Michael Levin - VP, IR Ernest Garcia - Founder, President, CEO & Chairman Mark Jenkins - CFO

Ronald Josey - JMP Securities Derek Glynn - Consumer Edge Research Nathaniel Schindler - Bank of America Merrill Lynch Tania Anderson - William Blair & Company Seth Basham - Wedbush Securities Nicholas Jones - Citi Zachary Fadem - Wells Fargo Securities William Lowden - BMO Capital Markets Gary Prestopino - Barrington Research Associates Ryan Sigdahl - Craig-Hallum Capital Group

Good afternoon, and welcome to the Carvana Second Quarter 2018 Earnings Conference Call. [Operator Instructions]. Please note this event is being recorded. I would now like to turn the conference over to Mike Levin, Vice President of Investor Relations. Please go ahead.

Thank you, Austin. Good afternoon, ladies and gentlemen, and thank you for joining us on Carvana's second quarter 2018 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. And now, with that said, I'd like to turn the call over to Ernie Garcia. Ernie?

Thank you, Mike, and thank you, everyone, for joining the call. We had a great second quarter. We performed well against our guidance and more importantly, continued to make significant progress toward our mission of changing the way people buy cars. We exceeded the top end of our guidance ranges in both units and revenue and also comfortably beat our GPU and EBITDA margin targets. This was driven by stronger-than-anticipated top and mid-funnel demand as well as strong execution across the business. The experiences we deliver to our customers continue to drive high growth in market share across both markets and cohorts. Our oldest cohorts continue to show significant growth, and the other meaningful trends we have seen at the cohort market level remain unchanged. The progress we are making toward our midterm GPU goal of $3,000 is consistent and significant. In fact, in the five quarters since going public, we have made $1,000 of progress in this metric. This quarter, we also started to get visibility into 2 other meaningful GPU opportunities through a transaction to better monetize our finance receivables and very exciting progress in the business of buying cars from our customers. Mark will provide more detail on this in his comments, and the topics are also covered in our shareholder letter in greater detail. The finance transaction clearly demonstrates additional value in the finance receivables generated on our website. And our progress in buying cars directly from our customers is significant in several ways. It demonstrates that customers are willing to sell cars to us online, just like they're willing to buy them. It gives us another opportunity to serve our customers and deepen our relationships, and we provide them with the simplest and most fair experience available when selling a car. And it demonstrates the existence of a large additional GPU opportunity. Importantly, the progress also demonstrates our ability to execute against these opportunities. We believe both are very significant in the medium and longer term, but believe continued progress may be variable along the way, but it's still early days and we still have a lot of valuable learning to do before turning to optimization. Stepping back, with all this progress as well as the other GPU opportunities we have discussed in the past, we are feeling very confident about our path of $3,000 and beyond. The path to profitability is also crystallizing, with EBITDA margin improvement of 7.3% year-over-year. Narrowing our EBITDA margin loss is driven by rapidly growing sales volumes, increasing GPUs and cost efficiencies that we are finding despite supporting triple-digit growth and a doubling of markets, operating the business, significantly subscale, and still having all of the potential efficiency gains you would expect from a company that's only 5.5 years old. Our product pipeline is also stronger than it's ever been. We're beginning to roll out a site redesign that should be completed by the end of the year. This redesign will include many functional improvements that make it even simpler to find and buy a car on our website. It will also include many U.S. design and communication improvements that make it easier to understand our value proposition and every step of the process as well as an architecture that will make it easier for us to roll out additional functionality modularly and test those rollouts more quickly. In addition, we are deepening our use of data to personalize our customers' experiences, and several projects to improve the efficiency and responsiveness of our operations' teams are already in motion. We believe we offer the best customer experience in the market today, but also believe we can make it even better than it already is and are dedicated to consistently raising the bar. The operations side of the business - on the operations side of the business, we had an impressive Q2. We opened 9 markets in the quarter and crossed over a major milestone of now delivering the Carvana experience to half the U.S. population. We opened four vending machines, which further built the brand and - which are supposed to further build the brand and improve operational efficiency. As a testament to the brand power, we were featured in over 500 local news spots around the country in the quarter, in both markets where we opened the vending machines and in markets where we didn't. We also broke ground on our fifth inspection center near Indianapolis while continuing to aggressively build out our own logistics capabilities to connect those nine new markets and the Indianapolis inspection center into our broader logistics network and to add additional capacity in the constrained routes. In the quarter, we quickly mobilized our ops teams to alleviate many of the pinch points in our logistics network, which is a big part of the reason we performed so well. That said, our progress in logistics, paired with additional strong growth in top funnel demand, pushed pinch points in other parts of the business. We continue to view this as a high-quality problem, but it's a problem nonetheless. To better position ourselves to satisfy the demand we are seeing and even more importantly, to make sure we continue to deliver and improve upon our customer experiences under the strain of all this growth, we are elevating scalability as an even higher priority. The back half of the year has always been a time when we invest and prepare for the volume we anticipate coming in the first half of the following year. This year, that is especially true. Heading into 2019, we've begun to invest more to ensure we have sufficient operational capacity to handle meaningfully higher volumes. Those investments are taking many forms, but they're aimed at improving efficiency and flexibility as well as building and cushioning against our operating plan. The business is on a very exciting trajectory. We're delivering on our promise to provide customers with the best experience available in buying a car. This is resulting in continued triple-digit growth and the demand from our customers to be growing even faster than we are. The unit economics of individual transactions and of markets continue to progress nicely as does the visibility and the profitability, long-term unit economics and beyond. We continue to make strides in both technology and operations that separate us further from the competition and provide additional competitive modes. This all points to the direction of our primary business risk being execution. And our results show that our team is up to the challenge. Mark?

Thank you, Ernie, and thank you all for joining us today. Unless otherwise noted, all comparisons will be on a year-over-year basis. Q2 was the strongest quarter in our company's history, and we are excited about what this quarter means for our growth trajectory. Retail units sold totaled 22,570 in Q2, an increase of 111%. Total revenue was $475.3 million, an increase of 127%. Total gross profit in Q2 was $49 million, an increase of 206%. And total GPU in Q2 was $2,173, an increase of $672. We continue to see broad-based improvements, including a reduction in average days to sale to 66 days in Q2 from 105 days a year ago as well as gains in reconditioning, wholesale, financing and new products. EBITDA margin was negative 8.8% in Q2, an improvement of 360 basis points from the prior quarter and 730 basis points year-over-year. In addition to GPU gains, we are showing significant operating leverage while simultaneously expanding our geographic footprint. We opened nine markets in Q2 and have opened eight more so far this quarter, bringing our total to 73. We now expect to be near the top end of our range of 35 to 40 markets opened in 2018, bringing our end-of-year total near the top end of our 79 to 84 range. We also took several steps to increase our inspection and reconditioning capacity. In Q2, we acquired a facility near Indianapolis to service our fifth IRC. Similar to our other IRCs, this facility will have capacity to produce 50,000 units per year at full utilization, bringing the total annual capacity in our five IRCs to 250,000 units. We expect the Indy IRC to play a key role in our logistics network, allowing us to more efficiently service customers in the Midwest. We also, as expected, became the sole occupant of our inspection and reconditioning centers near Dallas and Philadelphia, which we previously shared with DriveTime. This transition allows us to expand our production in the Texas and Northeast regions to accommodate anticipated sales growth as we look toward the first half of 2019. As we ramp production, we expect this transition to have a small transitory negative impact on total GPU. This impact is reflected in our guidance. On August 7, we earned a fee of approximately $4 million for facilitating the refinancing of $236 million of Carvana-originated finance receivables that we had previously sold. The refinancing provides the purchaser with more efficient financing than at the time of their initial purchase, and we are sharing in the value created. The fee will flow through our other revenue line item in Q3 and increase our third quarter other GPU by $160 to $170. We believe this transaction demonstrates our ability to better monetize the finance receivables originated on our website. In terms of outlook, for the third quarter, we anticipate continued gains across our key financial metrics. We expect retail units sold of 23,000 to 25,000, an increase of 96% to 113%. We expect total GPU to be $2,100 to $2,350 as we march toward our midterm goal of $3,000. We expect EBITDA margin to be between negative 9.5% and 7.5% in Q3, an improvement from negative 15.9% a year ago. You should use approximately 145 million weighted average shares on a fully exchanged basis in Q3 and Q4. For the full year, we are raising our outlook to 91,500 units to 95,500 units and $1.85 billion to $1.95 billion in total revenue. We are also raising our GPU outlook for the full year to $2,000 to $2,200, based on the strong execution we are seeing across the transaction. As we look forward to the second half of 2018, we are excited about continued progress toward our financial goals. We intend to continue to progress in our ability to efficiently scale our business, with an eye towards preparing for a seasonal surge in the first half of 2019. Thank you for your attention. We'll now take questions.

[Operator Instructions]. And our first question today will come from Ron Josey with JMP Securities.

Two please. With days to sale outstanding and presales, that's the topic, so down at 66, clearly, an improvement from 1Q and certainly from the back half of last year. We're getting closer to that 45-day goal. So Ernie and Mark, you mentioned a few reasons why DSO improved. Just wondering if how the impact or the benefit from faster vehicle turn times that you mentioned due to less expensive cars helped. Or perhaps, better merchandising presale or coming-soon inventories helped on the site. And then, Ernie, you mentioned some improvements in the pinch points from last quarter, but I didn't quite catch what you said, how you fixed it, and then the newer pinch points that arose.

Thanks, Ron. So on your first question, I think retail GPU increased by $415 per unit year-over-year. That's comparing Q2 '18 to Q2 '17. Average days to sale over that same period declined by 39 days. And so that reduction in average days to sale does explain the majority of the gains that we've made in retail GPU over that period. I think we made some small gains in other areas. But I think the biggest driver, like I said, is the reduction in average days to sale. The way we achieved that, of course, was through our inventory strategy that we executed throughout 2017 and will continue to execute to a degree going forward. And that's basically taking advantage of the e-commerce model that we have. We have pooled inventory and can very quickly add markets and increase demand in those markets, the drop on that pooled inventory. And that gives us good control over average days to sale that we've leveraged over the past year and over time, we'll leverage going forward.

Yes, Ron, and this is Ernie. So I do - we made a lot of progress against those pinch points we called out in the last call. I think that target took many forms. We invested quite a bit in the logistics network to try to alleviate some of the routes we were constrained. We started to invest in building out flex teams that enable us to kind of push places where we have excess capacity over to places where we are constrained in logistics or with inside advocates or with outside advocates. So there's been a number of different initiatives underway to try to improve our ability to satisfy this demand. That said, we continue to see demand outstripping our ability to fulfill it. I think if we made it through Q2, we're very happy with those results. But we alleviated many pinch points and sort of created additional pinch points kind of further down the purchase funnel as we alleviated the initial pinch points. So I think that's a high-quality problem, for sure, but it definitely requires focus. And so we are now making a very concerted effort to make sure that we're preparing ourselves for what we expect to be a very high-growth Q1 and Q2 of 2019. We're growing at triple-digit rates, despite nearly being at a kind of $2 billion run rate of revenue. And I think that definitely creates some operational difficulties in the business, but I think we're doing a really good job addressing them.

Our next question comes from James Albertine with Consumer Edge Research.

This is Derek going on for Jamie. Just a couple of strategic questions. Based on what we can tell, since the majority of the inventory listed on the site is young, you used sort of late-model vehicles, those model years 2014 or younger, do you see an opportunity in the long run to expand the composition of the inventory into older vehicles, those that are four years or older? As you gain more scale, would it make economic sense to try and capture greater share in that segment of the used market?

Yes. So I mean, I would say we started kind of initially with the expectation that it would be easiest to sell the latest-model used cars online because they were closest to new and therefore customers buying on the site, as I'm seeing, would have the easiest time buying later-model used cars. I think if you look over the last several years, we continue to sort of spread out and move to more older cars, less expensive cars, also up into more expensive cars. I think our fundamental value proposition is delivering customers a great experience, a very fair price, a broad selection. And I don't think that that's specific to a particular segment of vehicles. So I think, over time, we are likely to continue to expand. We will expand at the rates that we see our customers demand in cars from different segments. We're constantly monitoring all the demand we're seeing from the website for customers, and then we're migrating in the directions that they're telling us they want to go. But we're seeing good signs. We don't see return rates at elevated levels for older cars or reviews that are lower in any material way for older cars. So we think there's a big opportunity there for us to move, I would say, throughout the spectrum. I don't think there's a single direction that they can do for us to move up as well.

Okay. Got it. And then just on demographics, have you seen any meaningful change over the last quarter or two in terms of the makeup of your customer base?

No. Nothing. It's been very stable.

And our next question comes from Nat Schindler with Bank of America Merrill Lynch.

A couple of quick questions. One, you described that you're currently servicing 52% of the population with 65 markets. Could we then assume that your target would be roughly 120 markets? Or does it get to the point where those - the incremental markets are just not worth the effort? And then I have some follow-ups.

Sure. So Nat, the framework we used to think about market expansion over time is the following. So there's about 200 or so markets in the U.S. that have 200,000 or more people in that market population. We think, over time, all those markets, or at least the vast majority of those markets, are a good fit for the Carvana model. I think our model has several nice features when it comes to serving markets in many different sizes. Because we do have a capital and headcount-light expansion model, we can very efficiently launch new markets with just an extension of our logistics network and a limited headcount in the market. We also - because we have centralized inventory, we're able to make that inventory available to all the customers in the market right from day one via our logistics network. I think in midsize and smaller markets, that's a big advantage because we're able to serve that market with a very large inventory, a much larger inventory than customers in those markets are used to seeing. And so I think as we look forward, we think there's a lot of room for us to expand markets over time. And the 200 markets with 200,000 or more people is - are the key stats that we think about in that framework.

Yes, Nat, I know you already know this, but I can't resist the opportunity for more top spin here. So I was saying, we're really excited about the opportunity to open more markets. We also believe that there's an even bigger opportunity to continue to catch up in market share in the markets we've already opened to Atlanta and for our most mature markets to continue to grow. We continue to see really exciting signs there. And so we think we kind of have those three major growth drivers, oldest markets continuing to grow, already launched markets catching up to older markets, and then we still have half the country to go with the rest over time.

Okay. And then just a follow-up from that. You mentioned that - the markets of 200,000, but you guys go - obviously, CarMax has to actually physically be close to those 200,000 people. You have 100-mile radius usually surrounding your markets. So when you opened in San Jose, California, you effectively have San Francisco. You don't need to open San Francisco, even though it has another 800,000 people just 40 miles away, or more, if you want to count the whole Bay Area, I guess. But is that how you would think of it? Or would you say, "Well, that's another market. I can open in San Francisco, and I can open in the East Bay. And that, this whole, the Bay Area, is 5 million people, so I could have many Carvana markets in that area."? Or do you think of it as, effectively, you already have that area with that one opening in San Jose?

Yes. So that's a good question. So the way we think about markets is we define markets as having the full Carvana experience. And what that means is you've got free local delivery with the branded Carvana hauler and a Carvana employed delivery advocate. You've also got our free - our full suite of local advertising to make sure that customers in that - any market that we open are gaining awareness and finding access to a Carvana at the top of the funnel. And those are the two things that we look toward when we're defining a market. In this particular case, we've opened San Jose. Some San Francisco and other customers may find us because we're in San Jose. But we haven't officially opened that market and because we're not doing the full Carvana experience in those markets yet.

Great. And then just a follow-up on something else you said about growing within your older markets. I think, in the last call, you gave us a stat that you were growing roughly 67% in the - your market cohorts that were older than a year old or something roughly. Do you have a follow-up to that? Is that number sitting in roughly constant? And can you characterize the growth relative to other markets that you're seeing in your oldest cohort, let's say, Atlanta and Nashville?

Yes. I would say we continue to see trends that are very similar. We continue to see very significant growth in our oldest markets. New markets are ramping similarly to how they've ramped in the past, faster than those in the market. So all those trends that we pointed to in the past generally are still holding up. So just to restate those: older markets are growing; the new markets now, they're growing faster than older markets; smaller markets tend to grow faster than larger markets; markets that are more approximate to preexisting markets tend to grow faster. All those trends, for the most part, remain very consistent.

Okay. And then just one further follow-up, so I can completely monopolize this call. You have - I noticed your ASP growth was roughly seven - it had turned around from actually a slow downtick to slightly up. Is that because of this move towards higher-end cars? I did notice that you have now Tesla on the drop-down, which makes sense if you're in San Jose.

Sure. So there are a couple of things going on there. First, ASP typically increases seasonally in the second quarter. So that was a component of the sequential increase. The rest of the change was driven by mix, as you point out. So we basically sold more inexpensive cars, relatively inexpensive cars, in Q1, which resulted in relatively more expensive cars on the site in Q2. We saw a strong demand for those relatively more expensive cars, which pushed ASP up. I do think, as we look forward, we expect ASP to moderate a little bit as the mix normalizes.

Is that something that we should see, that seasonally normally, because of tax refunds occurring, having probably a disproportionate effect on the lower side of the market?

So one thing that happens to our mix in Q2 typically is we'll start to have more one-year-old cars on the site from the previous model year. And so those are newer and typically lower mileage that can have - that can push up ASP. And that's the driver that we've seen over the past several years.

And the next question comes from Tania Anderson with William Blair.

Could you - I kind of missed. I didn't really understand what the other pinch points down the line were that were different from the first quarter. And then those dedicated flex teams, like I'm sorry if I didn't understand this, so what exactly, what functions are they doing out there to help you? And then on the - purchasing direct from your consumers, you hit 11% this quarter. Do you have a near-term goal for that? And do you advertise it? Are you planning to advertise it soon?

Yes. Sorry, I'll start with the pinch points thing again a little more clearly. So what I mean by that is just as we start to alleviate pinch points, what we generally see is higher conversion rates on the site. When you start to see those higher conversion rates on the site moving down through the funnel, it puts more pressure on other parts of the transaction. So customers may show up on the site. If they don't find the car they want, they may bounce off the site. If they do find a car they want, then they want to understand where they can have it delivered and they want to schedule delivery. So that can put pressure on the logistics network or out in the field for last-mile delivery to their door. Or they may have questions. They may want to call an advocate on the inside and talk to them. So just as we make progress in any area, we just tend to see higher conversion rates with any given level of demand, but then starts to put pressure elsewhere. And so we make that progress and then we see kind of additional pressure elsewhere. And again, I think that's a high-quality problem, but it's a problem that we have to get in front of. When I say dedicated flex teams, what I mean, for example, is we have a scene out in Phoenix now that is able to fly out to any given markets out in the field that are having constraints with last-mile delivery to customers' doors, and they can go kind of provide some relief there so that we can make sure that we can handle deliveries to those markets and we don't see delivery times extend out too much. That gives us flexibility because when we're trying to predict where demand is going to grow at these kind of various high rates, it's hard to get it right at every single market and then across every single kind of edge of the logistics network. And so we need to build in flexibility so when we start to see a demand show up, we're able to quickly move resources to handle it. And so I think that's a more concerted effort, that we've started to dedicate more resources to more recently. Firstly, from our customers, I think that's an extremely exciting opportunity. I think for most automotive retailers, that is one of the major differentiators that they will point to. There's not a lot of places in standard automotive resales that differentiate theirselves in meaningful ways. And I think acquiring cars from customers is one of those areas. And so it's an area that's spoken about a lot in the industry. I think prior to our progress here, we've always viewed our biggest differentiators as being kind of more in the transaction itself and less in the acquisition of cars. And so this has just more recently become a focus for us. I think we're very confident that the experience that we deliver to customers is very good. Customers can go on our website. In a couple of minutes, they can get a value for their car. They can schedule a time for us to come pick it up. And we pick it up and give them a check and drive away if they're just selling a car to us. Or if they're also buying a car from us, we deliver their new car and pick up the old car and drive away. So it's a very simple experience. And I think we've known that customers really have enjoyed the experience for a while, but we just haven't put as much operational attention on that area of the business until more recently. I think the early progress is really exciting and significant and makes us believe there's a lot of opportunity here, but we're also just kind of in the very early innings. And so I think in both the percentage of cars that we're buying from customers and in the margin that we're getting from those - the incremental margin that we're getting from those transactions, which can either flow through wholesale, if we sell it at an auction, or can flow through retail GPU, if we sell directly to customers, we just think that there's a lot of variability that could show up there over time. And so we don't want to promise anything in the immediate term. But we think it's a huge opportunity, that in the longer-term, we expect to make a lot of progress.

And our next question comes from Seth Basham with Wedbush Securities.

My first question is looking at the flow-through of the top line to gross - to EBITDA. And we didn't see that much flow-through of the top line gross profit strength. Can you point - pinpoint where the biggest shortfalls were because of that?

Sure. So we talked in the shareholder letter about investing for scalability. I think you can think of that as starting in Q2. And then those investments in scalability took a variety of forms, and we expect them actually to continue to take a variety of forms. So the first form it takes is investments in technology and process improvements that we believe will increase our efficiency and long-term scalability. That's things like investing in management to focus on operation, strategy and process improvement as well as increasing product engineering and data science to also improve and automate processes. We've also started investing in Q2 and plan to continue investing in the rest of the year in investments that increase our flexibility. So Ernie talked a little bit about flex teams. We've also, in Q2, spent on third-party logistics that we used as spillover to help relieve pinch points. We expect to do that to some degree in the back half as well, while we build out capacity in our logistics network, particularly on constrained routes. A third category that I hit is we do - we did begin to and expect to, for the rest of the year, make additional investments in staffing to help meet the demand that we're seeing at the top of the funnel, that staffing and last-mile delivery network, first-party logistics as well as customer service. And all of that is geared toward meeting the demand that we're seeing at the top of the funnel, which is significant, and also importantly, making sure we're really fully scaled and capable of meeting the expected demand in the first half of 2019. If you've seen the seasonal cadence of our business, we typically have experienced the largest sequential gains in unit sales and showing the most leverage in the first half of the year. And so we want to make sure we're fully prepared for that.

That's helpful color. The one item you didn't mention was advertising. Advertising per unit increased year-over-year for the first time, at least for the past 6 quarters. Is this simply because of the increasing mix of immature markets? Or is there something else going on here?

Yes. So that is all mix. I think Ernie alluded to this, but just to hit the point again, we're seeing very positive trends within cohorts in advertising expense per unit, both year-over-year and more recently. And so with positive trends in all of the cohorts, the company-level metric is primarily going to be driven by mix of newer versus older markets.

Got it. And my last question is looking at the gross profit per unit guidance for full year 2018 that you increased by $25. That's less than the value from the incremental monetizing of financing receivables that you're planning in the third quarter. I assume that was in the guidance before. Can you please help me understand the moving pieces there?

Yes. I would say it was, to some degree, in the guidance before, and I think it's hard to precisely walk through exactly the guidance because there's so many different buckets for GPU improvement. So we definitely model in improvement. As I've said, we have improved by $1,000 over the last 5 quarters. And this is one of the buckets that we've known for a long time was an opportunity for us. We just had to put our focus on it and go get it. So I think that largely was in the guidance before.

And our next question comes from Mark May with Citi.

This is Nick Jones on for Mark May. On, first, I think, cars from customers, I guess to expand on that. Are there any advantages Carvana has over its competitors that make Carvana a more compelling place to sell a car if you weren't going to do a trade-in? And then secondly, is there any new color you can provide on repeat customers in your early markets?

So I think the first thing I would say there is that I think our logistics network and kind of our offering of going and picking up the car from you, I think, are big advantages versus what most competitors out there are offering. I think the ability to get a value from your computer, where you're not having to take pictures and calls and talk to someone on the phone and answer a bunch of questions and then go into the dealership and have it physically inspected to get a value, I think there's a lot of value in that. I think we're leveraging a lot of data to enable that. So I think those are differentiators that make the experience much, much simpler. I also think that that's one of the areas where the positive feedback of brand can show up. We've talked for a long time about the leaders there. Three kind of major areas of network effects in the business: one is brand; one is marketing; and then one is inventory size. But I think this is the place where brand is showing up. If we build a big brand where customers understand they get a great experience and then we add a new offering that makes sense that fits alongside our existing offering, we believe that we'll be able to make progress much more quickly. So I think those would be the major differentiators that we would point to, and we think they're clearly showing up in the results. On the new metrics on repeat customers, I would say we've only moved forward in time kind of 1 quarter. And again, just to remind you on the numbers there, 5 years ago, we sold about 100 cars. And so to give concrete data on repeat customers right now I just think would be really, really early because it's so thin. But as I said, we look at a lot of metrics that we think are rough proxies for repeat customers, and we feel very good about that data. One of those metrics is just the quality of reviews that we get on our website, and it continues to be very strong. So I think we're very excited about what that could be in the future, and as soon as we got any of it stable, we'll make sure we share that with you.

The next question comes from Zach Fadem with Wells Fargo.

On your out-of-market sales, is 2018 to date still tracking at that 20% level? And any change to the impact of national advertising? And then also with respect to economics, can you talk about the margin profile versus in-market and how that's been trending?

Sure. So in terms of the metric, yes, we haven't seen any meaningful changes there. I think there's certain regions that are outside of our current logistics network where there's increase in national advertising. We've seen small pickups in demand outside of the network, but nothing in particular, sizable that I recall out there. In terms of the margin profile, you can broadly think of that as being similar as, of course, order approximation.

Okay. And then just a bigger picture on your newer market trends, could you talk a little bit about the maturation curve for your 2017, '18 cohorts or maybe 2016, '17 cohorts and how that's compared to some of your earlier markets in terms of sales uptick and market share? And then, as we move forward, is there anything you've learned along the way that you can improve the customer response and maturation curve in future markets?

I would just say that the trends that we've seen continue to be very consistent, and it's the same kind of drivers that we've seen in the past. I think one thing that's interesting to note is we definitely see improving unit economics at the newer markets because they start with much higher GPU. So for example, our 2015 and older cohorts are all contribution margin-positive today. And the more recent cohorts are getting there much, much faster because they start with higher GPU and they also start with generally lower CAC because we're leveraging national marketing. But broadly speaking, I would say, outside of kind of the economics of the cohorts, the underlying unit ramps are following similar trends to what we've seen in the past.

Okay. And then quickly on macro, any impact, headwinds from rising gas prices, rising interest rates are you seeing anything in your business today?

I think, in general, I think we'd end up by saying we don't think we're immune to macroeconomic forces. So if there are forces that are impacting the broader industry, we're likely to be impacted in a similar way. I also think that it just does not get a lot of our attention because we're growing really, really quickly and we feel like we've got control over forces that move the business much more strongly than these forces that we don't have control over. So I would just say, whatever your expectations are for how those things are going to impact other automotive retailers, I would just build that in your expectations for us, and we'll be focused on all the things that we can control.

Our next question comes from Dan Salmon with BMO Capital Markets.

This is William on for Dan. I have 2. So first, can you go into more detail about the impact your acquisitions of Car360 and Carlypso are having across your business? And then second, on your floor plan facility, you're pretty much maxed out there. What are your plans for increasing that going forward?

Yes. So as I said, Carlypso is serving us in 2 ways right now. I think one is - one of the founders is focused on the business of purchasing cars from public and has obviously done a great job, making a lot of progress there very quickly, and we're very excited about that. And so I think that's great news. The other area is a little bit less visible, but I think every bit as important and in many ways, powers progress we're making in many areas of the business is the data that we got from Carlypso that enabled us to understand the specifics of a vehicle much more granularly without having to physically see it. And we're making a lot of steady progress there. I would say we're just starting to see some of the benefits of the data rolling into many different parts of the business, be it merchandising, the way that we're purchasing vehicles, the way we're valuing vehicles, the way that we price vehicles on our site. So that's all kind of starting to flow through. Car360, there's nothing you can see yet, but we're extremely excited by the progress that kind of we're looking at on our side, and I don't think it will be long until we have something fun to share.

Yes. And then, William, on the floor plan, so we've upsized the floor plan several times historically and do intend to upsize it again as we ramp inventory in the back half of the year in preparation for the first half of '19.

The next question comes from Gary Prestopino with Barrington Research.

A couple of questions. Ernie, I didn't catch it. When is your IRC going to be completed in Indianapolis?

So we expect to be producing cars out of there by the end of the year.

Okay. And where were you servicing the Midwest prior? What IRC is servicing the Midwest?

So all IRCs were serving the Midwest prior to the launch of Indy and will continue to do so. Prior to the launch of Indy, the closest IRCs, depending on exactly where you are in the Midwest, could either be Delanco, Blue Mound or to a lesser extent, Winder. And by that, I mean, Philadelphia, Dallas and to a lesser extent, Atlanta.

Yes, and I would say I think that there's something kind of interesting that emerges from that thought exercise. Obviously, putting down that inspection center in Indianapolis, one, it increases our capacity, which is great, and helps us stay in front of the growth that we anticipate coming in the future. But I think, two, it also reduces the delivery times and delivery miles to customers because we now have an inspection center that is much closer to customers. Reducing delivery times generally results in higher conversion. Reducing miles generally results in lower cost. So we think that that's something that will continue to happen as we add additional inspection centers around the country over time. So I think that's another kind of nice harder-to-see benefit of opening an additional IRC.

And could you maybe just comment on what you're seeing as far as supply particularly coming out of the auction market? And then are you able to use some of the newer technologies that some of the auction companies are offering their clients to buy cars, and not at the physical auction location, but maybe on a dealer-to-dealer basis?

Sure. So no major patterns of the auctions to call out. I think we're seeing a great selection of cars for our customers and are purchasing accordingly at the auctions. In terms of technology, obviously, we're - we've invested a significant amount in building out our proprietary purchasing algorithms which take into consideration a wide range of information and using sophisticated analytics to decide what we're bidding on all the cars that are deployed through those auction lanes. In terms of looking for additional sources of supply, and that's something that we're always interested in looking to test, to the extent, there's new products or new channels, but nothing specific to call out there.

And just to remind you, Gary, we do all of our purchasing from the home office in Phoenix. We don't send anybody to the auctions physically as it is today.

And our last question today will come from Steve Dyer with Craig-Hallum.

Ryan Sigdahl on for Steve. So retail GPU was up about $300 sequentially from Q1, although days to sale will only decrease by 4, from 70 to 66. Any additional color there would be helpful.

Sure. So one driver of the sequential increase and position of the days to sale declining was our Q2 inventory faced lower average depreciation rates than our Q1 inventory. We talked a little bit about the hurricane effects on Q1, for example, where wholesale prices were particularly high in Q4 of last year and then depreciated late in Q4 and through the first part of Q1, leading to lower-than-otherwise gross profit in Q1. So that would be the second - lower average depreciation rates in Q2 than Q1 would be the second driver I would point to.

Okay. And then lots of discussion about new linchpins - pinch points. And just looking at the GPU guide, is it correct to assume that retail GPU will be down quarter-over-quarter from Q2 to Q3?

So we don't give specific guidance on components of GPU looking forward. One point that we did call out is we have taken over DriveTime facilities. And as we ramp, we expect transitory negative impacts on retail GPU from taking over those facilities. But we do expect that to dissipate as we ramp production. And then, I think, generally speaking, we're very excited about the progress that we've made in GPU across all parts of the transaction. Obviously, with $2,173 in 2Q and then strong guidance for Q3, we're feeling very good about our path towards the $3,000 midterm goal.

And this will conclude our question-and-answer session. I would like to turn the conference back over to management for any closing remarks.

Thanks, everyone, for joining the call. We're extremely pleased with the quarter and with the broader trajectory of the business and the enormous opportunity we have in front of us. Most importantly, thanks to the Carvana team. We had a great quarter. We continue to build the foundation for a very big future. A special thanks to the ops teams for everything you did this quarter. You had a lot coming at you and did a tremendous job. Thank you.

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