Carvana Co. (CVNA) Earnings Call Transcript & Summary

Really happy to have with us today Ernie Garcia, Founder and CEO of Carvana, one of the fastest-growing companies over the past 5 years since its IPO. I'm sure Ernie is going to talk through a lot of the dynamics that the company is seeing right now, including just what's been a really difficult backdrop in the used car arena. Having said that, the company has announced a lot of plans to aggressively cut costs and reach free cash flow positivity in 2023. So I'm sure that will come up. I know Ernie has got some prepared comments, and then we are going to do a fireside chat up here. I do need to tell you, there's a complete list of research disclosures and potential conflicts of interest at williamblair.com.

Good job.

Thank you.

All right. So I'll just go through this presentation pretty quick. The goal is to bring people up to speed to where we are and what our plans are. Safe harbor, all the things that you're used to seeing from companies that we are and are not allowed to say. So as is always the case, we've got one of those. I think to try to set this up, I think what are all the questions that we faced in order to get from where we were when we started to where we want to be, which is a massive player in an enormous industry. And I think the questions that we faced was: One, do people you want to buy cars online at all? Two, is it going to require such a discount that you can't do it economically? Three, can we monetize the remainder of the transaction in a way that drives good unit economics? Four, could we execute and build in a way that made kind of the fundamental trade inside of Carvana, which is effectively a variable cost for fixed cost trade, actually work and be economic? And then five, can we drive down variable costs per unit? I think as we've kind of marched through our -- wow. Was that you? Good job. I'm feeling my talent in moving it forward. So I think as we've marched to our company life, I think when we went public, we probably had answered basically 1 and 2, but I think we still face questions on 3 and 4. We -- in 2016, just right before we went public, we sold 18,000 cars. We had about $1,000 GPU. So we really had -- kind of the questions that we answered to then were mostly around 3 and 4. And I think the question that we face today, given the macroeconomic environment and the capital markets environment, I think, is really question number five. So I think that's kind of the last important question that we've got to answer, and then I think all the questions get a lot more fun from there, which is just kind of how big can we be and what exactly can we build? We've had basically 3 financial objectives since we IPO-ed in 2017. The first was grow units. The second was grow GPU, and then the third was demonstrate operating leverage. There's a number of reasons why we prioritize it that way. One is the size of the opportunity. Two is just kind of like what would be most cash consumptive and what was the easiest gains to get, which we thought GPU was easier than SG&A. And then I think the third was basically of those questions that we faced, while they're all hard, and it's important that we answer all of them, we did think that the fifth question was the easiest to believe in early on. I think in this environment, that's a very important question for us to answer. But I think if we go back to the beginning, there were the fewest questions around, will you be able to operate at costs that are similar to or lower than brick-and-mortar retailers? And so that wasn't our primary objective. In this environment, we're elevating the priority -- the previous priority, number three. And so we're going to focus a lot on SG&A per unit. And then secondarily, we're going to focus a lot on just generating free cash flow. I think the most important part of generating free cash flow undoubtedly today is driving SG&A down per unit. We've operated in the past at SG&A levels that are much lower than we operated to in Q1 for a number of reasons that I'm sure we'll hit. But definitely, that's kind of the biggest delta versus -- SG&A per unit is the biggest delta versus what is, a, proven; and b, possible. And so that's got -- we got a lot of focus there. And then to get to free cash flow, interest expense is kind of what it is. We've got the capital structure that we've got. But then I think the next one is just being judicious on CapEx as we head through this period to make sure that we're preserving cash and being smart about where we spend money in a more uncertain capital market environment. Just at this stage for kind of where we are from a liquidity perspective, we -- post ADESA acquisition as of March 31, we had just under $3 billion of cash. We also had around $1.9 billion of additional assets that are not pledged, and therefore, can't be immediately turned into liquidity. But our high-quality assets that in the past we've been able to finance assets like that as well, it's -- that brings us to total potential liquidity resource of about $4.9 billion, again, of which $3 billion is cash. And then as we think about navigating through this environment and putting ourselves in a spot where we don't want to be beholden to capital markets, the way that we can kind of fully free ourselves of capital markets is to generate free cash flow. And so we try to provide this definition below that uses a lot of our language to just kind of create the simplest equation we can so we can separate concepts out and get to what is free cash flow for us. So it's basically retail units, GPU is gross profit per unit, minus cash SG&A per unit, I'm using cash SG&A to just kind of stay it more concisely, ex D&A and ex SBC. And so then your GPU minus your cash SG&A is basically your kind of positive contribution cash per unit. And then you've got to subtract out CapEx and interest. We don't include working capital in here because working capital has historically not really been a meaningful source or use of capital as it's financeable. And in this environment, we're carrying a large inventory versus what is necessary given that we saw fewer sales than expected in Q1. So we don't expect that to be a major driver, and it just simplifies the equation to have fewer concepts. This is a very simple table. But basically, what it does is in the footnote at the bottom, we outlined that we have about $600 million per year of interest expense, and then we have about $100 million of priority CapEx that we intend to invest. We have another $100 million of growth CapEx that we may choose to invest in this environment. And then after the ADESA acquisition, we have on the order of $100 million of positive EBITDA there. And so when you put all that together, that gets you to around $600 million of kind of -- beyond EBITDA cash that you need to cover to get to free cash flow positive. And so then what this table basically just does is it shows across the top, units, on the left side, GPU. And then it fills in the table with what would the cash SG&A per unit need to be in order for us to get to no longer consuming cash. And then the kind of blue area that's filled in at the bottom and to the right is basically areas where we've executed in the past. So in the past, we've had -- our highest GPU quarter was $5,000. Our lowest cash SG&A quarter was right around $3,750. And so you can kind of see what we've done in the past and what would be necessary to generate positive cash flow, obviously, making meaningful strides toward that number where it would dramatically reduce our burn. But those are kind of the numbers that we would need to hit to no longer consume cash inside the business. I think this chart is somewhat useful because it helps to explain where we are and also why we're so focused on SG&A per unit. So what this blue chart shows is basically SG&A per unit for every quarter going back to Q1 of 2018. And there's a couple of things that are notable in here. First, I'll just kind of like point out that we've had cash SG&A in the $3,750 to $4,000 range many times in our past. And during that period, on average, we are growing at approximately 100%. So we were supporting kind of that SG&A per unit while also making the investments necessary to grow very quickly. So we certainly think the business has proven the ability to get down to those levels in a pretty straightforward way. Then I think if you look at the shape, you kind of -- as you follow all the years, you see these different spikes in Q4 in general across all these different years. Generally, what we're doing is the biggest, most discontinuous growth that we face every year happens at the end of February when customers get their tax returns. And that growth happens fast enough that it's very hard to just kind of build capacity for instantaneously. So we're generally investing for 6, 12 months ahead of time to prepare for that point in time. And then the way that plays out in our SG&A per unit is you can kind of see that it tends to peak in Q4, and then you see the volume show up in late Q1. And then you tend to have like your best quarter in Q2. And then you start to make investments again, and you see that over and over again throughout our history. I think this year, you can see that in Q4 and in Q1, we've really shot up. In Q4, we started to have some affordability issues that popped up. Car prices appreciated again. Interest rates had just started to move up, and we had Omicron that also hit us. And so we had less volume than was anticipated. And then in Q1 is where things were really different than expected. I think the affordability issues really showed up, especially for kind of lower socioeconomic portions of the economy. And that is generally the customer base that is most kind of related to tax season sales. They tend to be most reactive to getting money in their accounts. And so we saw significantly less volume than we anticipated in Q1. And as a result, we had built a machine for many more sales, and our SG&A per unit was significantly higher. So our primary goal is, first step, just get down to where SG&A was per unit in the past. That alone would be an enormous step that would dramatically change our cash consumption, put us in a spot where we have a pretty long runway. And then we also obviously have gains that we can make from there given that we've had these SG&As in the past when we were really focused on growth. So we've set a near-term target of $4,000 of cash SG&A for Q4. We're calling that a stretch goal. I think -- we've obviously run a lot different scenarios to try to arrive at that goal. I don't think it's going to be easy to achieve. It's also certainly not impossible to achieve, but it is reliant on execution. And execution is always a bit uncertain in the precise timing that it will show up with. I think given that we've achieved these numbers before, the real question we face here is how quickly we want to try to get back to those numbers. Moving very, very quickly is better from a cash consumption perspective, but it generates risk that you will derail operations in some of your locations. Many of our different business functions operate in several locations. We have over 100 market ops locations. We have hundreds of logistics legs. And so if you try to instantly just kind of drop back down to where you were, you take some risks that you'll derail some of those locations, and therefore, see volume get impacted, which is not helpful in trying to drive SG&A per unit down. So we're trying to manage those different offsetting risks, and we think Q4 is a good time to shoot for that target. We then expect in 2023 to continue to make gains in GPU and to get SG&A to a level where we're EBITDA positive for full year 2023. This outlines our midterm goal of $3,000 and gives you kind of an area where we would expect each of the underlying components to be in order to achieve those goals. I think if you look through those, most of these goals can be achieved with internal benchmarking that we'll walk you through in a couple of slides, just by looking at different locations where we operate more effectively than our average location across all of our different operational functions. Advertising, for example, and here, we have about $800. We've achieved numbers as low as kind of in the low 500s in our oldest cohort of Atlanta in the past. So while we think there's a lot of work to get to $3,000, it's a number that we've never achieved in the past. We think the pathway there does not require that we invent new process or technologies. It's more just doing things we've already done in the past, spending a little bit less in anticipation of growth and then executing a little bit more consistently across different locations. And we think that, that will get us to this goal. Being at this goal would equate to $2,750 of cash SG&A and puts you in a spot where you're generating pretty significant cash at any reasonable level of volume and GPU. There's 4 drivers, I just kind of walked through some of those, that will get us to this -- kind of this lower $3,000 level. The first and most important by a long way is just balancing expenses with sales. In Q1, our SG&A was around $6,700 per sale. We've been around $4,000 in the past. And so that is, by far and away, the single biggest step is just to get back to where we were, which, again, we think is straightforward in concept. And we think what stands between us and achieving that is execution. So we have to stay focused there. I would -- actually, number 2 would be internal benchmarking. I'll come back to scale in a second. We'll provide some data in a moment there. But I think we do have meaningfully different cost performance across different locations based on the effectiveness with which they're run and their maturity. And one of the major areas that we're focusing across the business right now is trying to drive our 50th percentile up to our top quartile, so basically up to the 25th percentile. And we think there's pretty significant gains that can be made there. Scale is obviously helpful in our business. And then the hardest thing to do but the thing that there's also room for improvement in is just in our processes and in technology, just getting more of our processes more efficient. And I would say that, that is more getting from $3,000 to our long-term goal, which is 6% to 8% of revenue in SG&A is more reliant on kind of the fourth bullet point there. Perfect. So this just provides some -- there are obviously many ways to benchmark internally, but I think these are some of the areas that are useful. If you just look at some of these areas, in comp and benefits alone, this would be enough to do $200 to $300 better than we've ever done historically. Obviously, we've done much better historically in some of our older cohorts in marketing. There's a number of different metrics in here that also point to many logistics locations operating at much lower cost than other logistics locations. So this is just trying to give a sense of what the variability is across different operating locations for cost efficiency in all of our different functions. And as I said, this is a key focal point of the company now to try to manage kind of the mean up to the 25th percentile. Next several sides are focusing on the objectives that we've had for ourselves in the past. And I think the goal of them is to -- we're in a moment where I think, appropriately so, everyone's focused on where capital markets are and on costs and on cash and making sure that kind of everyone gets to the other side and can be excited about the future when we get there. But I do think it can be easy in those moments to forget how exciting the future can be. And so I think one of the goals of these next couple of slides is to talk about the size of the prize on the other side. And then the other is to point out that our team has been pretty effective when we set our mind to something at moving those numbers that we've aimed at. So this is objective number one, which is grow retail units and sales. As I said, we had 18,000 sales in 2016. We went public in Q1 of 2017. We sold 44,000 cars then. We continually grew it to where last year, we were at 425,000 cars. We've rapidly grown market share inside of our markets. If you look at that last data point, our oldest cohort, which was 2013, that's a 3.5% market share. If you extrapolate 3.5% nationwide across 40 million transactions, that would get you to $1.4 million in sales per year. Even that last data point grew at about 50%. Last year, it had a 2-year CAGR in the 20 -- mid-20 range. Probably the kind of -- the 50% is a bit high because we are alleviating some constraints. Probably the mid-20s is a bit low over the last couple of years because we were much more constrained at the end of 2021 than we were at the end of 2019, but those points are still clearly growing. So I think there's a clear path to very meaningful volumes in the future. That would be about 3.5x the volume that we achieved at the company level last year just by getting to that market penetration across the country. If we go to number two, this shows GPU. I think we've, again, appropriately had a lot of conversations recently about, "Hey, like where is GPU and what were some of the tailwinds that you might have had last year?" And Q1 was tough for a number of reasons that I -- maybe we'll spend a little time on in Q&A. So how far back is this all going to go? And I think that those are, again, reasonable questions, and we're happy to answer all of those. But I also think that it's useful to look at this over a longer period of time where we've added about $400 continuously year after year after year after year. And that's been by getting better across all the different GPU line items. And we do think that the underlying business has additional gains from here moving forward. Again, when we went public in Q1 of 2017, basically, the only question we faced then -- or the primary question we faced from everyone was just can you get GPU up to a level where you can make this work. And we're coming off a quarter with $1,169 GPU. We set a midterm target of $3,000. We ended up hitting that target approximately 2 years later. In November of 2018, we had an Analyst Day. We set long-term financial goals for gross profit of 15% to 19%. And in the third quarter of 2020, we had our first quarter where we were at 15%. For full year '21, we had our first year where we are at 15%. So we've set these targets in the past, and we've been able to march to them. And we think that's because we've built very careful bottom-up models. We've generated paths toward those models, and we've executed against those paths. And that's what's enabled us to grow units and GPU. If we head to the next one. I don't know if there's a lag or if you guys -- perfect. Thank you. Yes. This basically shows EBITDA margin. So the majority of the gains here -- almost all of the gains here have been scale from units and then improvements in GPU. And then we've basically been kind of -- on the SG&A side, we've been fighting back the expenses associated with growth, and we've been fighting back the expense associated with building a very large business of buying cars from customers over the last several years. But if you look at that, over the last 4 years, we've also averaged about 400 basis points of improvement in EBITDA margin over that period. And so this gives you a sense of what kind of the bottom line impact is of the targets we've set for ourselves in the past and that we've continually marched toward. A data point that's worth noting because I think it provides a pathway forward and also is useful when thinking about our SG&A objectives is that our 2 oldest cohorts had EBITDA margin that was over 4% last year. So at the company level, we were approximately 0%. Our 2 oldest cohorts were about 4% better. GPU is generally very consistent across cohorts because the entire kind of country shares a pool of cars. And generally speaking, there's not materially different propensity to finance or to get a warranty across geographies. And so the vast majority of that difference in EBITDA is differences in underlying SG&A per unit. And then I think going forward, just to kind of put like -- what are the things that are happening? I think undoubtedly, there's questions about what's going to happen macroeconomically, what's going to happen with inflation, what's going to happen with vehicle production, what's going to happen with interest rates. There's all of these questions and what are the impacts to our GPU and our volumes and our expenses, and I think that we'll be doing our best to keep an eye on that and manage to whatever comes. And we'll try to do it in as risk-averse a way as we can as we head through this period. I think a really big question is, what's our operational effectiveness going to be? We definitely are selling fewer cars than we believe that we could today if we converted our demand in a more effective way, given that we have many areas in our logistics network that are kind of constrained, and therefore, leading to longer delivery times and making it harder for us to display our entire inventory to customers across the country. So we're working on that. And I think the more gains we can make there, the more we can unlock volume that is kind of already present in terms of demand but is not being converted to sales. And then I think we also have to navigate this transition in focus. Any time that a company changes focus, you have to make sure that you build processes to enable that. You have to change what your focal points are. You have to build your plans to kind of step along that path and achieve your goals. You have to make sure that you communicate internally in ways that are effective, that keep everyone excited and moving forward. And then you have to do all that without derailing anything. And I think that those are kind of the 3 key pillars of what we're going to be working on over the next however long as we're marching toward positive free cash. And then I've got a handy-dandy Q&A slide, for which I will sit.

That's just code for me asking you questions.

Yes. Please lay it out to me.

Okay. Look, I mean, Ernie, I think I get asked tons of questions all the time about Carvana. But I think you could boil it all down to mainly, are you hitting an air pocket? Because the industry is down 15% to 20%. So even as you gain share, naturally, growth comes down. Or are you facing an existential crisis where your competitive advantage was somehow exaggerated during the pandemic, and there's a re-rating of your ultimate growth characteristics and your competitive profile?

So I sure hope it's the first one.

It may take 7 minutes.

I definitely think it's the first one. And I think that there's -- I'll tell you what evidence we look at to convince ourselves that it's the first one. And I think that one piece of evidence that -- there's probably 2 key pieces. One is I would go and look at our cohort market share curves that I think have grown pretty consistently across many environments and I think offer a very clear pathway to at least 1.4 million sales with significant growth still happening in our oldest cohort. And I think that unless you think that something has materially changed post-pandemic that will be different from pre-pandemic in persistent way, I think that, that's like a pretty compelling kind of pathway to material growth. I also think that there is this question of like, which are these companies that benefited in disproportionate ways from the pandemic and from changes in behavior that resulted from the pandemic? And I think it's reasonable to ask, are we one of those companies? And I think probably from a demand perspective, we are. I think the actual sales that are realized by Carvana, though, are not just a function of what's the demand. It's a function of what's the demand and also what's our ability to supply that demand. And I think that managing through the pandemic with an operationally intensive business was very difficult. And so the actual sales that we fulfilled, if you kind of like looked at our growth rate trajectory from 2013 when we launched through 2019 pre-pandemic, you would see kind of a growth rate that was slowing as we got larger and larger in a pretty nice, consistent way. And I think if you took that curve and then you estimated at that point in time, what's it going to be in 2020 and what's it going to be in 2021 and 2022, you can go do this exercise and you may arrive at a different place. But where I arrive is you would have actually expected higher growth in 2021 and 2022 than we observed. And I think that's because the operational complexities of managing through the pandemic overwhelmed the demand increases that we saw in the pandemic. And so I don't think that -- because of that evidence, I don't think that the argument that we were kind of a pandemic beneficiary in any disproportionate way is super compelling. And then I do think that we're in a unique environment where we are transitioning, where I think kind of the lower tiers in the socioeconomic spectrum had been really impacted by affordability. There's less cars being manufactured by the OEMs, and that's leading to car price appreciation and more sales happening at franchise dealers. We've seen interest rates go up. And as a result, we've seen a lot of kind of pullback in car purchasing in kind of the lower socioeconomic tiers. And I think we're trying to manage through incorrectly assessing what demand would be in Q1 by not foreseeing all those forces and then just kind of managing to whatever the new normal is as that starts to unwind at some period of time. And I think that, that's really the set of problems that we're facing today.

So the other question, because we have 4 minutes, has to be about ADESA. So the reaction when you first announced it was positive and then kind of turned negative as it became more apparent what the interest carry would be on that. So can you talk about ADESA and why you feel strongly that you should have went forward with the acquisition, and you obviously completed it, and talk through the willingness to carry that extra interest burden as you're striving to get to this cash flow positivity?

Sure. So I think there's 2 ways to think about ADESA. And I think -- I am not objective, but I think both are compelling. But I think maybe one will be more compelling to the average investor today. And I think that the one that's less compelling is maybe the way that most people are thinking and talking about it. So the one that is maybe less compelling today is it lays the foundation for a tremendous amount of growth in the future to make it easier for us to have the infrastructure necessary to go unlock 1.4 million sales per year and beyond as we try to build the kind of infrastructure and foundations that are necessary to build a very large business and go grab the price. And I think that, that -- it will aid us in both reconditioning and logistics in ways that are, we believe, highly differentiated and very difficult to replicate otherwise. And I think that, that's kind of the key story that has been focused on. I think even when you look at it from a pure cash flow perspective, though, ADESA was a $2.2 billion acquisition. We financed that with high-yield debt that cost us just over 10%. So that's around $220 million per year of interest expense. You've got a business that has historically had EBITDA in the several hundred million range and more recently has kind of been tracking at approximately $100 million. So even if we use kind of the recent levels and -- to offset from a cash perspective, you kind of have around $120 million additional that you need to offset. And then there's a bunch of cost benefits that we can get from integrating ADESA in ways that would have been very difficult to get if we didn't have ADESA. One of those ways is taking many more of our wholesale cars that we buy and today transport all around the country to get to the nearest inspection center from wherever we bought them and just delivering them directly to ADESA locations. That's actually pretty meaningful volume. We sold 50,000 wholesale cars last quarter, and many of those cars were kind of shipped very large distances that they won't need to be shipped in the future as we go open up those ADESA locations. And then there's also, on the retail side, a lot of logistics savings. We put out a number that says, if you look at sales that occurred within 200 miles of the customer, where the car was within 200 miles of where the customer was, we spent about $750 less in total than our average sale today. So having these ADESA locations that are much closer to customers, I think we have something like 58% of the population is within 50 miles of one of these locations, whereas I think we were in the high teens or mid-teens for our footprint prior to ADESA, that's a pretty meaningful gain as well. So you kind of start with $220 million of interest expense. You subtract out around $100 million of EBITDA, and then you've got around $120 million left over per year that kind of can be offset by these savings in wholesale and retail, which we think will not necessarily take us that much time to get to a place where that equation is positive.

One last question I want to sneak in, which is about the asset-backed market. So obviously, a big shutdown in that in '08 and '09. Can you talk about how exposed you are to the ABS market?

Sure. So I think the ABS market is generally the most efficient way to monetize receivables. It also is a more volatile way, with volatility being defined as availability and like proceeds that you're likely to get at any point in time. Now that volatility only shows up in relatively short windows of time. In 2008 and '09, prime securitization market shut down for a very short period of time. Subprime shut down for more like a year. Much of that market sort of shifted off-line. At the time, I was Treasurer of DriveTime, which is a subprime -- a retail and finance company in automotive. And we did many private securitizations. It's kind of how that transitioned. I think today, we're a much more notable company than we were as DriveTime in the past. And so my hope would be if we found ourselves in the spot where things shut down, we'd be in a better position than DriveTime was at that time, and we figured it out. But then we've also built this relationship with Ally, which has been really beneficial. They've been an incredible partner for us in 2020 when we had a very short window of time where securitization markets were extremely choppy. We sold all of our loans to Ally across the credit spectrum for a period of time. We have arrangements with them where we have commitments, where we were able to sell a lot of loans to them as well. And I think that we've tried to build that dual channel structure to kind of minimize our exposure or to reduce at least our exposure to the ABS markets because we've kind of always known that ABS markets are great 95% of the time. But you have to be thoughtful about what the other 5% looks like. And I think in an environment like we're in today, where interest rates are moving up very quickly in highly liquid markets like the ABS market, but they're moving up more slowly in like bank markets where they are more tied to Fed funds rates or to just how quickly they want to pass along rate increase to customers, given that they're funding with deposits, I think it's helpful to have a partner like Ally, who is not a direct kind of liquid market participant, and therefore, has more of the banking considerations. So I think that gives us a lot of options as we head through this period.

Great. So we'll see everyone on the breakout.

Thanks.