Carvana Co. (CVNA) Earnings Call Transcript & Summary

Great. Thanks, everybody, for joining us this morning. Hope you're safe and healthy in whatever locations you happen to be in today. I'm Rick Nelson. I cover the car dealers, automotive retailers online-only for Stephens. This is the fireside chat with Carvana. Super excited to have with us Mike Levin, who is VP, Investor Relations. [Operator Instructions] I thought the best way to start this, for people that were not familiar with Carvana, although there's fewer and fewer of those every day that passes, but I ask Mike if he could give an overview of Carvana's business model, the value proposition and how they frame up the growth opportunity in the used car market.

Sure. Thanks, Rick, for having us, and thanks, everybody, for joining in today. So Carvana was basically created to create a better way to buy a car. And essentially, we think what's been at root of the relatively poor experiences in automotive retail for decades is you have 40,000 to 50,000 kind of dealers in a mature, saturated market, and the cost structure and offering is not particularly differentiated. So the way that dealers are able to kind of make their required return, cover their kind of cost base is to kind of monetize every customer as much as possible, which leads to multiple hours in the back room, talking you down on your value of your trade-in, and the 4-square model and all these different things that people have come to dislike over time. So to create a better offering, our belief is that you have to change the cost structure and change the offering. So we are going about that here via vertically integrated e-commerce, where we have moved the transaction online and made it automated and self-service, which allows you to remove the retail footprint from the last mile, which allows you to then bring your inventory into large, centralized reconditioning and distribution centers and create a differentiated supply chain where we've kind of exchanged what was kind of a low fixed cost, high variable cost model for a high fixed cost, low variable cost model. And this centralized inventory that we have and now 10 inspection centers around the country is available to customers in all of our different markets cover about 73% of the population today. And we are bringing that larger selection to bear for all those different customers, where we're moving costs from the system from both the real estate infrastructure and replacing those with trucks in a logistics network and getting rid of kind of the salesman and redundant headcount in each given store. Those cost savings allow us to offer lower prices to our customers while still achieving the same or even better margins, we think, in the long term than traditional brick-and-mortar dealers. And it just allows for a much more seamless, easy experience. So from our perspective, the idea was never about sort of people wanting to get a car delivered or fill forms out online necessarily. It was about sort of leveraging modern technology to create a differentiated supply chain to give a better customer experience.

That's a great, great summary, Mike. Thank you for that. So the used car market, there's north of 40 million units that trade every year. The characteristics of the vehicles that Carvana sells, if you could elaborate a bit on that and what you see as your TAM.

Sure. So I think, for a number of reasons, we sort of think about the whole market as being addressable. And we began kind of selling 0 to 2, 0 to 3-year-old cars when we were originally trying to see if there's product market fit here. The easiest thing to sell online would be new cars. It's just a commodity of known quality. But being unable to do that, we were sort of starting with late-model used cars as kind of the closest thing to that commodity. And as we've kind of grown in scale and expanded our inventory, we have basically broadened our selection to both higher-end vehicles as well as older and higher-mileage cars. So today, we sell vehicles up to about 10 years old, over 100,000 miles and see very consistent demand and reviews and return rates kind of across that entire age spectrum and feel like we're going to continue to expand our inventory breadth and depth over time as we move forward. I think there are certainly good questions about, can you sell a 15-year-old car or a 20-year-old car? What does that mean for customer experience or your brand? I think all of those are totally fair questions. And I think it's, to us, a question of how well we handle merchandising and customer acquisition and setting expectations, but there's no fundamental limitation from the capability of the platform. That car is going to be sold somewhere in the economy. And the odds that somebody is going to get a worse experience as you're probably getting to an older, higher-mileage vehicle is probably even more likely. So we think the opportunity to bring better experiences kind of across the distribution, it is really a big one. I think some people also kind of break up the market in terms of dealer versus peer-to-peer sales. I think that is, to some degree, kind of emergent from different customer preference weighting. So if you are looking for a very specific vehicle and you want to save money and those are kind of your top priorities, you might buy peer to peer, and you're going to give up quality assurance and being able to buy a warranty and probably ease of financing. If those things are more important to you, you might go buy that car from a dealer channel. We think we offer a very broad experience, great values and the ease of financing, ability to offer vehicle service contracts as well. And I think maybe, lastly, thinking about the TAM, this is an interesting market where it is usually, in most cases, a market defined by production and then consumption of that output. This market is -- there's 270-million-plus vehicles in the U.S. economy. And based on the average ownership cycle, which is about 6 to 7 years, people are transacting, that creates that 40 million used vehicles. So if people have a lower cost of ownership, faster time-to-build positive equity, and lower switching costs in terms of time, money and frustration, I think there's lots of signs that people would transact on a faster basis. And you have the opportunity to have x percent of customers but 1.2, 1.3, 1.5x transactions.

That's great, Mike. I'd like to talk about the organic growth opportunities. You're serving a pretty wide swath of the population, 73% today. Your goal, I know, is to get to 95%. But if we could talk specifically about growth within the markets that you're already serving, Atlanta being your most mature market, maybe you could discuss the share gains you're seeing there and how that may translate to other markets.

Sure. I think we've seen very consistent share gains across our markets and cohorts over time and continue to see very strong demand growth, and I think that is driven by a number of different factors that generally get powerful -- more powerful the bigger we get over time. So first is just growing brand awareness and acceptance and understanding of our model. That's driven to a lot of share gains over time, just as you have a better truth. and offer kind of a fundamentally better economic proposition for customers. As we sell more cars, we can hold more vehicles in inventory and increase the selection available to customers, increase the likelihood anybody finds the vehicle that they're looking for, that raises conversion over time. So as you are opening new markets, generally, you've seen them ramp to market penetration levels faster than older markets have. And then older markets also benefit from that continued broadening of selection over time. As we build more inspection centers around the country, our pools of inventory get more proximate to all customers, and average delivery times come down, which also increases conversion. And just the product offering itself improving in terms of the functionality of the website, the ease and the understandability as well as just this behavior of buying online, gaining acceptance and understanding and normalizing, which I think COVID has been an accelerant for. So I think all of those things grow in power as we get bigger and has led to continued share gains over time.

Thanks, Mike. Yes. The consumer has certainly pivoted toward online in a lot of ways. So demand for vehicles, overall, has been strong. It's new cars. It's used cars, seems to be all price points from entry level to $200,000 vehicles. If you could talk about what you think the drivers are of that demand and how sustainable do you think this demand is. Is it a multiyear phenomenon or more short-lived once we come out of the COVID situation that sales are more challenged?

Yes. Honestly, I think you're probably in a better position to forecast that and kind of had some opinions on that. I think there's no doubt that there's probably a lot of different forces at play here in terms of more people looking to get personal transportation as they look to kind of safely make it to their jobs during this time. New vehicle production has been constrained, so supply there has been limited. I think all of those forces are probably leading to a lot of the different macro dynamics that you've seen. We're swimming in the same macro pool as everybody else, so all those waves we're going to feel kind of be subject to. But generally, and I think still today, those forces are just much smaller in magnitude versus the secular share gains that we're seeing. So I don't think we've got a particularly insightful or unique opinion in terms of what's going to be happening from kind of a macro perspective. I think we're just focused on trying to scale our operational supply chain as quickly as we can to meet demand and support growth.

Okay. So all of the franchise dealers that we've got with us here at the conference are all talking about their e-commerce initiatives, and several have freestanding used car concepts where they've got an e-commerce angle as well planned for those. What do you see, I guess, is Carvana's biggest competitive advantages? And has this concerned to you that we're seeing others start to ramp online?

Yes. I mean I think the starting place, it is just that the market is enormous and 40 million transactions and incredibly fragmented. The largest player is less than 2% share. The largest 100 have 7% or 8% share. And the operational intensity to growth means that the ability for any single player to grow into the market limits how much they can actually impact anybody else. Last year, we had the fastest growth in absolute units of any auto retailer in history, and we captured 0.2% of the market. And we think we have clearly the most scalable model and can continue to push that bar higher. So feel like we're in a really good spot there. I think as more and more people recognize the value of kind of doing business in this way, I think that's probably helpful in terms of normalizing that behavior, bringing attention to it. I think it's important for us to continue to differentiate our position within the market and let customers know that this is the best place to buy online and that we created this category, and this is still the highest NPS scores within the industry. I think maybe more fundamentally, we think because the root of the problem was the lack of differentiation and cost structure and supply chain, not so much a consumer demand for extra features that there's, I think, more to do to offer kind of truly differentiated customer experiences here, where adding delivery or sort of filling out forms online wasn't really the pain point for consumers or what they were kind of ultimately kind of looking for. It was just easier overall experience where you don't need to kind of deal with all the kind of things that kind of go on in a dealership and putting together a deal that can often be kind of painful for customers. So I think that people are going to continue to move in our direction, and I think that probably helps normalize behavior, and that's all very good. But beyond that, I think it's just a massive market, and everybody's scale relative to the opportunity is very small.

Okay. So the strong demand for vehicles has led to some outsized GPUs across the sector. I know Carvana saw a nice improvement in retail GPU as well. You've indicated that, that should drop sequentially in the fourth quarter. I'm curious the driver to that. Is that because you're flowing through higher-cost inventory? I know we had some spikes in the wholesale market, and now you're flowing that inventory through cost of goods sold with some reference also to seasonality. I'm curious about the drivers and where you see GPU going over the longer term.

Sure. So I think first is, for anyone who's maybe not familiar, the normal kind of seasonal pattern across the used vehicle market is generally when tax refunds drop, you see kind of a big spike in demand for used cars. And then that kind of level sort of holds through Q2 and then into the beginning of Q3 and then kind of drifts lower from there through the end of the year and into the beginning of the next year. Now vehicle depreciation, sort of how much they depreciate per day is going to be related to that. When demand is highest at the beginning of the year, the depreciation rates of vehicles are lowest. And then generally, in Q4 and early Q1, when industry demand is lowest, vehicle depreciation rates are highest. Simply because of those factors, you will see just regular seasonality in retail GPUs over the course of the year, even if nothing else is changing. So when we were talking about the retail move from Q3 to Q4, we were saying that the move you should see should kind of represent just kind of normal seasonal factors and look more like the move from Q3 to Q4 2018 as we had some specific factors last year that was kind of overwhelming that normal seasonal pattern. I think in Q3 itself, there definitely was an elevated price environment, particularly within the wholesale market. And we definitely saw benefits in our wholesale GPU that we don't think are staying sustainable in the near term. So in wholesale, we expect that to go down Q3 to Q4. Normally, that will be down seasonally. And then I think on top of that, there will be even more kind of correction than usual as price levels are kind of normalizing right now. Now in retail, I don't think we actually saw that much of a benefit from the elevated price environment in the quarter and a couple of reasons behind that. So one is we just bought many, many fewer cars at the trough of the market than many others within the industry. We also have particular pricing policies where we do not write up vehicles in price, even if you are in an appreciating environment, which you saw in certain segments within Q3. So for customer experience reasons, if people are favoriting a car or looking at it or looking to transact, having that car suddenly go up in price the next day is kind of a weird thing to see. So I think because of that, that's one of the things kind of embedded in our pricing logic. Whereas if a dealer doesn't sell a car and they go out on the lot, they can mark it up $500 tomorrow and take advantage of that. So I think that and probably buying more wholesale vehicles as market prices were rising probably led to kind of the fundamental drivers of that benefit in Q3 was really just sourcing many more vehicles from customers up to 56% versus 31% a year ago. So I feel like that is a generally sustainable level and sort of fundamental improvement in our sourcing and other factors.

So as we look at GPU, the major drivers is lowering the days to sale, sourcing more from customers and lowering your refurbishing cost. Where do you think the biggest gains have been achieved? And where are the biggest opportunities for retail GPU?

Yes. I mean I think the biggest gains have very clearly been in just buying more cars from customers. I mean we kind of laid out our long-term goals there, sourcing kind of 38% to 52% of vehicles that we sell retail. 2 years ago and I think we were at 11% or 16% or something like that at the time and are already kind of slightly above that range at 56%. So I think we just have a very strong customer offering there that is leveraging the infrastructure that we've built around the country for the retail offering that is able to scale incredibly quickly. Where moving from Q2 to Q3, we purchased over an incremental 50,000 units from customers quarter-over-quarter, which is just a very large operational feat. So I think the demand, customer experiences we're delivering there is really compelling. We continue to drive that forward. On days to sale, I think we've been about 60 days or so, and I think that feels like a decent near-term level. And then as we kind of move forward in time, we will kind of reduce the operational components of that days to sale, so transportation and reconditioning cycle times as you get shorter distances from any vehicle you're acquiring to the inspection center as you build more of them. And as you kind of scale up your reconditioning facilities and have more tenured employees and lines running at kind of steady state and not always ramping up, you can reduce those cycle times as well. And then on the website, which is really just a choice of how quickly do you want to turn those vehicles and what proportion of vehicles do you want to hold versus the national demand that you're seeing. As you get to larger and larger absolute counts of inventory, you can turn that inventory faster and faster without impacting the effective selection for customers as they come to the website. So I think, structurally, I think that our days to sale can be lower than traditional brick-and-mortar models, especially controlling for similar reconditioning quality being done to the vehicle. And then just as we get larger and more efficient over time, I think there's additional efficiencies to pick up in terms of lower inbound transportation costs, lower direct reconditioning costs in labor and parts as well as higher overhead absorption of the inspection centers as we just put more volume through those facilities.

That's great. [Operator Instructions] Got a couple that I want to ask about. So -- but before we get there, other gross profit, finance fees you collect from lenders, GAP insurance, vehicle, service contracts. You were $1,900 per unit in this segment, the outlook there, the opportunities, as you see it, to expand that other GPU.

Sure. So I think within finance, we were a little over $1,400 in Q3. I think that feels like a good, sustainable level for us long term, especially as we move back into the ABS market and continue to kind of mature within that kind of capital channel. I think in the remainder of that kind of other GPU line is really mostly just vehicle service contracts and GAP waiver coverage that we sell today. Our penetration of GAP waiver coverage, I think, is relatively similar to the industry. Vehicle service contracts, we've made a lot of progress over the last couple of years. I think, historically, our penetration was probably around 25% or so and today is more like high 30s, around 40%, give or take. CarMax on their extended protection plans has penetration of about 60%. Many dealers are in the 50% to kind of 70% range. I think there's certainly additional opportunity to increase penetration of that offering over time as well as broaden the selection of plans that we offer customers for vehicles. I think as we look forward, there's probably many other value-add services that customers are looking for at the time of making a retail purchase that could be interesting to explore, whether that's prepaid maintenance or vehicle insurance or things like that, that I think could also be interesting to add to that offering. But I think we are probably going to offer fewer products there than many traditional dealers, I think, as we want to make sure that we're always kind of balancing our unit economics with customer experience and what really adds value to the customer. And we don't want to sort of fill the transaction and then also the loans with kind of just sort of air products that we don't think really are desirable for customers.