Carvana Co (CVNA) 2020 Q3 法說會逐字稿

Good afternoon, and welcome to the Carvana Third Quarter 2020 Earnings Call. (Operator Instructions) Please note, this event is being recorded. I would now like to turn the conference over to Mike Levin, Vice President, Investor Relations. Please go ahead.

Thank you, Hallie. Good afternoon, ladies and gentlemen, and thank you for joining us on Carvana's Third Quarter 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 on the IR website. Joining me on this 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. Unless otherwise noted on today's call, all comparisons are on a year-over-year basis. Our commentary today will include non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our shareholder letter issued today, copy of which can be found on our Investor Relations website. And now with that said, I'd like to turn the call over to Ernie Garcia. Ernie?

Thank you, Mike, and thanks everyone for joining the call. The third quarter was an exceptional quarter for us financially and operationally. Let's start with the financial highlights. The biggest headline is that it was our first EBITDA positive quarter as a company. We also crossed $4,000 GPU for the first time. These are incredible milestones and both carry significant meaning and implications for our long-term financial performance and our strategic flexibility.

The numbers themselves are exciting. We think they are even more remarkable when putting them in context. In the last 4 years, we've improved gross profit margin by almost 12%. SG&A as a percent of revenue by over 12% and EBITDA margin by about 25%. We've made all of that progress while simultaneously making the investments necessary to grow the business more than 10x. In addition, the third quarter also saw incredible operational achievements. The first and most notable of these was that we bought more cars from our customers than we sold them for the first time in our history.

This is an amazing accomplishment that was only made possible at the quality of experiences we deliver to our customers, the quality brand we've created, the infrastructure we are building and the Herculean effort put forth by our team. In the third quarter, we bought almost twice as many cars from our customers as we bought at our previous peak in Q1 of this year and over 3x as many cars as we bought from our customers in the second quarter. That kind of sequential growth is pretty unbelievable, but we think it's even more impressive at our current scale.

In order to achieve it, our operations team had to handle a 1 quarter sequential increase in cars bought from customers of over 50,000 units. In addition, the ops team made quick progress toward alleviating our inventory constraints. During the quarter, production caught up for retail sales then grew further, enabling us to substantially increase our inventory by the end of the quarter. While this is impressive progress, we still ended the quarter with just half the inventory that was immediately available for our customers as we had prior to the pandemic.

At our current demand levels, we prefer to be much higher than we were even then, and therefore, still have a lot of work to do in this area. The team is doing an unbelievable job, and we expect to continue to make rapid progress over the next several months to position us very well for another big growth year in 2021.

Now I'd like to take a step back. While the results of the third quarter are exciting and significant, it's important to maintain our perspective and to focus on what is underneath the numbers. Our progress over the last 7.5 years from a fledgling start-up in a single market that delivered 1 car to 1 of our customers once every several days to becoming the fastest-growing automotive retailer in the U.S. with $4,000 gross profit per unit and breakeven EBITDA has come from 4 powerful and persistent forces: one, our clear mission, vision and values. We know where we are going and we know how we're going to get there; two, our unwavering focus on our customers. We know that what really matters is making our customers' lives a little better. We know their expectations are always rising, and we are always working to keep us and to surpass those expectations; three, our long-term focus. We aren't afraid to lay the foundations today that are necessary to build big things tomorrow, even when we don't see immediate results; and four, the strength of our team. Companies are collections of people. Our people choose to care. They believe in what we are doing and find meaning in it. They dream big, they aren't satisfied with what they've done, they don't think it's good enough, and they can't wait to make it better. These 4 forces are powerful. They're what really matter. They provide the relentless pressure that drives long-term progress. They are what brought us here.

So what does this mean for our future? Customer preferences have accelerated their shift in our direction. Our financial performance and positions are stronger than ever before. Our team is executing at the highest level in our history. We're delivering the best customer experiences available in buying or selling a car. We have a scalable offering to get better as it gets bigger. We've demonstrated the power of our mission, our customer-centricity, our long-term focus and our incredible team. We couldn't ask to be in a better position. Our job is to keep marching, to never allow ourselves to be satisfied, to overcome the challenges that will continually arise along the way, to always get better.

If we do that, we're going to hit our goals of selling 2 million-plus cars per year and of becoming the largest and most profitable automotive retailer. And we're going to fulfill our mission of changing the way people buy cars. Mark?

Thank you, Ernie, and thank you all for joining us today. Q3 was a record quarter for Carvana. We made significant progress on all key financial metrics and reached several exciting financial milestones. We achieved positive EBITDA for the first time in our company's history, a significant step on our path toward achieving our long-term goals.

Retail units sold in Q3 totaled 64,414, an increase of 39%. Total revenue was $1.5 billion, an increase of 41%.

Retail unit growth accelerated from 25% in Q2, but was nonetheless limited by constraints on production capacity, which impacted the selection of vehicles available on our website throughout the quarter. Total gross profit per unit was $4,056 in Q3, the highest level in company history and an increase of $1,093 year-over-year. Our record GPU this quarter drove our gross margin to 16.9%, near the midpoint of our long-term model. Retail GPU was $1,857 in Q3, an increase of $552. Growth in retail GPU was driven by a significant increase in the share of our vehicles sourced from customers.

Vehicles sourced from customers reached 56% of retail units sold in Q3, exceeding the high end of our long-term target laid out in November 2018. Total vehicles acquired from customers grew by 128% in Q3, leading us to buy 114% as many cars from customers as we sold to customers, up from 69% a year ago. We believe the high-quality experience we provide to customers creates a sustainable sourcing advantage, which we expect to continue to benefit GPU over time.

In Q4, we expect retail GPU to have a seasonal change that more closely resembles Q4 2018 than Q4 2019. Wholesale GPU was $266, an increase of $146. This was driven by record gross profit per wholesale unit sold of $1,113 and record wholesale volumes. Record gross profit for wholesale units sold was primarily driven by strong industry-wide wholesale pricing. In Q4, we expect the wholesale market to transition to a more normalized depreciation environment, accentuating normal seasonal trends in wholesale GPU. At a higher level, we continue to see improvements in our wholesale channel and expect gains over time toward our long-term goals.

Other GPU was $1,934 in Q3, an increase of $395. The gain in Other GPU was primarily due to a $337 increase in finance GPU to $1,415 from $1,078, driven by tightened credit standards and lower benchmark interest rates.

EBITDA margin was positive 1.4%, reflecting a nearly 7% improvement from a year ago. The significant milestone of positive EBITDA margin was driven by our progress in both GPU and SG&A.

SG&A as a percentage of revenue levered by 1.6% in Q3, primarily reflecting benefits of increased volume and cost efficiencies gained during the pandemic, partially offset by renewed investments in growth for 2021. Looking forward, we expect to make seasonal investments in SG&A expenses in Q4 as we have in past years to prepare the business for 2021, leading to a similar seasonal pattern, while also continuing to show leverage versus 2019.

In addition to strong GPU and SG&A leverage, EBITDA was positively impacted by a $10.8 million increase in the carrying values of finance receivable related assets on our balance sheet. This increase partially offsets the decrease in carrying values of these assets that impacted EBITDA in Q1.

During the quarter, we opened our ninth inspection and reconditioning center, or IRC, near Columbus, Ohio. And following quarter end, we opened our tenth IRC near Orlando, Florida. We also expect to open our 11th IRC by year-end, increasing our total annual production capacity to approximately 600,000 vehicles at full utilization. IRCs play a key role in our model by allowing us to expand our selection, while also speeding delivery times and increasing logistics network efficiency. We remain focused on our plan of building out our network of IRCs and scaling production capacity to meet demand for our offering.

As we continue to focus on building inventory, we have elected not to run our Cyber Monday promotion this year. As a result, we expect retail unit growth to face a headwind around the time of Cyber Monday and a tailwind as we continue building selection on the website.

Following quarter end, we issued $1.1 billion in new senior notes and redeemed $600 million of existing senior notes due 2023. The new notes reduced our annual interest rate by more than 3 percentage points and extended maturities to 2025 and 2028. As part of the refinancing, we incurred a onetime debt extinguishment expense of $33.7 million that we expect to be included in interest expense in Q4. Following the issuance of our senior notes, we had more than $1.8 billion in total liquidity resources on our balance sheet, giving us significant flexibility to drive toward our goal of becoming the largest and most profitable auto retailer.

Thank you for your attention. We'll now take questions.

(Operator Instructions) Our first question today will come from Zack Fadem with Wells Fargo.

And congrats on all the progress. So first question with respect to the inventory constraints and IRC bottlenecks that you've been dealing with. Curious if you're willing to comment on how much your demand outpaced vehicles on the website in Q3? And then could you talk about some of the things you're doing to return the inventory back to pre-pandemic levels? And if you think Q4 is the right time frame for your supply to catch up with that demand.

Sure. I'll try the best I can on that. So I think we wouldn't want to quantify precisely what demand would have been in the third quarter had we had significantly higher inventory levels, but there's no doubt that customers are sensitive to selection. And if you have more cars on the site, they're more likely to find what they're looking for, and therefore, conversion goes up.

And so I think there's little doubt that it impacted things in a pretty material way, most likely. I think our inventory, as we exited the third quarter in terms of cars that were available for a customer to buy that were kind of ready, they could click on and start the purchase process with, that was roughly half what it was pre-pandemic. Pre-pandemic, we had significantly lower sales volumes than we have today. So our ideal inventory would be higher than we were pre-pandemic. And then we believe we have that inventory, our sales would be higher still, and so our ideal inventory would probably be higher again.

So I think right now, the goal is to build inventory as quickly as we can. The team has done an incredible job. We've made a ton of progress in the last 3 months. And I think we've laid the foundation to make even faster progress moving forward. Mark talked in his prepared remarks about everything that we're doing in terms of opening new facilities: we opened our ninth in the quarter; we opened our tenth at quarter end; we plan to open our 11th before year-end.

So that's going to bring us to a total capacity of about 600,000 units. And then we're doing our best to staff up and train and get ready to build out as many units we possibly can to serve the demand that we expect to come in the future. So I think we're really optimistic about where we are. Right now, we're focused on continuing to build out that capacity as quickly we can and to then fill it out with actual production. But I don't know that we want to yet quantify the precise impacts.

That makes sense. And then a quick one for Mark on the GPU line. I'm curious if there's any color that you can provide to bridge the gap from the Q3 GPU to your Q4 expectations. And to what extent you would describe any of the Q3 upside drivers or market forces as transitory across all of the various GPU buckets? And how would you go about quantifying those impacts?

Sure. Yes. So starting with Q3 retail GPU, I mean I think we sort of looked at in a number of different ways, bridging to previous historical period. So I think if you look back in Q3 2019, the biggest driver since then was our progress in buying more cars from customers. I think we saw a significant step-up to our best-ever 56% consumer sourcing rate this quarter, which is a meaningful increase from a year ago. And we also saw strong margins on the cars that we source from customers this quarter.

I think if you look to Q2, there's a couple of other dynamics that enter into the bridge. I think, one, average [age of] sale move down meaningfully on the order of 30 days quarter-over-quarter as we started to resume purchasing and we began selling cars in a much more normalized pace following that peak pandemic period. I think bridging to Q1, if you do that, I think in Q1, we were roughly $1,700 in retail GPU prior to our end of quarter COVID-related balance sheet adjustment.

And so I think you're bridging from Q1 to Q3, sort of a similar story. I think most of the change, primarily driven by more success buying cars from customers. So I think that's sort of -- I think the major drivers as we're looking from Q3 2020 retail GPU back at some previous periods, I think when -- as we look forward to Q4, I think the way we're currently thinking about is we're expecting seasonality that's more similar to Q4 2018. Then Q4 2019, I think -- in Q4 2019, we're still at the sort of early phases of optimizing our buying cars from customers, bidding and pricing. I think that led to some non-normalized dynamics. So when we're thinking about Q4, we're really thinking about seasonality that's more like Q4 2018 than Q4 2019.

When seasonality, you mean as GPU or gross profit as a percentage of the year? I'm just trying to get a sense of what you mean by that seasonal step down?

Sure. I mean, I think we're sort of thinking about it in terms of seasonal changes in retail GPU.

Driven by variation in depreciation rates, et cetera.

Our next question comes from Nick Jones with Citigroup.

The first one, I guess, focusing on 4Q again. How would you kind of rank the magnitude of the supply constraint impact on kind of unit growth versus not running the Cyber Monday deal? Like does kind of finding supply kind of maybe overshadow the headwind from Cyber Monday? And the second question is, I guess, maybe trying to retain the lines on investment in IRCs and opening 2 in 4Q. I mean is that -- is it fair to assume from a profitability standpoint, I guess, kind of back to kind of aggressive investment, maybe not really replicating what we saw in 3Q from here? Any color there would be great.

Sure. So first, with Q4 volume, I think the kind of outlook that we provided in the shareholder letter was that we expect normal seasonal trends across all the major line items, and so that would also relate to retail units. And I think you outlined kind of 2 big differences that will be going on this year versus years prior. One, we don't plan to run the Cyber Monday promotion, as we focus more on building our inventory and preparing for a big first half 2021. And then two, as we continue to build inventory, we do expect that to be a tailwind to conversion. And so kind of that's what sits underneath our expectations for a similar seasonal shape to years past.

And then as it relates to inspection centers, I think we're clearly in investment mode in terms of trying to build out capacity to deliver cars to customers and produce cars that are ready for customers and have been for a while, and that's been part of our long-term plan. Building out inspection centers is the most operationally intensive and kind of longest lead time lever in the business. And so in general, we're always trying to stay out in front of that as best we can. Now from an investment perspective, as you're thinking about cash, we've got a lot of different ways that we can finance that. We've obviously got roughly $1.8 billion of liquidity resource on the balance sheet so we can choose just to financial those with our own cash or we can choose to finance those in the sale-leaseback market if we don't want to use cash for that purpose. So I think we're flexible there.

And then in terms of the way that it impacts earnings, as we open more inspection centers, there can be marginal impacts as that kind of the cost of those centers and the depreciation of the centers flows through fewer units, but that's not a super material driver. So I wouldn't expect that to be a huge part of our story.

Our next question comes from Ron Josey with JMP Securities.

This is Andrew Boone on for Ron. If I think about the flywheel that you guys laid out of kind of great selection, better [conversion], bringing more users to the platform, it seems that adding third-party inventory would accelerate this effect and really improve selection and therefore conversion. We're recently seeing third-party dealer inventory on your site. Can you just talk about Carvana's willingness kind of at a high level, if not this test specifically, to be more of a marketing platform for dealers?

And then additionally, now that you're acquiring more cars from sellers versus buyers, can you talk about your ability to convert those to a trade-in, meaning you get the sale and the car buyer at the same time? As well as any marketing efficiencies you're seeing leading with a kind of why your car first message?

Sure. Thanks for the question. So on the first, I think we're very, very focused on continuing to build out the business to deliver constantly incredible customer experiences at growing scale and to get that flywheel going where we get better as we get bigger. And I think there's lots of forms that takes. In the letter, we discussed kind of growing inventory, giving customers more selection, having inventory in more locations to the inspection centers, shortening delivery times.

So there's a lot of different forms that takes. I think there are also a lot of other theoretic reforms that could take, and there are a lot of other things that we could do to build out additional benefits for our customers. We'll continue to invest in those tests and to kind of look at opportunities to make our offering even better for our customers. But at any point in time, we're running several tests that I would say fit into the bucket that you're describing.

In general, when we're running those tests, we're going to value those tests through the lens of what do they mean for our customer experience, what do they mean from a scalability perspective, what do they mean from an economics perspective. And when we feel like we've run those tests and got them to a place where we feel really good about the outcome through those 3 lenses, then we'll likely pour resources on them and we'll talk about them a lot more, and we'll seek to grow them out in the same way that we've grown buying cars from customers. But I don't think we have any projects that have yet risen to that level of discussion.

Going to your second question, I think there's enormous opportunities there. We're still very, very small compared to the retail market in terms of the number of customers out there buying cars and the fraction of those customers that are buying cars from us. We're also very, very small compared to the market of customers that are disposing of their cars. And so I think those 2 markets are, to some degree, independent and then they're connected when a customer chooses to do 2 transactions simultaneously through a trade-in.

What we are going to seek to do is to make sure that we're giving customers the best offering available on both sides of the transaction, if they want to buy a car or if they want to sell a car. And then as we get those customer interactions, we have an opportunity to explain to them the other things that we do as well. And I think that can happen simultaneously and would look like a trade-in, or I think that it can happen with lag where a seller becomes a buyer 6 months down the road or a buyer becomes a seller 6 months down the road.

And so again, I think the thing that matters most there is we want to focus on our customers, make sure their experiences are great, make sure our offering is in a really good spot. And then just scale both as quickly as we can because we're very, very small compared to the opportunity.

Our next question comes from Adam Jonas with Morgan Stanley.

So Ernie, they say imitation is the ultimate form of flattery. If anyone's listened to these franchise dealer calls or any other auto retail call, it's like you've really woken them up, right? They're always -- they're talking about omnichannel, digital touch list, et cetera, et cetera. Are you seeing any sign at all that some of their admittedly nascent efforts, even if they're nowhere near the full flywheel that you have and may never get there, have the attempts at least impacted your business in any way, including inventory acquisition or pressure at the auction, share, anything?

I think the simple answer to that question so far is no. And I'm not sure that should be that surprising because I do think that we're very small compared to this market. In the quarter, depending on what I'm using from a market perspective, we are probably around 0.6% of the total transactions in the market. And then probably the list of competitors that you're referencing, if we add them all up, might add 5 to -- probably not a lot more than 5% additional market share on top of that.

So I think all of the bigger players that I think are starting to get more focused in this direction, even if you sum them, are still small compared to the market. So I think the overall opportunity is really, really large. And I think when we think about what's most important for us, you have to kind of decide what you focus on, and we would like to think that we're the clear leaders in this space, and we think that the important area for our focus to be on is just on our customers because they're telling us where we need to go.

And so we're continuing to really focus there. And then we think that the dynamics of the market that we've discussed in the past, the fragmentation and the relatively few players that really have the capacity to make large financial investments and building out a true e-commerce platform suggests that that's not necessarily the area where we should be putting most of our focus.

We think that we offer the best customer experience, that we have the most scalable model and that we've got the capacity now through our balance sheet and also through our financial performance to be able to invest in our offering. And so we're focused on us. We're focused on execution. We're focused on our customers. And I don't think we're yet seeing anything that's impacting us from any competitive moves.

Okay. Great. And my follow-up kind of is in the topic of these test buckets that you talked about. I know you guys look at a lot of different things to improve frequency and all sorts of different ways to improve experiences beyond just buying and selling of used cars. There are a lot of car companies with increasingly, these electric car companies, start-ups and even some legacy companies that could get carved out, and may want to engage with the consumer in a B2C way in a far better way than the traditional model, right?

And I'm just wondering if there's an opportunity, and I know it doesn't exist right now, it's not part of the 2021 story, perhaps. But are you even having discussions internally, if not externally, with how the Carvana platform could help a new entrant into the U.S. market engage with the consumer in a digital vastly superior way? Is that an opportunity? We're thinking 5, 10 years out and is something relevant to talk about today?

Yes. Thank you. So what I would say is our focus has been and remains on delivering great experiences to customers that are buying used cars. I think our platform is, you can think of as kind of being the thing that sits between cars and customers. And I think everything that we do, our strategic lens is we're looking at different ways to acquire cars and make sure that we've got inventory available for our customers, and we want to make sure that we have the best and simplest transaction that can connect them -- connect our customers up to those cars.

So I think that's kind of interesting opportunities theoretically in the future. But I also think it's important for any given company to make sure that you stay focused on the things that are most important at any point in time as you continually drive toward those things. And so for us, that thing is making sure that we're delivering great experiences to our customers who want to buy or sell used cars.

Our next question comes from Daniel Powell with Goldman Sachs.

One and a follow-up. Maybe another run at the retail GPU in 4Q. I guess based on the drivers of the outperformance in 3Q being more consumer sourcing, faster sell-through and more reconditioning efficiency, I guess which one of those drivers would you expect to step back enough such that the depreciation schedules and things that you're talking about catching up and for you would actually drive some of that seasonality that you're expecting? Because at first glance, it looks like all 3 of those should still be nice tailwinds for you in 4Q.

So I would just start with normal seasonality. Normally, there is variation in depreciation rates across the year, with the fourth quarter tending to have the fastest depreciation rates. You can think about depreciating -- depreciation happening in time. And kind of depreciation per day tends to be the highest in the fourth quarter. And so if you look back at our historical numbers and kind of massage out cars bought from customers and days to sale, you just kind of see those sorts of trends in there. And so I think that's the major driver that's underneath our expectation, and that's something that we've seen over many years since launching in 2013.

Okay. And then on finance GPU, obviously, really strong performance. Similar type question there. With the interest rate environment that we're in, any reason that the strength that you're seeing on the sale of the receivables there would deteriorate for any meaningful reason? And on a somewhat related note, was there anything in the tighter credit standards that you felt like impacted conversion on the site in the quarter?

Sure. So what I would start with there is, I think we also gave some directional color for our expectations for finance GPU in the shareholder letter and said we expect it to follow normal seasonal patterns. So there is some pattern there which relates to a number of relatively small factors, but the data to create the seasonal pattern. So I won't dive into all those underlying factors right now.

I think in terms of what happened this quarter with finance GPU, I think there were a lot of things that were moving. And so I'll try to give you a quick list, and hopefully, that would be helpful in figuring out what we believe is sustainable for the long run. So first, and I think most importantly, I do believe that we made some fundamental improvements to our credit platform. That includes improvements in credit scoring and credit pricing and credit structuring. And just getting more efficient at all those things.

So I think there was some fundamental improvement that we would certainly expect to persist. One is going through the pandemic, we tightened credit pretty dramatically. We've maintained tighter than normal credit standards and -- that are continuing today. When you tighten credit, all else constant, you generally would expect to be originating more valuable loans. You also would tend to expect lower sales volumes. But so I think as it relates to finance GPU, that would be a tailwind to finance GPU that at some point, we would likely unwind as credit markets normalize and kind of the current environment becomes a little bit clearer. And then you'd expect that to be a tailwind to sales.

I think the market clearly got better in terms of the execution of just credit markets in general. So I think that, that was a tailwind going from Q2 to Q3, but I think it's that would kind of be just moving more to a more normal function market. Interest rates clearly went down. All else, constant interest rates reducing and kind of benchmarks and spreads reducing generally would lead to increases in the value of finance receivables because consumer prices are -- interest rates are generally stickier than the underlying benchmarks. And so I think that, that would be a tailwind. That one can play out over a little bit longer period of time, but I would not expect that to be persistent in the long run.

And then I think that's offset to some degree by the fact that we didn't monetize in the ABS markets this quarter. We monetized the vast majority of those we originate in our platform through our partnership with Ally, which is the great partnership we've discussed in the past that provides with so many benefits, including flexibility and strength and reliability that we saw going through the pandemic. But that channel, generally speaking, is not as economically efficient as securitization.

So I think that was something that went a different direction and kind of worked against us in the quarter. I think in the immediate term, there's a lot of things that potentially could move that make it a little bit hard to forecast, but operating under the assumption that kind of big macro factors remain in a similar spot to where they are now, we would expect to see somewhat normal seasonal trends going from Q3 to Q4. And then taking a longer lens, where we found ourselves from a finance GPU perspective this quarter was basically at the high end of the long-term model that we put out in November of 2018. And I think that we clearly vetted that at the time and thought about it pretty carefully, and we believe that those are kind of long-term sustainable numbers connecting the dots between here and there. I'm sure there's going to be some bumps along the way, but we don't think there's anything that's clearly not sustainable in the long run in finance GPU.

Our next question comes from Rajat Gupta with JPMorgan.

Just a follow-up on the previous question on finance GPU. You mentioned the long-term target had high $1,400 at the high end. But I think when you put out those targets, your benchmark rate assumptions were higher, your spread assumptions were also higher. Both are much smaller and tighter now. So is it fair to assume that if rates stay lower for longer, your upside potential there is meaningfully higher than that $1,400 or not? And I had a follow-up.

First order, I think we haven't seen enough yet to make us think that we should materially change our long-term financial targets. I think that we should think about kind of the value of the loans being driven by the spread and capturable value between the consumer interest rate and the underlying benchmarks. And so as benchmarks move down, the spread between those and kind of the loss adjusted kind of interest stream that we get, it doesn't necessarily move. And so I think that when interest rates are moving, that can have a short-term impact, but it probably shouldn't have a sustained long-term impact. So I think that in a vacuum would not be sufficient to change our long-term model.

Got it. Got it. Okay. Makes sense. And then just on SG&A leverage, a nice progress here in the third quarter. Within the different buckets, are there any areas maybe around advertising or other overhead expenses where you might have tightened spend temporarily that should start to come back in the future? Or should we view it as being a more normalized level in the third quarter? Just curious as to the puts and takes there as we go forward.

Sure. So I think on SG&A, if we're during into the fourth quarter, I think we're sort of expecting normal seasonal patterns in total SG&A as we start to ramp up for 2021 first half. And normally, Q3 and Q4 are investment periods for us, where we start the seasonal investment process to prepare for the next year. And so we'd expect that to be the case this year as well on total SG&A, while also showing leverage over 2019.

So I think that's sort of the leading pattern that I would suggest for Q4. I think in a bigger picture sense, I think we're very excited about the leverage that we're now showing year-over-year now that we've moved a couple of steps away from COVID. We certainly expect to show more leverage looking forward into the future. I think one thing that we're very excited about is the patterns that we're seeing in SG&A in our cohorts, where older markets tend to have significantly lower advertising expense per unit than newer markets, significantly lower last mile delivery and logistics expense per unit than newer markets, and then much higher market shares, which means lower overhead per unit.

And all of that, I think, all those patterns that we're seeing in the cohorts, I think give us a lot of excitement about where we expect SG&A to head over time.

Our next question comes from Colin Sebastian with Baird.

This is Dalton Kern on for Colin. Just another question on the kind of reconditioning capacity looking forward to Q4. I know it sounds like just acquiring inventory and getting it through the pipeline seems like the primary constraint rather than absolute capacity in the network. But can you give a sense for how much, if anything, the Ohio and Orlando IRCs are going to be able to contribute near term? How long does that take to kind of ramp up and bring them online? And when should we expect that to start to help flow through?

And then separately, just on the topic of the cohort profitability and SG&A specifically. A lot of the scale benefits in proximity makes sense in terms of lower logistics expenses. But can you talk about how advertising in your older cohorts is trending relative to your long-term targets and kind of anything you've learned there that you think is applicable to your longer-term strategy in driving advertising leverage?

Sure. So on the new IRCs, what I would say is, we're working hard to make sure that we're laying out that facility capacity as quickly as we can. And then we have kind of our goal is to ramp that up. We're working to make sure that we build the process to enable it to ramp it up as quickly as we can, but it also becomes a choice on how quickly we ramp it up once we have those facilities up and running. Now right now, it's a little harder to forecast than normal because I do think coronavirus has impacted things from a hiring perspective, from a training perspective and in certain cases, from an efficiency perspective. And I think that is more true when caseloads get higher, which we've seen recently. We haven't yet seen that impact the inspection centers today, but there's certainly some risk of that in the future. And so I don't know that we want to be forecasting precisely how quickly we'll be ramping inventory. Right now, our expectation is that it will continue to ramp as it has over the last several months and potentially a little bit faster, but I don't think we want to quantify it more than that. And Mark, can I hand it over to you?

Sure. Yes. And then on advertising expense in older cohorts, so I think we're feeling pretty good about the patterns that we're seeing there. I think we've always seen cohorts that are older and have had more time to accumulate awareness in word of mouth, have meaningfully lower advertising expense per unit or ad ex as a percent of revenue than newer cohorts where we're just sort of getting ramped up in terms of raising awareness of our offering.

I think what that means is our older cohorts are significantly closer to our long-term model than the company average as a whole or certainly newer cohorts that are in that early phase. I think in Q3, our older cohorts are probably on the order of maybe 50 basis points or so outside of our long-term target. And so I think obviously, we feel pretty good about that just given all the opportunities for continued improvement and continued growth that we have. And as I was mentioning earlier, I think we're very encouraged by the trends that we're seeing in cohort unit economics, which give us a lot of conviction about our path to the long-term model and SG&A.

Next question comes from Chris Bottiglieri with Exane Paribas.

Congrats on your early profitability milestones. By my math, I have you, your earliest markets with EBITDA margins of around 7% to 7.5%. If your overhead was at your long-term target, you'd essentially be at the midpoint of your long-term margins. Yet your market share in these early markets, it's still a little bit below your long-term targets. So I guess the question is, with that backdrop, is there anything you've seen today that's made you more constructive or bullish on your long-term targets? Are there's sort of like certain segments or cost structures where you have better visibility into than you might have had 2 years ago?

I think we're obviously extremely happy with where we are. We're extremely happy with how the company has performed. We hit a bunch of milestones this quarter. I think the underlying cohorts look great. We're extremely happy about that. I also think it's early to be focusing too much on kind of raising the long-term model. I think that long-term model we put out there, we put a lot of thought into. We believe we've got a very clear path there. And we think that path is becoming clearer, at least to the outside world, all the time. We feel like it's been clear for a while, but I think it's getting clear to the outside all the time.

So we're going to keep marching toward that. At the company level, we still obviously have a lot of work to do as we continue to [abandon] and lever that SG&A further. But I think things are looking really good. We're very confident. I'm not sure it would be prudent to raise the long-term financial model at this point, though.

Our next question comes from Lee Krowl with B. Riley Securities.

First question, in the shareholder letter, you mentioned consumer preference migration several times. I wondered to get a little bit more detail on that, particularly against the dynamics of kind of new versus used mix as we've seen it shift over the summer, but also just kind of the dynamic of purchasing digitally versus the dealership. What have you guys seen? What seems a little bit faster than expected? And what seems a little bit more durable versus kind of the near-term dynamics?

Sure. So I think it's probably a number of things that have happened. I think one is clearly is a move toward digital in everything, not just in buying a car, but in everything. I think as people choose to stay home more and choose to distance more from others, I think just more things in general are happening online. And as more things happen online, people get more comfortable doing things online. And I think that generates habit formation. I think generally speaking, the biggest inhibitor to change is pre-existing habits. And when you go through something like this where habits are forced to change, people try new things. They like new things and new habits are created.

And so I think our expectation there would be that, that is likely to be a persistent change, and it's an acceleration of trends that were already clearly underway, given the results that we've shown over the last 6.5 years prior to the last year. But I do think that's likely an acceleration. I think another thing that is happening is there seems to be a shift toward personal car ownership. I think it's hard to say exactly how persistent that will be over time, but that seems to be a trend that is real and seems fairly strong at the moment. It seems like the underlying drivers of that have a high likelihood of being able to persist over time.

I think the shift from new to used is most likely transitory. I think that the biggest driver is just the mechanical driver of there being fewer new cars and production being impacted with the manufacturers. And then that's just kind of putting a, just, cap on how many sales you actually can have. So that's probably the biggest driver. The other driver is, generally speaking, in periods of economic distress, there tends to be substitution into less expensive cars. And so that is probably also at play to a lesser degree. I think both of those we should think of as transitory. And those probably have implications for all kinds of different things, including the way that car prices will migrate over time. And I think it's hard to forecast.

But I think those are the major trends. I think the other trend that was present early in the pandemic was basically a lack of other options. There just weren't other places to buy a car. I think we've moved through that. And so I think that was the most transitory of forces, but I think that, that one is behind us at this point.

Got it. And then just a quick follow-up. Obviously, the priority is to scale inventory and broaden the offering. Could you maybe just talk about shifting from your core kind of center inventory out into the value and luxury type SKUs that some of your competitors focus on and perhaps the GPU implications of those?

Sure. At a high level, going back to the spring that we discussed earlier, I think we view our job as building the best transactions just between customers and cars. And then we don't think our offering is specific, in any way, shape or form, to a specific type of car. We think that we could go way up to more expensive cars. And we think we can go way down to less expensive cars. We think customers, when they're looking to buy a car, they want a great selection. They want to find the car they're looking for. They want a fair deal. They want a simple process. They want a process that's fast once they've decided what they're going to do. They want confidence the car is in good shape. They want confidence that if they make a mistake and doesn't fit their life, they can return it.

And those things really have nothing to do with the specifics of a car. So I think we feel like we've got an opportunity to serve a very broad swath of customers looking for all types of cars over time. And I think we're just trying to do the best job we can with all of our other priorities just managing which things we're tackling at what point in time. And we have been steadily broadening our car offering. I think there's a lot of room to go, but especially buying cars from customers has given us access to both cheaper cars and more expensive cars and more diverse cars. So I think that that's -- that is happening today, but there's clearly more room for that.

And then from a GPU perspective, at a very high level, I think the simplest way to think about it is the cars that we're selling don't impact GPU in a super material way. They would flow through retail with a kind of moderate impact that would be related to the kind of expense level of the car. More expensive cars do have moderately higher GPUs and cheaper cars can have moderately lower GPUS. But that's not a huge impact, and it's -- you can kind of think of it as being a large portion of the margins being fixed and then a relatively small portion of the margin being related to the price of the car.

And then from a finance GPU perspective, I think, generally speaking, that one is more proportionate to the price of the car. So I think you can kind of think about that as having the relationships that you'd expect there. Wholesale, I don't think has a very strong relationship to the car nor does the other. So it's not a huge driver of that. And I think inside of the likely bands that we would move at an aggregate company level, I don't know that you should expect that to be a huge driver of GPU migrations in the long run. But I do think as we broaden, that's an opportunity to just broaden our offering and positively impact more customers.

Our next question comes from Brad Erickson with Needham & Company.

Just a couple for me. I guess another one on the oldest 5 cohorts in Q3 you called out in the letter, which I guess were solidly EBITDA positive. Can you give us a sense of growth rates going on in those markets just relative to corporate average? I get that they may be lower just because inherently they're more mature, but just curious if that's wrong, any color you can provide there?

Sure. So I think generally, that frame is correct. I would say the trends we've historically seen is newer markets tend to grow faster. They benefit from pre-existing awareness and a lot of inventory early in their life. And then older markets tend to grow slower. As they get bigger, they're just working off of a bigger base, so they continue to grow quickly. I think right now, we're in an especially unique time. When you're looking at different markets, the different markets are impacted in different ways by the different operational constraints that we've got in the business.

And so I think when we're looking at this internally, what we're always focused on is what's happening to underlying demand. And we can kind of massage out what's happening to underlying demand by looking at what's happening to sales and looking at the drivers of conversion that impacts sales, which include where inventory is, how much inventory is close to different markets, what delivery times are in different markets, what our inventory distribution looks like relative to the distribution that is traditionally sold in different markets, what our credit policies are relative to the distribution of credits in different markets.

There can be a lot of different drivers there that can impact things between those markets, and that's why we generally, outside of annual and aggregating at the cohort level, we haven't provided a ton of detail on individual markets. But when you look at all that together and just say what's happening to demand generally across the country, whether it's looking at newer cohorts and older cohorts or looking at different geos or anything else, we're seeing very consistent trends in demand growth.

Got it. That's helpful. And then just wanted to follow up. I'm still a little confused on kind of what you're saying about the sustainability of the strong pricing you're seeing lately. So on the one hand, like it seems like some of the inventory constraints are going to persist a little bit here. You're not committing the sort of a magnitude of improvement in Q4. On the other hand, you mentioned wholesale channels should sort of display a little more normal back half depreciation characteristics. So can you just tell us like which of those is a more important force for pricing into Q4?

Sure. So let me -- I think let's start with wholesale, I think that's simpler. There was clearly abnormal dynamics in the way that vehicle pricing migrated really over the last probably 6 months, give or take. And I think that, that was -- probably if you had to look at what was the driving force, it would have been stimulus probably provided an extra demand for cars. I think people spending more time at home and less time on services, it gave them more money to spend on cars. So it was probably a positive boon to demand. I think potentially the shift toward personal ownership was helpful there. But then I also think that there was less production capacity with the manufacturers, which meant that there was fewer new cars, and so that was helpful for new car pricing and then that kind of flows over into used cars and kind of down to cheaper used cars as well.

So I think we've clearly seen abnormal dynamics in the whole market that we would not expect to persist in the long run. If you look at our wholesale profit per wholesale unit, we showed record numbers by a long, long way in the third quarter. And so I think what we called out in terms of our directional expectation there heading into the fourth quarter as we would expect it to be normal seasonal patterns, but more pronounced than normal.

And the reason that they would be normal directionally is because normally you do move into a higher depreciation environment going from Q3 to Q4. But this year, we expect that delta in depreciation to be significantly higher. And in certain segments, there was even vehicle appreciation in the third quarter. And so from a relative depreciation perspective, we expect it to be materially higher. And so we think in the long run, the types of property we were able to show for wholesale unit this quarter are likely achievable. That's what we've got at the high end of our long-term model. But we think that between here and there, we're clearly in a step down and it starts to build again as we've been building for the last several years because it just wasn't a normal environment.

When you're looking at retail, I think dynamics are different. In theory, they should be related because as we buy cars from customers, something we've historically said is that the incremental margin that we get on cars bought from customers is similar to the wholesale margin that we get on cars that we buy from customers and sell wholesale. In this quarter, I think that, that relationship wasn't as tight as it's historically been just because different segments of cars were appreciating and depreciating at different rates. And then also because we weren't buying cars in the absolute trough that was in kind of late Q2 and Q3, we just weren't -- or sorry, late Q1, early Q2, we were buying cars during that period, and that would have been the period where you would have bought cars and seen the most appreciation.

And then there's also some dynamics with just the way that we handle pricing. We generally don't write cars up even if there is appreciation, and we have time-based write downs that are a function not only of time-based depreciation, but also how that car is performing relative to other cars on our site. And so the way that we handle our retail pricing provides a lot more stability and kind of less exposure to the way the wholesale market is operating. And so we expect retail to move back, like we said, similar to the seasonality we saw in Q4 2018, but not nearly as dramatically as we would expect wholesale to move back, which we do think was powered by clearly abnormal forces.

Our next question comes from Brian Nagel with Oppenheimer.

So one question I really want to ask, just dive deeper into resisting the whole -- the effort to buy more cars from customers. I mean clearly, a big driver here for a while, particularly in the quarter. But how significant -- how large could that become for Carvana? And as you continue to drive that effort stronger or higher, are there incremental benefits to your company, meaning are you going to be able to buy cars even better over time from customers?

Sure. So I think -- at this point, I think the business of buying cars from customers, that team has done an absolutely unbelievable job building out that customer experience, the technology that supports it and the operational capacity to handle it. I really -- I want to talk again about that stat that I gave my prepared remarks. It's pretty remarkable when you think that from Q2 to Q3, we increased the number of cars bought from customers by 50,000 units in 1 quarter. That's a pretty incredible number.

And I think the team just did an unbelievable job to enable us to do that. So that's really exciting. I think because of all that growth and because they were able to grow on top of the infrastructure that we've been building over the last several years to service the retail side of the business, they were able to catch up to retail really fast. We've only been really investing in this business for a couple of years and it's made a ton of progress. But I do think these businesses are, in theory at least, largely independent. As we discussed earlier, they're connected through trade-in, but the majority of the volume in both cases is independent. And so I think from here, our goal is to grow each business as quickly as we possibly can and effectively have those businesses race each other, so everyone's winning, I think, is effectively what we're trying to do.

And so I think it's -- when we're stating these businesses as a ratio of one other, that's going to be a function of how one business is progressing relative to the other, but we're going to be trying to move them both forward as quickly as we possibly can because we're very small compared to both opportunities, and so we're excited about that.

As it relates to opportunity to get better, there's no doubt. I mean I think across the entire business, there's opportunity to get better everywhere. And I think every group inside the business on the retail side and on the purchasing side, I think has 1 million things they want to get done and looks at what we currently have and is mildly embarrassed by and really excited about what's about to come next.

So I think there's a tremendous amount of opportunity. I think to some degree, there's debatably even more opportunity in buying cars from customers just because that's a younger business. And so we haven't had as much time to invest in that offering as we've had to invest in retail. But I think we're very early in both. And then both businesses also benefit from volume. As we get more volume, we get smarter at bidding and we get smarter at pricing, we get smarter at merchandising. So I think we get better at a lot of different things. So I do think volume will make us better.

Our next question comes from Nat Schindler with Bank of America.

Just a couple ones. I don't mean to belabor the point, and I know you've gone over this. But can you be a little more specific on the GPU guidance? And I know you're saying it's going to be more like 20 -- Q4 2018 than Q4 2019. But I got to be clear, is that gross profit per retail unit inclusive of the other revenue, i.e., the total gross profit per retail unit, which really wasn't all much of a difference in the move sequentially in '18 versus '19? '18, it was down 7% sequentially and '19, it was down 5%. Are you talking just the retail sales -- just the retail part, not including the other funding?

First, so I'll take a crack at that. I think the biggest thing that we expect to see is that wholesale GPU per wholesale unit number to return to more normalized levels, which is a sort of accentuated seasonality relative to what we've seen in previous years because the wholesale market, we do think was in an abnormal depreciation environment through most of Q3, and we expect to see return to sort of normal seasonality. So I think -- with that, I think we expect to see a change in wholesale GPU that's going to be driven by that wholesale gross profit per wholesale unit sold number.

In retail, I think the way we're thinking about it is sort of within the range of normal seasonality and by normal seasonality, what we mean by that is sort of just looking at sequential changes in retail GPU. And we think the normal seasonality is best reflected by 2018 because 2019 had a number of other factors going on with us scaling up, buying cars from customers, for example, that we think led it to not have what sort of a normal seasonality in the retail component of GPU would look like.

And so I think those are some of the trends. I think there's some little seasonal trends that can happen in financial GPU as well, which Ernie pointed to. So I think like at a very high level, we think, generally speaking, we expect to see normal seasonality. We expect to see it sort of extends seasonality or accentuated seasonality in wholesale gross profit per wholesale units sold. And when thinking about what historical seasonality looks like, we think '18 is probably a better benchmark than 2019 on the retail side just because of the other dynamics that were impacting Q4 '19. That's hopefully a helpful summary.

I'm still actually confused. So what I'm confused about is just simple definitions here. When you're talking about retail units sold, you're not talking about total retail gross profit, which is inclusive of other revenue, the financing. You're talking about just the retail sale of the car. And in which case, that fell sequentially in 2018 -- Q4 2018, sequentially, it fell by 22%, while in 2019, it grew. So -- but there isn't much of a seasonal pattern here to look at because the earlier patterns are much -- are all over the place. But -- so -- but are you just saying just the retail side? That number, I guess it was like $1,900. That number goes -- is going to go down similar to what it fell in 4Q '18 from 3Q '18.

Yes. We're talking -- with that piece of the commentary, we're talking about the retail component of GPU, that's correct. And then we gave some separate commentary related to wholesale GPU, which is another separate line item. And then with respect to finance and other GPU, we gave a little bit of commentary as well.

That piece is related to the retail component GPU.

That's correct.

Okay. Got that. Okay. Then just a couple more questions, just very quickly. In the last quarterly note, you gave us a July unit growth to just tell us where trends were for the third quarter. I know you're not giving actual much guidance in -- for the whole quarter, but is there any -- can you give us anything that happened in October, just to give us a baseline? Or I know it's not totally over, but

I think we want to stick with the commentary that we've provided so far. We think that the Q4 sales growth is going to largely be a function of our ability to produce more cars, and we're working very hard at that.

Okay. And then just one final question. This is more of a kind of a long term how margins might flow out question as you do a whole lot more of this buying from consumers. So obviously, you've talked a lot about the GPU value. If you purchase a car from consumers, you get it for a much lower price than from auctions. It's less efficient market, totally makes sense. But how much incremental operating expenses are there? When you purchase a car from auction, it's not a lot of -- it's a lot of cars in a single place. They're often right near your IRCs. So if you -- but if you buy cars from consumers, they're all over the place. You then -- and they also -- it takes people to negotiate in that and talk. Is there a lot of extra operating expense in that relative to buying at the auction?

Sure. So I think the big expenses associated with buying cars from customers is an inventory source is exactly what you laid out. It's basically the labor and delivery/pickup costs of going out to a customer's driveway to pick up the car and then bringing it into our logistics network and then transporting it into one of our IRCs. All of those expenses are actually -- those are reflected in our wholesale gross profit number as well as in our retail gross profit number.

And when we say that buying cars from customers is more profitable than buying cars from auctions, that's -- what we mean by that is it's more profitable net of those expenses of going to get the car from the customer and bring it into the IRCs. So I think that's the most important point is the most significant expenses. There are some other little expenses in call center and the like. But the most significant expenses are in picking up the car from the customer and transporting it to our IRCs, which are all reflected in our GPU numbers and are also reflected when we say that those cars are more profitable than the cars sourced from auction.

And our last question today comes from Alex Potter with Piper Sandler.

I'll just ask one here. Average days to sale, and I guess the inventory dwell time, they've been also really strong lately. And obviously, that was a contributor to your retail GPU. Just wondering if you could comment on how much of that is sticky? I mean part of that could just be that consumers are snapping up cars as soon as they show up on the website; the other part maybe is your own ability to process vehicles more quickly through inventories. I guess if you could just maybe talk through both sides of that equation and let us know which ones are sustainable versus not?

Sure. I think instead of using the language of sticky, I think I would use the language of it's largely a choice. I think our demand is growing. And at any given level of demand paired with any given level of inventory, you're going to have some turn time that will emerge in that. And then as you grow your inventory from there, you're going to increase conversion, but you're unlikely to increase conversion by as much as you grew inventory, so your turn time may get a little bit longer.

And so I think we can use -- in a normalized environment where we have as much inventory as we would like, we can use inventory as a lever that we can think of similar to marketing, to kind of have some incremental expense that flows through retail GPU that helps us to acquire more customers.

And so I think that's a lever that, in general, we have today, we don't have that lever. It's not really an active lever because we're just trying to produce as many cars as we possibly can to catch back up to the demand levels that we're seeing. But moving forward, I think that, that is largely a choice and then at kind of equilibrium in the long run, once you've got a fulsome selection of cars, which takes many more cars than we have today, there's no real need to continue to build out selection of cars when you have, again, fulsome selection that's located well across the country to minimize delivery times.

And so at that point, I think there would be a kind of meaningful and secular decrease in turn times. But between here and there, it's largely a choice and it's one of the many levers that we have to drive customer demand.

This concludes our question-and-answer session. And I would like to turn the call back over to management for any closing remarks.

Thank you, everyone, for joining the call. We really appreciate it. To everyone out there on Team Carvana, I hope you're proud of the achievements that we had this quarter. It's pretty remarkable for us to have our first EBITDA positive quarter as a company, for us to hit $4,000 GPU. But please remember that as incredible as those achievements are, it's all a function of the things that we do every day and have been doing for the last many, many years. And it's all because of how much you choose to care and how you've made this mission, your mission. So thank you so much. I hope you take a moment to enjoy the view and gloat a little bit, but then I hope we get back to it because we still have a lot of work to do. Thank you for all that you do.

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