a question to ask across time is, what has really changed versus the first eight years of our life where we grew very, very quickly? Because I think that it's as hard to answer that question as it is or harder to answer that question than it is to answer how are we going to grow from here? During the early parts of our life, we grew at triple digit rates for a very long time and that took many forms. It took – but I think the most exciting forms were basically a migration of people's preferences toward e-commerce. It was a building of the car model brand. It was the positive feedback inherent in our business model. We are a business model that gets better as it gets bigger, and so it is true that when we grow inventory, selection goes up, conversion goes up. When conversion goes up, your advertising costs go down. You spend more on advertising, you sell more cars. That allows you to carry more inventory and there are several other examples of that as well. So I think those persistent drivers of growth led us to grow by multiples over an eight-year period of time. And then I think over the last year and a half, we've clearly been in a highly distorted environment, and I think that makes it a little bit harder to kind of read through and take the current trend lines and figure out exactly how to extend them. But we don't think anything fundamental has changed, and we think that we put a lot of pressure on ourselves over the last 18 months as we focused on getting more efficient. And just to use those same terms, the primary form that's taken is we've shrunk our inventory by 50%, meaning that our selection of customers has gone down, and we've pulled back on marketing spend by 50%, and those things have fed back negatively. So, I think when it's time to restart the engine, we expect that our offering will be received the same way that it was before. We expect that positive feedback will once again show up, and we expect that some of those things will lead to us selling millions of cars again in the future. I think we've spoken in the past about the distortions of the market that we're in today, and I think that can sound a bit conceptual and kind of unclear. But, I think, our statistic that I think is useful to think about as it relates to, how have things changed over the last 18 months, and is the last 18 months really the timeframe we should be looking at to extrapolate the long-term potential of Carvana, should we be looking at the earlier periods. Pre-pandemic, the average car that we sold was approximately a three-year-old car, and that car was for sale for approximately $19,500. More recently, the average car that we've sold has been 5.7 years old, and it's been sold for $25,000. That's a dramatically different situation, and I think there's many, many reasons for that. One is, there was obviously a huge supply chain disruption. Most fundamental is it was a huge supply chain disruption that led to car prices going up relative to other goods. It remains the case today that for consumers buying cars, they are spending approximately 50% more per month by the same car than they were pre-pandemic. They are spending on the order of 10% more on everything else they are buying. So cars have gotten relatively less affordable, and that of course has a very large impact. Our expectation would be that at some point in the not-too-distant future as supply chains get resolved, cars will probably have a similar cost relative to other goods as they had for a very long period of time before, and that's a very big difference that we would expect. I think it's also useful to note that one of the other reasons why the average car that we're selling today is 5.7 years old versus 3 years old is because an entire portion of supply that we used to have ready access to, which was maybe two portions of supply, off-lease and off-rental, that we used to have ready access to, have largely evaporated because of the distortions that are driven by the price increases over the last year and a half, give or take. And the reason for that is that all the cars that are coming back off-lease today, those were leases that roughly, if we assume the average lease is 36 months, they were leases that were written three years ago. Three years ago, car prices were in a completely different spot, and so the expected value of cars being returned was in a completely different spot. And as those cars are being returned today, every dealer where those cars are landing is very smartly picking those cars up and benefiting handsomely to the tune of thousands of dollars per unit from all those cars that are showing up on their lot. And so basically, off-lease cars are just not flowing through auction today. And because the OEMs have been constrained, there are very few rental cars that have been flowing through auction today. So those are big, big distortions that lead to a meaningful change in the portion of the market that we are accessing, but I don't think that there are strong arguments that we should expect that to persist over any meaningful period of time. And I think there have already been early signs of some normalization. We've seen depreciation pick up a little bit more recently. We would love to see that continue. I think there is nothing that will lead to the normalization of the market more quickly than car prices coming back into line with where they historically were or even close to where they historically were. And so I think we very much look forward to that. But as those distortions unwind, I think the more likely scenario, or at least our belief, and everyone can make up their own mind, is that we're back in the world that we were for eight years before. And in that world, we have in offering customers love. We have a highly differentiated model. We have a ton of infrastructure that we can grow into. We have a business that gets better as it gets bigger. We have a team that is hustling fast and has always hustled fast and has changed the direction of that hustle over the last 18 months, but is ready to change it back, and we're going to go hit our goals. So I think when we look forward, we don't know exactly what the growth rates are going to be. We're not ready yet to provide that. I think it's going to depend on all sorts of factors, macro factors, industry factors, our own execution. But we do expect it to be considerable, and our expectations for the future are absolutely unchanged versus what they were before.

Got you. Very clever. Thank you, Ernie.

Thank you.

The next question comes from Michael Montani with Evercore ISI. Please go ahead.

Hey, good evening. Thanks for taking the question. Just wanted to ask if I could, if you could discuss a little bit what you saw in terms of the wholesale market, both with respect to volumes, but then also I think some of the relatively wide spreads between wholesale and retail that you referenced and if there's any update into October on that front?

I can take that one. So I do think we saw an increase in wholesale market depreciation in Q3 relative to what we were seeing in Q2. I do think the way that flows through our business, it certainly impacts wholesale GPU, where the cars that we're buying from customers that are wholesale eligible experience a little bit more depreciation. That tends to push down margins, and then that can also flow through to volume, if we adjust in order to take into consideration updated depreciation expectations. And so I would say that's probably – that's the pattern that we saw in Q3 and then the most important impact that it had on our business. I think if you move to the retail market, I guess I would also say the retail depreciation that we saw in Q3, based on our data, was actually the highest level of retail depreciation that we've seen in the third quarter going back at least through 2018. So it really was a somewhat unusual quarter for retail depreciation, and so that definitely had an impact on our retail GPU. I think we feel particularly pleased with our retail GPU in light of the fact that we saw many year highs in retail depreciation in the quarter. And so that was just to add a little bit more color. Another dynamic that we saw in Q3 was elevated retail depreciation in addition to elevated wholesale depreciation.

That's great. And just to follow-up on some of the commentary about focusing on enhancing levels of profitability, how – I guess how urgent or how much opportunity do you see potentially, Mark, kind of over the next two years to potentially start eating into the debt stack to avoid some of the higher interest payments a couple years out?

So I think our most important priority, and this has been the way we've thought about our most important priorities for many years running now, is improving the operations of the business. And obviously we've made tremendous gains there. As we pointed out, the last couple of quarters generated over $300 million of adjusted EBITDA, of which on the order of 100 was one-time items. But that still leads to a very significant generation of adjusted EBITDA over just the last couple quarters, with we believe, opportunities for further improvement in unit economics as we continue on with our Step 2 initiatives. And so I think our focus is going to be on continuing to improve the fundamental strength of the business, continuing to make progress on unit economics in Step 2, and then when it's time to return to Step 3, taking advantage of our significant excess capacity to efficiently grow, while also significantly levering overhead expenses. And we believe that combination leads to growth, being profitable growth that I think we're very excited about.

Thank you.

Our next question comes from Daniel Imbro with Stephens. Please go ahead.

Hey, good evening everybody. Maybe I wanted to add two follow-ups on inventory. Maybe first just on inventory efficiencies. So retail GPU, I guess partially is benefiting from just carrying fewer cars, having less depreciation. Do you think Carvana can keep this near 66 days of inventory as we return to growth? Like, have we learned to be more efficient with days on hand or will we need to build that inventory back before we can return to that unit growth?

So I would offer a couple of perspectives on that. One, I think during some very significant growth years prior to the pandemic, we operated in a low to mid-60s average day sales. In some quarters, it even ticked down to the high 50s. So I think we feel very comfortable operating in this range of average day sales, and indeed have done so for many years while executing very high growth rates. So I think that would be the most important point on that particular question.

Got it, okay. And then maybe a follow-up on inventory. As we think about off-lease availability, Ernie, you just talked about returning supply. But if the average lease is 36 months, we're about to come up on a period where fewer units were sold and fewer leases were generated. And so if we don't get back to pre-pandemic off-lease for a couple of years, I guess, what other sources can you lean into to support that growth?

Sure. So yeah, first of all, I think you're right. So just to set the table, I think lease rates were relatively consistent through at least the beginning of 2021. And then I think as you headed through 2021, lease rates dropped to about half what they were prior as a percentage of new car sales by the end of 2021. And so as we head into 2024, the number of cars coming off-lease will decrease. That is correct. We view that first and foremost as a positive, and the reason is because those off-lease cars have not been making it to us. So first order, the effective way to think about our access to off-lease cars over the last year and a half has been that it has been roughly zero. And so we have not benefited from the existence of those cars coming off-lease. And I think as we head to a comp period, as we head into 2021, another thing that happens kind of in another dimension is we start to comp over periods where car prices were higher. So the expected value of those cars were higher, and so the expected residual value of those cars were higher. And therefore, they are kind of less valuable cars relative to expectation that are showing up on dealers' lots, and therefore most likely, less likely to be held by dealers and so they will probably start to flow downstream to us. So, I think we view that as a definite positive, because it gives us access to additional inventory. We also think that that same dynamic has probably been something of a competitive headwind for us over time, because there's been a single channel of cars that was just materially better priced than all other channels of cars that were in a competitive market, which is either buying at auction or buying from customers, which is where we've been sourcing cars. And so we think with that kind of competitive headwind being pulled away from us, that's also positive. So, we look forward to that. I think the point that lease rates going down may change some of the dynamics across the sum of the industry. If we move away from just kind of Carvana specific impacts or independent dealer-specific impacts to the entire industry, I think that point is correct, that fewer cars coming back will change things. But what you have to keep in mind is, those cars were still sold in 2021. Just a larger proportion of them were sold with a loan or with cash. And so what ends up happening then is instead of having a scheduled return of that car to the used car market in three years, you instead have kind of a customer making a decision about when do they want to get their next car, and that will of course have a distribution over time. But in general, customers still desire to swap between cars. And so we do ultimately expect those cars to come back into the market at the industry level as well, maybe just with a little wider distribution of time. So just to summarize that again, I think these points about the kind of impacts that may occur over the next couple years to the sum of the used retail market, because of reductions in new car sales or reductions in lease volumes, we believe those are correct for the sum of the market. We believe that they are actually benefits to us and to other independent dealers, who have effectively been locked out of access to those cars over an extended period of time. So we look forward to that and we view it as the alleviation of a competitive headwind.

Understood. I appreciate the color.

Thank you.

This concludes the question-and-answer session. I would like to turn the conference back over to Ernie Garcia for any closing remarks.

Great. Well, thanks everyone for joining the call. To the Carvana team, thank you. This is another incredible quarter. I said it in four or five different forums already, but I want to keep repeating it. I think it was incredibly hard for anybody to see us making the kind of progress that we've made over the last 12 months from the perspective of 12 months ago. And that's because we've stayed focused, it’s because we've kept our heads down, it's because of the incredible work that you've done, and we couldn't be more grateful. Let's make sure we keep doing it. Let's make sure that we learn the lessons that we learned over the last 24 months and just become a better version of ourselves into perpetuity. Thank you. We'll talk to you guys all again next quarter.

The conference is now concluded. Thank you for attending today's presentation. You may all now disconnect.