Good day, welcome to the Carvana second quarter 2019 earnings conference call. All participants will be in a listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key on your telephone keypad. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star, then one on your telephone keypad. To withdraw your question, please press star, then two. In the interest of time, we ask that you limit yourself to one question and one follow-up before reentering the queue. Please note this event is being recorded. I would now like to turn the conference over to Michael Levin, Vice President of Investor Relations. Please go ahead.
Thank you, Sean. Good afternoon, ladies and gentlemen, and thank you for joining us on Carvana's second quarter 2019 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 second quarter shareholder letter is also posted on the IR website. Joining me on the call today are Ernie Garcia, Chief Executive Officer, and Mark Jenkins, Chief Financial Officer. Before we start, I would like to remind you that the following discussion contains forward-looking statements within the meaning of the federal securities laws, including, but not limited to, Carvana's market opportunities and future financial results that involve risks and uncertainties that may cause actual results to differ materially from those discussed here.
A detailed discussion of the material factors that cause actual results to differ from forward-looking statements can be found in the Risk Factors section of Carvana's most recent Form 10-K and Form 10-Q. The forward-looking statements and risks in this conference call are based on current expectations as of today, and Carvana assumes no obligation to update or revise them, whether as a result of new developments or otherwise. Our commentary today will include non-GAAP financial measures, including, but not limited to, ex-Gift measures that exclude the impact of the 100,000 milestone gift to our employees. Reconciliations between GAAP and non-GAAP metrics for our reported results can be found in our shareholder letter issued today, a copy of which can be found on our investor relations website. Please note that all gross profit, SG&A, and EBITDA metrics mentioned by us on the call today are on an ex-Gift basis.
Now, with that said, I'd like to turn the call over to Ernie Garcia. Ernie?
Thanks, Mike, thanks, everyone, for joining the call. Q2 was a pretty remarkable quarter for us. We continued making significant progress across the business in achieving our mission of changing the way people buy cars. Notably, the quarter included the very important and very visible milestone of crossing our midterm goal of $3,000 GPU. Let's start by putting that in context. In 2016, our fourth year in business, only three years ago, we achieved total GPU of $1,023. On our first public conference call in June 2017, we outlined our path to a $3,000 midterm target. We knew the business, we knew we had to do to generate those kinds of unit economics, we had a plan. Admittedly, that goal seemed a long way off to many. We sit here just two years later couldn't be prouder to report $3,175 in total GPU.
How did we get here? The first step was a simple one to plan and a hard one to execute. We built a customer-centric company that has the technology, operations, and culture to deliver incredible customer experiences. We are organized around this goal. Customer centricity isn't a compelling storyline for us, it is who we are and what we are built to be. Secondly, we think long term. We don't let expediency and ease push us into mediocrity. We are building the best solutions we can for our customers, and we aren't afraid to make the investments those solutions require. The more we learn, the more we appreciate just how important and just how big of a differentiator this is. Thirdly, we are thoughtfully designing the company to ensure great unit economics are possible.
Through vertical integration and through leveraging technology, we are taking traditionally high variable costs and lowering and fixing them. Next, we prioritize our goals carefully through our lens of customer centricity and long-term thinking. Finally, and very importantly, we've executed. That path has led us to where we sit today. What's next? What's the next goal we'll be focusing on? From here, our folks will be squarely aimed at achieving our goals of delivering 2 million-plus cars per year and hitting our long-term financial models. We arrived at these goals through the same careful bottoms-up analysis that led to our $3,000 midterm GPU goal, we believe these goals are similarly achievable with similar risk, namely execution risk. If we continue to execute the way we have, our future is extremely bright. I believe we will continue to execute.
The primary reason for this is that we have assembled a team that relentlessly strives to be a little bit better every day. We keep our eyes on tomorrow with the same scrappy tenacity we did at the very beginning and have every intention of continuing to do so. Seeing the quality of people we have across functions and physical locations is what drives my confidence. Mark went more on units, revenue, GPU, and EBITDA margin, all of which were exceptional this quarter. I would like to quickly call out one highlight. For the first time in our history, we had decreasing year-over-year EBITDA dollar losses. We did this while simultaneously posting 95% unit growth, 108% revenue growth, 134% growth in our total number of customers served, and 188% growth in buying cars from our customers. This isn't easy to do, and it isn't something that's seen very often.
We did all this while also accelerating our investments in our future through all the channels you'd expect. This powered the most exciting product pipeline we've ever had. All the product we have been building has been carefully considered and thoughtfully designed by passionate, capable people. We don't yet know what will come of all the improvements we're building today, but we're excited by their prospects. We do know that many of the products we have built over the last several years are starting to pay significant, visible dividends now. One of the most recent examples of this is buying cars from our customers. In early 2018, we began discussing the investments we were making in this area.
Since Q1 2018, we have made steady progress toward growing total units purchased from our customers from about 4,500 in that quarter to about 22,800 in Q2. During that time, we've moved from buying 24% as many cars from customers as we were selling to them to 52% in Q2. That is pretty significant progress happening over just five quarters, and it means that we are now buying as many cars from our customers as we were selling just one year ago. The contribution of these efforts to our total GPU have been significant as well and played a major part in hitting our midterm target this quarter. Just like on the retail side of the business, the root of this growth lies in a clean, simple experience for our customers. Just like in the retail business, we believe this is a powerful and fundamental root.
We continue to believe this is an area we are just beginning to really tap into and understand. Knowing exactly where we'll end up here and exactly when we will get there is difficult at this stage, but there's no doubt we have a lot of momentum and there's a lot of unrealized potential. I'll close the same way I opened. This is a pretty exceptional quarter across the business. While these results are extremely exciting, what we believe is more exciting is zooming out and looking at all the progress that is happening on a larger time scale. First of all, we are just a six-year-old company. It's worth pausing and thinking about all the implications of that much progress that quickly. Over a several year period, units have grown exponentially and GPU and EBITDA margin have both improved linearly.
What drives these results are incredible people delivering incredible experiences with a focus on the long term. What the consistency of the results suggest is that it is still very early days building into a very large opportunity. We are a team of builders and are excited by what we see in front of us. We'll keep building, experimenting, and learning, and we'll keep you informed along the way. Mark?
Thank you, Ernie, and thank you all for joining us today. Unless otherwise noted, all comparisons are on a year-over-year basis. We are pleased to report another quarter of exceptional growth in both retail units and revenue. Retail units totaled 44,000 in Q2, an increase of 95%. Total revenue was $986.2 million, an increase of 108%. This marks our 22nd consecutive quarter of triple-digit revenue growth. Total gross profit per unit exceeded our midterm goal at $3,175 in Q2, an increase of $1,002. The second quarter continued our theme of broad-based GPU growth, with record GPU across all parts of the transaction. Retail GPU increased by $302, reflecting gains in acquiring cars from customers, incremental shipping revenue, and lower average days to sale. Wholesale GPU increased by $88, driven by 194% growth in wholesale units, and a record gross profit per wholesale unit of $660.
Finally, other GPU increased by $613, reflecting gains in finance monetization and increased attachment of ancillary products. Our gains in GPU this quarter were bolstered by significant growth in buying cars from customers. To support this growth, we are making investments in advertising, technology, and operations related to this offering. These investments are already paying dividends in the form of meaningful contributions to revenue and total GPU. Even more importantly, we believe buying cars from customers is a significant fundamental step toward achieving our long-term goals. EBITDA margin was -3.3% in Q2, an improvement of approximately 5.5 percentage points, reflecting gains in total GPU and SG&A. Our 3,000 GPU milestone and record EBITDA margin make this a good time to look back at what we have accomplished over the past three years. In 2016, the year before we went public, we had an EBITDA margin of -23.2%.
This quarter, it was -3.3%. Over that time period, we improved EBITDA margin by nearly 20 percentage points, including a nearly nine percentage point improvement in gross margin and a more than 11 percentage point improvement in SG&A. We are proud of the progress that we've made and are excited about what this progress means for our goal of becoming the largest and most profitable auto retailer. We ended the quarter with more than $750 million in committed liquidity resources and held an incremental $80 million in real estate and securities on our balance sheet, giving us substantial flexibility to execute our plan. We opened 28 markets in Q2, bringing our current total to 137, up from 85 at year-end. We continue to expect to open 55 to 60 markets in 2019, bringing our year-end total to 140 to 145.
After a record first half of market openings, in the back half of the year, we plan to turn our focus to preparing for growth in the first half of 2020 and for buying more cars from customers. In Q2, we began production at our seventh inspection and reconditioning center in Nashville, Tennessee, raising our production capacity to approximately 350,000 vehicles per year at full utilization. We continue to view inspection and reconditioning centers as a long-term competitive advantage as we further expand our as soon as next day delivery infrastructure. Q2 marked the most successful quarter to date for our finance platform. On June 27th, we closed our second auto loan securitization, further expanding and diversifying our investor base. Finance GPU was a record $1,100 in Q2, an increase of $526.
This benefited from a $100-$150 impact from a decline in benchmark interest rates during the quarter, but was our highest ever by a significant margin, both with and without this effect. We are excited about what this means for our finance platform and expect to recognize additional gains over time. In terms of outlook, we are raising our full-year guidance for retail units sold to 167,500-172,500, and total revenue to $3.6 billion-$3.7 billion. We are also raising our guidance for total GPU and reiterating our guidance for EBITDA margin based on our strong results in the first half and our view of the many exciting opportunities ahead. We expect approximately 155 million weighted average shares on a fully exchanged basis in Q3. As we look forward to the remainder of 2019, we are excited about our progress toward our long-term financial goals.
Thanks for your attention, and we'll now take questions.
Thank you. We will now begin the question and answer session. To ask a question, you may press star, then one on your telephone keypad. If you are using a speakerphone, please pick up your handset before pressing the keys. As a courtesy to make sure we receive all questions, please limit yourself to one question and one follow-up. If at any time your question has been addressed and you would like to withdraw your question, please press star, then two. Our first question today will come from Sharon Zackfia with William Blair. Please go ahead.
Hi, good afternoon, and congratulations. That was a really big ramp-up in profitability really quickly.
Thank you.
I guess, question sequentially, though. If I'm doing the math right, I think the all-in GPU will decelerate by maybe $300-$350 in the back half at the higher end of your GPU guidance. In addition to the finance, where I guess you were indicating like $100-$150 may not be sustainable, where is that other couple hundred bucks coming from sequentially in the back half?
Sure.
Meaning decelerating.
Yeah, sure. I think there's a couple of things that we're taking into consideration in our GPU guidance, and obviously we're very excited about where we came out in the quarter and very excited about where we expect to come out for the year. The two things that we're taking into consideration are, one, seasonally, the used car market is typically weaker in the back half than the front half of the year. That relates to typically depreciation rates in the used vehicle market tend to be higher in the back half than in the first half, we're taking that into consideration. Moreover, we typically in Q4 have historically done a Cyber Monday promotion, which has an impact on retail GPU, other things equal. I think that's the first point.
The second point you alluded to, which is we think we benefited this quarter to the tune of approximately $100-$150 in Finance GPU due to a quick decline in benchmark interest rates within the quarter. Those would be two things I'd point to. Of course, we're obviously very excited with our GPU progress. Breaking through the 3,000 midterm goal was a very exciting milestone this quarter. I think increasing our guidance by $200 on both ends for the year relative to where we expected to be at the beginning of the year, I think is very exciting. I think we're looking at our third consecutive year of more than $500 year-over-year GPU growth as a public company, and so I think all of those trends are leaving us feeling very good right now.
Sharon, this is Ernie. All I would add is I think we improved by over $1,000 year-over-year this quarter. That's pretty incredible growth, and it was driven by, as Mark said, significant progress across all the different elements of GPU. I do think when we look at it on comp through the rest of the year, last year in Q3 and Q4, we made some pretty significant positive moves as well. It's important to take that into account. The moves we see from here are normal and seasonal with the exception of that $100-$150 delta because of interest rates.
I think secondarily, just to follow up, Ernie, last year you had talked quite a bit about roadblocks that you had to overcome in growth. I think a lot on the logistics side to fulfill customer demand. Where are you on roadblocks at this point? If you had to pinpoint where there are obstacles to growth, I'd just be curious on what those are.
Sure. I think the term that we used last year was pinch points. I think what we mean by that is basically throughout the business, there's a bunch of different operational areas that we need to make sure that we're tight on and executing well in order to deliver the best possible customer experience. The best customer experience is they go to our website, and they find the car they're looking for. If we're selling a lot more cars than we expected, sometimes our inventory can get lighter, and that can reduce sales versus what they otherwise might have been. We need to have a logistics network that's functioning well and is able to deliver cars quickly to customers everywhere.
If there are parts of the logistics network that are a little bit tight, we may have delivery times that are a little longer than they otherwise would be, and so that can reduce conversion. If we're getting a lot of calls into advocates with questions about how the process works or wanting to sell a car to us, and we're not able to handle all that volume, that can lead to reductions in volume. There's a lot of places where that can show up that really exists on a continuum. It's not really a binary that exists anywhere. I think today we're in a pretty good spot.
I do think that the 188% growth in buying cars from our customers has been faster than we anticipated. There are a couple of areas in the business where we're probably not running quite as tightly as we wish, I think that that's coming from a really great place. I think in general, we're in a good spot.
Okay. Thank you.
Our next question will come from Zachary Fadem with Wells Fargo. Please go ahead.
Hey, guys. Congrats on the progress. First question on the guidance. You took unit sales and GPU all up for the full year. Curious if you could differentiate which areas of the increases are driven by flowing through the outperformance in the quarter, and what part of the new outlook is driven by changes on your view on the back half of the year?
Sure. I think across the board, we saw a lot of progress in the first half in general and then specifically in Q2.
That we're very excited about. We raised units in Q1, we're raising it again now. We're passing through a big raise in GPU, we're holding EBITDA margin. Everything across the board looks really good to us, I think it's important to walk through the differentiators there. In units, we're seeing great volume across the board, across cohorts, across markets. It looks really strong. We're very excited about that, we're able to pass that through. In GPU, I think there's several contributors. One that is definitely worth calling out is buying cars from our customers. We posted our largest-ever Wholesale GPU contribution there. It was about $160, give or take. The other place that buying cars from customers flows through to GPU is when we buy cars from customers and then sell those cars retail.
That can't be separately broken out in our financials. It's also a big contributor of somewhat similar size. A big part of the beat in the quarter and a beat that we anticipate for the remainder of the year is coming from that business of buying cars from customers. That business is growing at 188%, and it also involves some investment that we're accelerating now given the strength that we're seeing there. On the order of a fifth, give or take, of our marketing budget, is now going toward the business of buying cars from customers. There are also technology investments and then operational investments and costs associated with running that business. It's clearly contributing a lot of GPU. It also has expenses associated with it, and that's why you're seeing the big GPU raise, without an equivalent push-through on EBITDA.
Overall, we view that as extremely positive.
Got it. That's really helpful. Bigger picture question, you plan to end the year at 67% population coverage, basically where we are today. Curious if you could talk through what's holding you back from accelerating to that 80%, 90% level, and maybe talk about game plan for bridging the gap in the regions like the Pacific Northwest where you may be under indexed.
Yeah. I think everything that we continue to see there is really strong, both in terms of the way that those markets are responding that we're opening up and the ease with which and speed with which we're able to open those markets. We opened over 50 markets in the first half. Those markets are our fastest-ramping markets that we've ever had, in the company's life. We're very excited by that. As we look forward to the first half of next year, we do expect to see significant volume increases in 2020. With all the progress that we're seeing in our business of buying cars from customers, we want to make sure that we're thoughtful, that we don't put ourselves in a spot where we are really constrained on the operations side.
We're putting more of our focus today on preparing for that growth and preparing for all the growth in buying cars from customers. We anticipate rapidly opening markets again thereafter. I think, given what we've seen over the last year, if anything, our eyes have gotten bigger for the percentage of population we can ultimately serve and maybe even the speed with which we can serve it. For now, we're going to slow down a bit and make sure we stay focused on ops so we don't get over our skis too far.
Got it. Makes sense to me. Appreciate the time.
Thank you.
Our next question will come from Chris Bottiglieri with Wolfe Research. Please go ahead.
Hey, guys. Thanks for taking the question.
Of course.
A two-part question. One on financing. Really impressive growth. Curious, though, why you expect to give back the 100-150. When I look at the interest rate curves, all I see is a pretty steep slope and down-looking curve. If rates stay where they are today, and we get another rate cut even, would you expect to still give back that 1-150? The message is that you're passing it back to the customer and APR is going lower. How do you balance that?
That's exactly right. I think you hit the answer to your question at the end. Were we to hold rates flat, a move down in benchmark rates would cause that 100-150 to persist. The reason we have called it out as being transitory is we've passed a lot of that rate decline along to customers.
Gotcha. Okay. Big picture question. Last year, for the debt holders, you gave them this pretty impressive chart showing market contribution profitability. It showed Atlanta with just under $1,300 of EBITDA and a fully loaded EBITDA of a little under $500. Given your total GPU is $1,000 higher today than it was when you gave that disclosure, I'd be curious to hear if you can give us an update on Atlanta profitability, at the contribution level.
Sure. I think at this point, we're at 3.3% EBITDA margin loss for the entire business, obviously showing a lot of leverage, and I think we're going to keep our EBITDA level disclosures at that level instead of diving into the market level. You can go look at that chart, and the way it works is not all that complicated. Generally speaking, the GPU is roughly shared across markets, so if we have gains in any given market, that's on average going to flow through to all the markets. Then as we get bigger and lever corporate costs more, that generally benefits the markets. As we continue to lever marketing, that tends to benefit the markets. I think when we look at it across the markets, there's really good stories there, and we think the path to profitability is clearer than it's ever been.
I think even stepping back and saying, why did we set this $3,000 goal to begin with when we first came public in early 2017? I think there were two reasons why we set that goal. One is we thought through it carefully. We felt like we understood all the levers in the business and where we could go and get money, and we felt that $3,000 was a level that made sense and was achievable given the business that we had built. Two, and really importantly, we thought the $3,000 was a very significant level in terms of being a number where you could build a sustainable and very profitable business, and we continue to believe that today. I think there's a lot of ways to look at our financials today. I'll call out three.
One is just to look at the momentum and progress that we've got. Mark talked about levering nearly 20 points over the last three years and being at 3.3% EBITDA margin loss today. That suggested that there's a lot of room that we're very, very excited about. Another approach is just like we approach the $3,000 GPU target, we can do a bottoms-up analysis, and that basically is our long-term financial model. We went through that very carefully, looked at all the expense items, all the areas we thought that we could add additional GPU. We put that together, and I hope at least that from an objective perspective, you could go through and do a similar bottoms-up analysis and arrive at similar places as we have in our long-term model. I think third, and interestingly, is do a relative comparison to dealerships out there.
If you go look at any given dealership, any of the publics, with some of them, this comparison will be very easy. With some, it's hard because you need to get their SG&A per retail unit and kind of massage out service ops. If you go look at that, most dealerships are spending around $2,300 in SG&A per car they sell. In the quarter, we spent $4,100 in SG&A per car we sold. That gap is paying for 95% unit growth and 108% revenue growth, and we're still significantly sub-scale. Even at that level, our losses, if you take our EBITDA loss in the quarter and divide by sales, they were about $750 per car.
If you basically assume in the long run we can get to a cost structure that looks similar to dealerships, there's an $1,800 benefit there, and now you're suddenly $1,000 positive, which is about 5% EBITDA margin. Obviously, we believe in the long run, our cost structure looks a lot different than dealerships, and it is much better than most dealerships. In addition, we still think there's a lot of room in GPU. We're looking at that and feeling very, very good, and we think just given the clarity of the story now, we're going to stick with speaking about it at a company level.
Got you. That makes sense. Thank you for the color.
Yeah. Thank you.
Our next question will come from Nick Jones with Citi. Please go ahead.
Hi. Thank you. I found in the letter to shareholders, the comment on serving over 60,000 customers in the quarter, and then doing 44,000 retail units. I guess, I interpret it that you maybe acquired more than 17% from customers, and that maybe there's some friction or inefficiencies or the days to sell is a little bit longer on sourced vehicles, and there may be ways to kind of ratchet up that 17% without necessarily sourcing significantly more vehicles? Or can you help me unpack that sentence?
Sure. There's a lot going on there. Let's start with what the 60,000 number is. That's basically the sum of customers that we transacted with. That includes all the customers that we sold cars to, that's 44,000, and then also all the customers that we bought cars from that we didn't also sell a car to. Those are kind of unique additional customers that are just selling us a car without also simultaneously buying one from us. When you add that up, you get to 60,000. That number of total customers transacted with is the number that grew by 134% in the quarter. That's kind of what that number is. The number of cars that we bought from customers and then sold to customers was 17% in the quarter, up from 14% last quarter.
On average, this is a heuristic, so this isn't exactly right, but on average, about half the cars that we're buying from customers meet our retail standards and can be sold to customers. In the quarter, we bought 52% as many cars from customers in total, including cars that we bought without an associated sale and cars that were traded into us. We bought 52% as many cars as we sold to customers. You'd expect the percentage of cars that we're selling retail to be a little bit higher than that 17%, but there's kind of a natural lag there because there is a lag time from when we buy those cars until they make it through the reconditioning center and get up on the website, or ultimately purchased by customers. There's some nice positive momentum there as well.
Got it. Thank you for taking my question.
Thank you.
Our next question will come from Seth Basham with Wedbush Securities. Please go ahead.
Thanks a lot, and good afternoon.
Hey, how you doing?
Good. My first question is around the strength in the quarter. Your units probably topped your internal plan. Can you give us a sense of how much they beat your internal plan? Secondly, what you think drove that upside? That would be helpful.
I think, we're going to stick with the guidance that we've provided in the past. I think the fact that we're raising does suggest that we outpaced what we were anticipating in our guidance. I think that's great news. We think that what's driving it really in any given quarter, I think you can zoom into all kinds of little drivers, but what we think is most important is we're delivering differentiated, very high-quality experiences to customers. We're giving them a bigger selection they see anywhere else. We're giving them more value they see anywhere else. The team's executing very well. That's showing up in broad-based demand across the market. I don't know that there was anything macro that felt unique that would raise the level of being called out separately in the quarter.
Okay. Fair enough. Secondly, in terms of your Finance GPU, should we think about $1,000 as your normalized run rate here? Is there any change in penetration rate that you experienced in the quarter or anything else to affect that run rate other than the interest rate call-out?
We printed $1,100 in the quarter, and then as we said, there was about $100, $150 impact due to interest rates that we would not expect to persist. That would get you down to about the range that you were talking about. We think that there are also additional gains to be had from here of the same flavor of the gains that we realized this quarter. I think we would point to our long-term financial model there. We think there's probably a couple hundred dollars of additional gains from Q2 levels. We would expect a step back of $100, $150, and then to, over time, continue our positive walk forward. There's a number of drivers there. In general, we've seen a very slow drift over time, positively in attachment rate of financing.
The far more powerful drivers have been just getting better execution over time, as we've outlined before.
Thanks a lot.
Thank you.
Our next question will come from Ron Josey with JMP Securities. Please go ahead.
Hi, this is Andrew Boone for Ron. Two questions please. Trade-in's just going really well. Can you talk about the drivers there, whether that's better conversion, the marketing campaign, just whatever's behind that? Secondarily, advertising came in a little ahead of us. Understood the trade-in campaign, is there a cap for national advertising or are you guys spending more on performance or is there anything there to call out? Thank you.
Great. Let's start with buying cars from customers. I think what's driving that at a high level is just that experience is incredibly simple, and that experience is also pretty easy to understand. I think buying a car online is not a super easy thing to communicate to customers. If you say, what does buying a car online mean to 100 customers, you might get 100 different definitions. If you say, "Do you want to go to the website and get a value for your car and then we'll pick it up and we'll just send you money right away," people understand how that works. That's a fairly straightforward thing to explain to customers, and then the offering, I think is about as good as you could hope for from a customer perspective. We think that's the root of what's driving it.
It's a very simple experience, and it's simple to understand, and we think that that's great. If we zoom in a little bit more, we're definitely making investments in telling that story in marketing. That gets to your second question, and we'll circle back to that. We're definitely making investments in technology throughout the funnel and the website experience. It's still very early days. We're still learning meaningful things every time we run a test on the website that are changing the way that we're approaching this problem. We're still learning a lot about pricing. We're not even leveraging all the depths of our data that we've got today to make sure that we're pricing as intelligently as we can. We still have a lot to learn because our historical volumes are pretty small compared to the volumes that we're seeing today.
We think there's tons of gains to be had there. We're still learning a lot on the operations side. We're making changes to our scheduler to make delivery and pickup simpler and faster for customers. I think there's just broad-based gains all over the place. I think, just to circle back to the advertising question, we definitely are leaning into this, and advertising is more because it is a simple story and because we are seeing incredible growth. When we look at the unit economics of incremental advertising in the business of buying cars from our customers, and then we compare that to the GPU gains that we're seeing, we think it's a really good trade, especially when you take into account all the growth that we're seeing.
That's definitely baked into our guidance for the remainder of the year, is the expectation that we will continue to invest aggressively this offering that we've got.
Great. Thank you.
Thanks.
Our next question will come from Rick Nelson with Stephens. Please go ahead.
Thanks. Good afternoon. As you source more cars from the customers directly, with half of those feeding retail, are you migrating to older, higher mileage cars on your website? If that's the case, are you seeing any change in the number of vehicles returned or the proportion of vehicles returned?
Great question. I think the answer to the first part of your question is largely yes. We're seeing a different distribution of cars that are available to us as a result of buying cars from customers. I think that that particular channel does weakly push you toward having more access to lower cost inventory, which is great. We also generally see a broadening of the distribution of cars that we're selling on the other side, where we're selling more expensive cars as well. I do think that there's overall a broadening of the cars that we're selling, which is happening slowly, but has been part of a continual migration over a longer period of time, and it's certainly been accelerated by buying more cars from customers.
In terms of return rates and then customer reviews, which are both things that we monitor pretty closely at the car level to make sure that we're understanding how different classes of inventory are causing different customers to get different experiences. We don't see anything material there. The directions tend to be what you'd expect. We see very slightly higher return rates on older cars. Oftentimes we see lower reviews on older cars, but that's not even a very strong relationship at all. I'm not even sure I would confidently call out the direction there. In general, through the broadening of the distribution of cars that we're selling, we're just getting more confidence that we can sell a very broad distribution of cars to customers over time.
Great. If I could ask as a follow-up, as you grow this internal sourcing, it's going to push more cars to wholesale. Is there a certain size or height that you need to get to operate your own auctions, or is that part of the long-term game plan?
That's an interesting question. First, there's clearly significant economics associated with selling a car wholesale through an auction that you don't own. There's a buy fee and a sell fee, which can add up to $300, $400. There's transportation on both sides that can be pretty significant. That's undoubtedly an area of opportunity. We also have great partnerships with the auctions with which we do business today. I think, in general, our instinct will always be to seek to work with our partners first and find arrangements that work out well for both of us, so we can stay focused on the many other exciting opportunities that we have in front of us. I think that would be our starting point, and that's not an enormous strategic focus today, but the point is a good one.
Thanks. Good luck.
Thank you.
Our next question will come from Colin Sebastian with Robert Baird. Please go ahead.
Thanks. Good afternoon. As you shift growth from new market openings to market penetration. Wondering how you will alter the approach in terms of engagement with potential customers, and how much of the investment in the website and app are generating improved metrics such as conversion rates? I have one follow-up. Thanks.
Sure. There's a lot going on there. I think as we open up more markets, which to start with, we think there's still quite a few more markets out there. There's still significant swaths of population that we can service. Once we make it through that population, we transition to just growing existing markets. I'm not sure that fundamentally our engagement strategy changes materially. We still remain at very small levels of awareness and understanding, even in our most mature markets. We think there's just a lot of gains to be had there. There's a lot of advertising to be done and awareness to be accumulated over time. There's a lot of learning on our part to figure out how to communicate to broader swaths of customers more efficiently. I don't think that there's a ton that will change there.
I think it'll be part of a continual migration and a continual learning on the marketing side because we clearly just have a lot of headroom there. We're making all kinds of changes all the time across the website. I think, I said in my prepared remarks, but I really do think that this point is oftentimes at least lost a little bit. We are a six-year-old company, and that means that nothing we do is perfectly refined in any way, shape, or form. We're still very much at a stage where we roll out changes, and we see big significant moves that matter and move the needle, and that happens at a pretty regular cadence. I don't even think we've seen that cadence slow down because our offering is broad.
There's a lot of technology and customer experience associated with buying cars, reconditioning cars, offering a logistics network so we can deliver them next day, giving them a good experience on the website, offering a simple finance experience. There's a lot there. There's so many areas for us to keep exploring, and we've got great teams that are exploring all those areas, and we're continually finding significant improvements. If you define conversion strictly as website to sale, that's a complicated measure for sure, because the quality of traffic is constantly changing depending on what marketing channels you're using, and then depending on if you're making progress in areas like SEO, where oftentimes you're driving a lot of value, but you may be driving lower converting traffic.
I think when we think about conversion and all the comments I made about conversion before, those comments are kind of holding website traffic static. Over time, our website traffic is likely to move in both directions. As we get better at things like SEO, we'll probably drive more volume of traffic, but it may be a little bit lower quality, and therefore you may see lower conversion rates when you're measuring it that way. As we have other innovations that are more direct and hit in-market buyers, you may see conversion go up. I think there are those two separable effects, and I would just make sure that you're thinking about them separately because I do think they're pretty independent.
Okay, that's helpful. The clarification or the follow-up is on the GPU contribution from the finance platform. Just wanted to clarify, outside of the impact from the lower benchmark rates, should we expect the improvement in finance GPU to sustain itself on a sequential basis? Were there benefits from the timing of securitizations or other factors that mean that might be a little bit lumpy? Thanks.
Sure. We called out the benchmark interest rate impact in our remarks and I think also in our questions. I wouldn't go into more specifics in terms of breaking down our guidance for the rest of the year. I would say that over time, we certainly see more opportunity to improve execution. We're still relatively early in our life as a securitization market participant. We think that leads to a lot of opportunities as we look forward over time.
Okay. Thanks, Mark.
Our next question will come from Tom Champion with Cowen. Please go ahead.
Hi. Good afternoon. Thanks for taking the questions. Apologies if you've already addressed this, but I'm curious on the other side of the capital raise in May, where you're really leaning in terms of investment. It seems like this could be on the new market side or IRCs or logistics. Just curious, any comments on that? Second, maybe dovetailing with the last question. You completed your second ABS transaction. Can you talk about the implications of that and your reliance on that channel as a part of your portfolio of financing sources? Thank you.
Sure. As it relates to the capital we raised and what that's unlocking, it clearly does unlock additional aggression if we so chose. I think so far we have not elected to be more aggressive in any material areas. The one area where I think we're being a little bit more aggressive today is buying cars from our customers. That is largely self-funding and paying for itself. I don't think that's an incremental area of investment. I think we thought about bringing that capital as just buying us a lot of financial flexibility. There were two forms of capital that we brought in there. There was high-yield debt, there was a slug of equity.
The high-yield debt for us is very efficient capital because it basically replaces sale leaseback financing that we would do to finance our CapEx otherwise and has a similar cost and a shorter duration. It's just generally more flexible. We largely raised that for that purpose. It's just kind of a dominant form of capital that we like having on the balance sheet. The equity we took on the order of two points of dilution, give or take. I think we thought that was a smart thing to do just from a cushion perspective to make sure we're prepared for bumps in the road. We really haven't changed our tack there in any material way. On the ABS transactions, I think we're very excited about that.
In general, the way that we like to think about our finance platform is we want to optimize both the monetization of finance and the stability of our finance platform, our finance partners. I think, in the quarter, we ended with a mix of Ally and the securitization market on the order of 99% of the loans that we sold in the quarter went to those two purchasers. We think that was a good mix because we've got a lot of stability, and we are also able to show a lot of progress in monetization. We think that that looks good, and I think looking forward, we'll continue with our strategy of optimizing monetization and stability. We don't want to commit specifically to any given channel today. We feel great about the flexibility we've built into the program over the last several quarters.
Makes sense. Thank you.
Thank you.
Our next question comes from Armintas Sinkevicius with Morgan Stanley. Please go ahead.
Great. Appreciate you taking the question. When I look at $1,500 of other GPU, and I think this question's been asked in many different forms, but the $1,500 versus the roughly $1,000 in the first quarter, so a $500 sequential step up. You mentioned the $100, $150 from the benchmark. What's the other $350-$400 related to for other gross profit per unit?
Sure. Yeah. It's primarily just better execution on our loans. I think we introduced ourselves to the securitization market in the first quarter, with our inaugural deal. We went back to the securitization market for the second time this quarter, brought in more investors, diversified our investor base, and saw a higher gain on loan sale than we saw in the previous quarter. I think the better execution is the big driver. There's some other smaller drivers. For example, fees per unit were lower, as expected this quarter, because we had fewer first-time deal expenses than we did a larger deal. Just better execution on our loans was the biggest driver of the sequential increase in finance GPU.
Got it. Okay. Call it the $350-$400 then, sequential improvement, x the benchmark rates. At the Analyst Day, you talked about a $500-$700 improvement from lower cost of financing. Is this $350-$400 part of that $500-$700, or is this apples and oranges? You mentioned a couple hundred dollars, which would sort of line up with the commentary from the Analyst Day of $500-$700.
Yeah. I think you're thinking about that the right way. I think we've made a nice step here, both including and excluding the inter-quarter benchmark move, but certainly feel like we've got some potential increases in the future as we mature in the securitization market, continue to demonstrate strong performance, increase liquidity of our program, et cetera.
Got it. One more, with the finance receivables last quarter, that was about a $50 million burn for the quarter. This quarter, it was a $90 million source of cash. Maybe you could talk about some of the puts and takes there on finance receivables, and how to think about that as a source or use of cash going forward.
Sure. We review finance receivables very similar to inventory. That's working capital that auto dealers or other auto industry participants have as part of their standard operations of the business. Now, both inventory and finance receivables are heavily financeable with standard working capital facilities. Much like we've done with inventory over time, we expect to the extent that finance receivables are moving up and down, finance the vast majority of them with working capital liquidity from short-term working capital facilities.
Got it. Much appreciated.
Thanks.
Our next question will come from Lee Krowl with B. Riley FBR. Please go ahead.
Great. Thanks for taking my questions, and congrats on a very solid quarter.
Thank you.
Two questions. First, just on the buying versus selling ratio, obviously, doubling year-over-year and up from 40% last quarter to 52%. You guys highlighted that the opportunity out in the market, from competitors is 1 to 1. Thoughts on if you believe that's achievable with Carvana, and if so, maybe perhaps a timeline and the GPU implications of that?
Sure. I think the market leaders buy from customers about as many cars as they sell to customers. That's clearly achievable in the sense that it's been done before. There's not really a logical ceiling on what that number can be. Trade-ins are mechanically related to the number of cars you're selling, and so you're necessarily limited to less than one there. Buying cars from customers that are not associated with the sale, leaves you in an unbounded place. I don't think we know exactly where it can go. What we know is that we're delivering great experiences to customers. It's growing very quickly. We're learning very fast. There's clearly a lot of momentum.
When we look at what the market leaders are doing and say, given what we've seen so far, do we think that over time there's the opportunity to buy as many cars from customers as we're selling, and do we think that over time there's the opportunity to make about $1,000, give or take, per unit that we purchase from customers? We think that that looks very achievable from where we sit today. We also think that there's a lot of work left to do. We've made a lot of progress recently. We're going to keep working at it. I don't think we want to give any more concrete guidance than we provided at Analyst Day with our long-term financial model, which made those same assumptions that I just outlined.
Got it. Just on the locations increase, obviously heavily weighted to the first half, just your thoughts on the investments as it relates to fewer location openings in the second half and whether the dollar value of investments remains the same and is just allocated elsewhere internally, be it technology or marketing, or if the dollar investments to scale the business step down?
What I would say is that I think we think about those as somewhat independent choices. What the choice to open markets more slowly in the back half is largely driven by operations and by preparing ourselves for the first half of 2020, and wanting to make sure that we stay in front of ops there and continue to deliver a very high-quality experience to customers. That's what's driving that. The investments that we make internally, in process improvements or operations or technology, we tend to make those pretty independently, and we don't manage with a quarter-to-quarter total investment budget in mind. We're trying to do the right thing for the business at all points in time and then trying to recognize that we do have limited resources at our disposal, and so we need to allocate them as intelligently as we can.
I think we think about those things separately, but in the back half of this year, we've elected to put more of our operational focus on preparing for growth in early 2020. That probably, in a vacuum, will not impact the way that we're thinking about investments in technology.
Got it. That makes sense. Thank you for taking my questions.
Thank you.
Our final question will come from Dan Salmon with BMO Capital Markets. Please go ahead.
Afternoon, everyone. Just curious, Ernie, as you really expand the footprint out and cover a bigger part of the country, obviously we're not going to talk about 2020 just yet, but expecting that your launch rate picks back up again after the preparation that you're doing for the first half next year, it's not very long, I think, before you do complete your goal of covering the entire U.S. footprint, or at least the sort of top 200 markets that you've talked about targeting traditionally. With that goal becoming increasingly in sight, do you have any different thoughts on expansion beyond the U.S., even if it's just in North America or the Americas more broadly?
Second, just maybe Mark, you could remind us the long-term financial goals, and that target model, if I remember correctly, was just simply for, the domestic footprint in mind that would not consider any sort of international expansion in it.
Let me start with this. I think we clearly are getting very close to that mental model that we provided earlier about there being about 200 markets in the U.S. that have over 200,000 people in those markets, serving about 80% of the U.S. population. We're approaching that pretty rapidly. I would say that the experience that we're having in terms of our ability to open markets and the way that those markets are responding to us and the way that we're seeing smaller markets respond is, if anything, kind of broadening our ambitions about what we think is possible. We'll take the remainder of this year to think through carefully what we think the right plan is. Directionally, if anything, our ambitions are being broadened there. As it relates to thinking about outside the U.S., I think we've got a long way to go here.
This quarter, we're on the order of 0.5%, less than 0.5%, of the used car market. We think there is just so much room to grow in our most mature market. We're still a relatively small part of car sales happening there. We still have relatively low awareness. We think we've got a lot of room to grow. I think we've got a machine that works really, really well. We think there's a lot of room to just continue to market that machine and tighten that machine up since we're only a six-year-old company and to see a lot of growth.
We also think that the business of buying cars from customers that has been effectively launched off of this infrastructure that we built is good evidence of the types of things that we can do here domestically as well to further monetize the transaction and service all the needs the customer has when they're buying or selling a car, and give them a better experience. I think our focus is going to definitely be there before it moves outside of the boundaries of the U.S. That is what our long-term model presumes as well.
Great. Thank you.
This concludes our question and answer session. I would like to turn the conference back over to management for any closing remarks.
All right, thanks, everyone, for joining the call. We really appreciate it. To everyone out there on team Carvana, thank you so much for everything you're doing. This was an absolutely incredible quarter. Please take a moment and take pride in what we've built. It's pretty incredible, but we've still got a lot of building left to do. Go get a good night's sleep, and let's come back tomorrow and keep on keeping on. Thank you, guys.
The conference has now concluded. Thank you for attending today's presentation, and you may now disconnect.