Earnings Call: Q1 2017

Jun 6, 2017

Good afternoon, and welcome to Carvana's First Quarter 2017 Earnings Conference Call. All participants will be in listen only mode. After today's presentation, there will be an opportunity to ask questions. Please note this event is being recorded. I would now like to turn the conference over to Alex Wellens, Investor Relations. Please go ahead. Thank you, Austin. Good afternoon, ladies and gentlemen. Thank you for joining us on Carvana's Q1 2017 financial results conference call. Please note that this call will be simultaneously webcast on the Investor Relations section of the company's corporate website at www.investors. Carvana. Com. Joining me on the call today are Ernie Garcia, Chief Executive Officer and Mark Jenkins, Chief Financial Officer. Before we start, I'd 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 may cause actual results to differ from forward looking statements can be found in the Risk Factors section of Carvana's prospectus related to its initial public offering filed pursuant to Rule 424 under the Securities Act with the SEC on April 28, 2017. The forward looking statements and risks in this call are based on current expectations as of today and Carvana assumes a no obligation to update or revise them whether as a result of new developments or otherwise. With that said, I'll turn the call over to Carvana's CEO, Ernie Garcia. Ernie? Thanks, Alex, and welcome everyone to our first earnings call. We're very excited to share strong Q1 results with you today. Hopefully, you've all had a chance read our shareholder letter, which is available on our Investor Relations site. We've decided to issue a shareholder letter prior to our calls in an effort to give you all the information you need to ensure you have time to digest it prior to the call, so we can focus more time on answering your questions. Prior to jumping into questions, I'd like to briefly frame up the way we think about the business and our priorities and then Mark will give a high level overview of our financial results. Carvana's mission is to change the way people buy cars. Everything we do comes from the belief that the experience of buying a car can be improved upon dramatically. We've built our e commerce offering to give customers what they want, lower prices, wider selection and a simpler no hassle experience. Our core belief is that consistently poor customer experiences in automotive retail are the direct result of high variable costs and a lack of differentiation between traditional dealerships. Our insight when building Carvana was that to truly address the customer experience, we had to change the economics of the business and culturally orient ourselves toward taking pride in providing exceptional customer experiences. Our business model, operations, processes, people and company culture are all purpose built to do this one thing, give customers great experiences when buying a car online. Our rapid growth shows that just as in other sectors, consumers are ready to buy cars online. We've made the process so simple that our customers can buy a car from Carvana on their phone or computer in less than 15 minutes. We deliver the car to their door or they can pick it up at one of our proprietary vending machines. And all that comes with the peace of mind of a 7 day return policy. We believe the evidence that our offering is resonating with consumers is clear and getting clearer every quarter. In fact, in the Q1, Atlanta became the first of our 23 markets to fulfill more than 1,000 sales in a month. To achieve this, we have made significant investments in technologies that reduce variable costs. This created a cost structure that looks very different from anything else in automotive retail. We have low variable costs and high fixed costs relative to dealerships. Low variable costs mean that we can sell cars at lower prices to customers. They also mean that we can invest in delivering high quality experiences to further separate ourselves from the status quo. Those high fixed costs create significant leverage in the business. By continuing to deliver on the best customer experience available when buying a car, we will continue to grow revenues. Given those fixed costs, that growth mechanically drives improvement through the rest of the income statement. Our mission and cost structure are aligned and they jointly drive our financial priorities for the business. Objective number 1 is growth. Growth is our first priority because in addition to being valuable in and of itself, it is also the primary driver of our second and third objectives. Growth has 2 primary drivers, growth in market share within markets and growth in the number of markets we serve. In the quarter, we grew revenues 118% year over year, which resulted from solid progress in both of those underlying drivers. We opened several new markets and our guidance calls for significant expansion of our footprint through the remainder of this year. Objective number 2 is growth in total gross profit per unit across all parts of the transaction. This is something that we got a lot of attention or that got a lot of attention on the roadshow, so I want to give you a quick outline of our plans here. The single largest driver of total GPU growth in our plan is growth in retail sales. As we sell more cars, we can turn cars faster with the same sized inventory. Because cars are depreciating assets, this will result in higher revenue, even when selling cars to customers at the same discount to market prices. Secondly, increase in unit sales reduced cost of goods sold. We've made significant investments in infrastructure to support our reconditioning and logistics efforts and higher utilization of that infrastructure reduces those costs on a per unit basis. These growth effects are responsible for about half of our walk from where GPU is today to our mid term goal of getting to $3,000 per unit. To date, we have grown total GPU significantly. This growth has primarily been driven by product and pricing enhancements and has only benefited from the growth effects just discussed to a minimal degree. This has been because we have historically made the strategic choice to grow inventory in lockstep with sales to give our customers a continually improving selection of cars to choose from. Over the last 6 months, we have determined that we now have significant coverage of the cars our customers are demanding. This means that growth in inventory is now less valuable to Carvana than it historically has been. A result, we are now beginning to leverage our growth to drive the average days to sale of inventory down, which will give us an incremental and powerful driver of total GPU that our prior growth has not benefited significantly from. We can add demand to that fixed supply in 3 ways open new markets, allow younger markets to catch up in market share to more mature markets and to continue to grow our more mature markets. Growing inventory more slowly than the sum of those three effects directly results in lower average days to sale. All that said, I want to make sure we are clear in setting expectations for the timing of this effect going forward. So I want to provide a bit more detail how this is likely to unfold. Given the delay in tax returns this year, along with our buildup of inventory in the Q4 in anticipation of that seasonal bump, we saw average days to sale increase a bit in the Q1 and expect it to increase somewhat more in the Q2. By the end of Q2, we expect to have worked through that aged inventory and expect average days to sale to start to decrease and provide a tailwind in the second half of the year. Our guidance for the 2nd quarter has this expectation of increasing days to sale built in. We view this as good news because it provides a clear path to total GPU improvement from here. Objective number 3 is to demonstrate operating leverage. The primary driver for this objective is simply to grow unit sales to spread our technology and fixed cost expenses across more units. This is another area where we have made significant progress historically. That progress slowed but did not stop in 2016 when we invested significantly in building out our corporate and product infrastructure, which was partially driven by preparations for being a public company. Looking forward, we believe we will be able to reaccelerate our progress in this metric given our expectation for growth in units and total GPU, as well as our expectation that we will slow investment in our corporate infrastructure. As you see in our shareholder letter, we are making meaningful progress in EBITDA margin. Lastly, I want to make sure I touch on the size of the opportunity we are tackling here. The market is enormous. The car market in the U. S. Is roughly a $1,000,000,000,000 market with used making up the majority of the sales. Consumers are responding. People are clearly expressing their desire to buy cars in what we call the whole new way. Hitting the milestone of 1,000 sales in Atlanta is 1,000 sales in a month in Atlanta is meaningful in any context, but especially so when considering the fact that we ended the quarter with 23 markets and anticipate opening at least 16 total this year because of our capital light rollout plan. The path to profitability is clear. The strength of the demand we are seeing within market and our ability to open new ones provides most of the fuel necessary to enable our total GPU lock. The brand we have built is powerful. Our website and brand assets like the vending machine make a strong impression and tell a clear story to our customers. We are executing. We are a 4 year old company and have achieved this scale in this timeframe because we are comprised of a group of people that believe in what we are doing and that pour themselves into delivering incredible customer experiences every single day. I couldn't be prouder to call myself a part of that group. We are well positioned. My view isn't that customers are just now becoming ready to buy cars online. My view is that customers have been ready for a while, but the right product hasn't been available to them and that demand has remained latent. I believe this because building an easy to use e commerce platform that lives up to the expectations of customers is very hard and requires expertise in many different areas. If it is hard the first time, it will also be hard to replicate. With that, I will hand it over to Mark to give an overview of our financial performance. Mark? Thank you, Ernie, and thank you all for joining us today. Full details on our Q1 financial results are available in our shareholder letter. We've put a lot of financial detail into the letter, so rather than repeat many of the numbers, I'll keep my comments brief and highlight a few aspects of our financial model and momentum. Unless otherwise noted, all comparisons are on a year over year basis. As the CFO of Carvana, I am very excited about our financial which combines rapid revenue growth, increasing gross margins and improving operating efficiencies. This yields an attractive long term model that we believe will result in significant profitability in the long run. We view retail units sold as the single most important driver of our financial model. Total first quarter retail units sold were 8,334, an increase of 120%, tracking closely with total revenue, which grew 118 percent to $159,100,000 While we are extremely pleased with our growth in the quarter, we believe that our growth would have been even greater were it not for the delay in IRS tax refunds, which affected consumer spending patterns. In the Q2, we see retail unit growth accelerating to 130% to 141%, further driving our revenue growth. As a management team, we are focused on improving gross profit across all parts of the vehicle transaction. Total gross profit in Q1 was $9,700,000 an increase of 146%. We improved total gross profit per unit by $123 to a total of $11.69 in the first quarter. We see this increasing another $200 to $2.50 sequentially in the 2nd quarter. We are also committed to improving EBITDA margin. EBITDA margin was negative 21.6 percent in Q1, a significant improvement from the prior quarter. We see EBITDA margin improving sequentially to negative 18% to 18.5% in the 2nd quarter. In addition to higher revenue and gross profit, we are also leveraging our cost structure and driving towards profitability. We are very pleased with our Q1 results and our Q2 guidance provides you an outlook into our first half financials. As we move into the second half of twenty seventeen, you should expect that we will continue to grow retail unit sales, open new markets, drive GPU and improve EBITDA margin. You should also expect that OpEx in the second half of 2017 will increase in dollar terms as we expand, but will decrease as a percent of revenue. OpEx includes anticipated stock based compensation of approximately $3,000,000 in Q2 $3,000,000 in total for the second half of the year. We expect very modest increases in share count in the second half. You should use approximately 137,000,000 shares on a fully exchanged basis for Q3 and Q4 after giving consideration to the IPO. In addition to GAAP net loss per share, we plan to report a non GAAP adjusted net loss per share that assumes all LLC units are exchanged to shares at the beginning of each reporting period and that shares issued in connection with the IPO were outstanding at the beginning of the period. We believe this adjusted net loss per share facilitates comparability across periods making it easier to track our progress. In terms of our full year outlook, we expect significant growth in retail units sold, totaling 44,000 to 46,000 for the year, up from 18,761 in 2016. We expect 16 to 18 new market openings, bringing our end of year total to 37 to 39, as we further penetrate existing regions and expand into the Southwest and West. We are executing against our key priorities, growing our market presence and leveraging our investments to realize efficiencies across our business. We are very excited about the opportunity in front of us and the strength of our financial model. Thank you for your attention and I think we're ready for questions. We will now begin the question and answer session. Our first question comes from David Lim with Wells Fargo. Please go ahead. Hi, good afternoon everyone. The question I have is on guidance. Ernie and Mark, can you guys give some level of detail on maybe what the swing factors could be where maybe those numbers that you've outlined could possibly get better or and maybe some of the assumptions that are underpinning that, maybe some color on what you're expecting for maybe market share in total for you guys? Yes, David. This is Ernie. I'll take first one at that. So, I think in general, the easiest way to think about what's underlying our guidance is to look at the market share curves that we put in our S1, and just kind of look at how each historically has grown in market share and then overlay the additional market openings on top of that. That's enough to kind of give you units and for the most part that is the set of assumptions that we are making to forecast the rest of this year. I think as it relates to total GPU, it's probably the next most important driver. We're basically looking at the impact that that growth will have to reducing turn time to increase total GPU through the rest of the year. And that's the primary driver there. So I think that gives you the underlying model. Got you. And then, how should we think about when we look at your Q1 results and the new markets that you've opened, how is that trending relative to maybe some of your more established markets such as Atlanta, such as Nashville? And can you give us some color on any kind of inflection where with the vending machines being open in Dallas and Houston, are you seeing like a little bit more of steeper curve up once that's open? Any color there would be appreciated. Sure. So, couple of questions in there. First of all, how are new markets responding relative to older markets? I think throughout the company's history, we've seen a slow progression where newer markets are generally responding better than older markets. I think that progression has continued and that continues to be what we see. I think as it relates to vending machines, we've now opened 3 additional vending machines aside from our first vending machine in Nashville and we've seen very similar results in those three markets. Those results are significant and more than enough to make the incremental investment of a vending machine very, very powerful inside of these markets. And as a result, we anticipate continuing to open these vending machines in all these markets as we move into them. I think there is another effect that's worth noting, but it's early to really rely on this. I think we're seeing that markets that are in closer proximity to markets we have previously launched are generally ramping faster than new markets that are not in close proximity to markets that we have previously launched. And we believe that that is a result of more seated awareness in the market prior to us arriving there. Something that will impact that going forward is we just started to test advertising on some cable channels over the last quarter. And we believe that that is now cost effective and will continue to get more cost effective through the remainder of the year. A byproduct of that aside from reducing incremental marketing costs as we go into new markets is that it also seeds awareness in markets that we aren't yet in, which could cause new markets to ramp more like the markets that are closer to pre existing markets. So that would be how I'd respond to that question. And thank you. My final question is SG and A per unit retail, it looks like there was some movement in that number this quarter versus last year. I was wondering if you could give some details. It like the number may have gone up on a year over year basis by just a shade. So I mean is there any kind of additional detail that you could provide relative to that? Sure, David. This is Mark. I can provide some color on that. I think the big transition that we went through between last year and this year is the transition associated with growing from $130,000,000 in revenue up to $365,000,000 in revenue and becoming a or preparing to become a public company. So 2016 was a big investment year for us. We built out our logistics network for the first time. We added significant amount of management to help us grow and really build this thing out. And we also added a lot of infrastructure and corporate headcount associated with preparing to be a public company. We saw that take the form of increased SG and A per unit in 2016, in particular in the second half of twenty sixteen as we really made our push to prepare to be public. We expect now going forward to really be able to leverage those investments and we expect to see declining SG and A per unit going forward over time. Yes, David, I would just jump in to add. I think if you look at the trend in our total gross profit per unit, you see that that's positive. If you look at the trend in our EBITDA margin, you see it slow down last year as we were making all those investments that Mark alluded to. And then you see it in Q1 and then in our guidance for Q2 start to reaccelerate, which gives you a sense of what we expect going forward. Great. Thank you so much. Our next question is from Nat Schindler with Bank of America Merrill Lynch. Please go ahead. Yes. Hi, Ernie. Hi, Mark. Congratulations on your first public comp your first public quarter. Your GPU, you certainly want a great progress in the GPU per retail unit in this quarter and your guidance calls for another, what, 250 sequential. And I'm assuming most of that is actually in the retail units as opposed retail GPU as opposed to in the financing and other parts of the business. So if that's the case, if it isn't on the timing that you're selling the cars, as you said, that it will actually pick up again. Where are you getting the leverage on to get that GPU? And additionally, on a longer term basis, you've had a lot of volatility in this number. Can you be confident in improving kind of sequential improvements largely from here on out? Yes. So Dave or Nat, I'll hit the first part of that question. This is Mark. So I think when you look at our Q2 guidance on total GPU relative to the $11.69 that we printed in Q1, those gains are across all parts of the transaction. We don't plan to give a detailed breakdown of what we expect on each element of the transaction, but I can tell you that we do expect that gain to be broad based across the different parts of the transaction, including vehicle and the other areas. I think as Ernie alluded to, the gains that we anticipate on vehicle will be in spite of slightly rising average days of sale. That reflects kind of efficiencies on the cost side of the equation as well as optimizations in purchasing and pricing. And then we also, like I said, expect to see some gains going into Q2 across the other parts of the transaction. Yes. And then I'll just jump in and add a little bit. I think the way I like to think about our GPU walk is to break it into those 2 buckets. There are the effects that are driven by scale and then there are the effects that are driven by product improvements. I think the effects that are driven by product improvements are generally harder and are chunkier and that for the most part is what has driven our growth historically. The effects that are driven by scale, once you choose to stop growing inventory as quickly and kind of benefit from that scale, they tend or we expect them to be much more consistent and straightforward because it's simply the effect of holding inventory flat and then selling more cars. And so, we believe that our GPU walk will be less volatile from here forward than it has been in the past. The last thing I would note is there were 2 one time effects in Q4 that I think when you look at our historical volatility in GPU, Q4 is a really big contributor to that volatility. Those two effects related to transitioning our finance deal and then related to offloading a bunch of recalls in the Q4. So we don't expect those 2 impacts to persist. Okay. And going to that timing element, how long you expect to be keeping cars and you're building that getting that scale advantage now, Your marketing basically says 7,500 plus cars available at any given time. If you doubled your sales, you'd obviously have and keep that same roughly, you'd have your amount of time you're holding your cars on average. Is that reasonable? Do you think 7,500 cars at any given time is reasonable to handle even double the business? So, yes, the way I would think about that is I think we still believe that there are gains in conversion rates to be had by growing inventory from here. We believe that those gains are less pronounced than they were when our inventory was smaller. And so, the balance of growing inventory and growing conversion versus recognizing higher gross margins by holding a smaller inventory and turning cars faster has shifted to where we now have more of a focus on driving down average days to sale and driving up GPU. So you can expect to see that going forward. We do expect also to be growing inventory over that timeframe over the next several years. We don't expect it to be completely flat. And then in terms of where that could go, I think if you look at other retail models across the board, you see kind of 30 to 60 days is probably the average window of average days to sale that you see for other groups out there. I think when you look at our business model, there's reason to believe we should be able to outperform that because we're holding a fixed supply group and we're making that available to many, many different buckets of demand. So, I think our medium term goal is to get in line with traditional retailers. And I think from there, there's certainly upside. And the way the math works is exactly as you outlined it. If we can sell twice many cars with the same inventory, roughly speaking, you have your turn times. Well, thank you guys. Good job on the Q1. Thanks. The next question is from Mark Kelley with Citigroup. Please go ahead. Hey guys, thanks for taking the question. First, you kind of hinted at it, but on the marketing spend, I know you're testing a national campaign. Are we at the right spot in terms of markets at this point where it's more efficient to roll that out? And then have you baked in any benefit from a national campaign into your full year guidance, since you'll be going in a bit warmer than what you've experienced in the past? And the second, just on the GPU side, can you talk a little bit more about the IRC utilization and when we might start to see you guys run 2 lines instead of 1 and kind of how that plays out? Thank you. Great. So I'll start this is Ernie. I'll start with marketing spend. So I think we are just now arriving at the place where some of the least expensive cable channels are efficient to buy nationally as opposed to buy in each of our markets and add that all up. And so we started to test that. The way that kind of pricing works across those channels is as we get larger and larger, more and more of those cable channels will start to make sense. And then as you get significantly larger from there, broadcast channels also start to make sense to turn to national. Our goal is to basically march through that as it makes economic sense and continue to add more and more national advertising as we scale from here. So, I think you should think about national advertising as being something that is sort of continuously rolled out over the next 12 months to 18 months, starting right now. As far as the impacts, that that will result in, I think we are hopeful and there is evidence as it relates to nearby markets performing better than markets that are not near pre existing markets that this could cause new markets to ramp faster than they historically have. That's not something that we're banking on or relying on. It's something that we will study and evaluate through the rest of this year. And then as we understand it better, it may or may not impact our guidance for our next fiscal year. Yes. And then Mark, I'll hit the point on IRC utilization. So our 3 current IRCs that we have up and running in Atlanta, Dallas and Philadelphia have collective capacity to produce about 150,000 cars per year. That's about 37,500 cars per quarter. If you compare that to the 8,334 units that we reported selling in Q1 of this year. That gives you a sense of current utilization levels. That puts you between kind of around the 20% to 25% utilization mark. All helpful. Thanks guys. I appreciate it. Thank you. Our next question is from Mike Levin with Deutsche Bank. Please go ahead. Good morning or excuse me, good evening guys. Congrats on the first call. I was hoping you could kind of dig into what are the pieces of days to sale as they kind of stand today? And how does that progress Q3, Q4 and into 2018? And what are the kind of the pieces that are on the reconditioning and prep side and what are the pieces on the days to sale and kind of market side? Yes, this is Ernie. So I'll take that. So days to sale, I think you can think about is broken up into those 2 components, debatably 3. So the first component is getting the car to the reconditioning center to get it refurbished. The second component is moving it through the reconditioning center and photographing it and then pushing it up to our website. And then the last component is when that car is sitting up on the website. And that last component is what is kind of driving the drive down as we put more and more customers through the same fixed supply. That's what causes turn times to increase. Roughly speaking, we're probably turning cars once they're up on the website in about 65 to 75 days, give or take. And that gives you a sense of what the other two components are, and that's where we think we're going to have our gains. Got it. So the kind of 20 to 30 days on the recon side, that is not expected to come down in the near term? We expect that to come down. So, I would break that into transport and then reconditioning. I think it's useful to break into those 2 steps. Transport, I think is something that we almost think of as not really being a cost in terms of time, because the way that we bid on cars is we're thoughtful about how much it's going to cost to cost of carrying that car during that extra transport period is already taken into account. Once the car arrives at an inspection center and is prepared for sale, that is definitely a very important time and we do expect to have gains there. I don't think that's a particularly large driver. We are already reasonably efficient and we do expect that to come down, but the gains that we will have there are small relative to the gains that we will get from just pushing more customers onto the website across the same fixed supply. Got it. Okay. And what in maybe just in Nashville at this point, in that market, what's kind of the mix of delivery versus pickup that you're seeing? And is it kind of trending in the same way with the latest vending machine openings? And have you kind of assumed those savings from pickups in your guidance kind of moving forward? Yes. So I'm going to answer that very high level. I think in the future, this probably isn't something that we'll continue to guide on, but I'll give you a baseline. So historically, we've seen about half of our customers choose to pick up when we have a vending machine option available and about half of them choose to have the car delivered to them. The trends in that have been relatively stable. We certainly believe that we have the ability to impact that based on our presentation of those options on our website. And we take advantage of that when it makes sense. So I think we expect that to continue to be very similar. The new vending machines that we've launched, we have not seen anything significantly different. We don't have reason to believe that it will be significantly different. Great. And maybe just kind of very high level here. During the roadshow, you'd kind of guided to 7% to 11.5% EBITDA margins and CarMax probably your closest competitor is around 7.8% last year. And if we kind of look at what you're thinking about in terms of GPUs for total, you're about excuse me, about 3,000 versus them at maybe close to 4,000 and that's just coming from the SG and A difference that you're applying to gross profit. So without opening a capital finance company or running your own auctions, how do you kind of think about getting to the midpoint or the higher end of your long term EBITDA guidance? Yes. So I'll take a swing at that. So I think you framed it up really well. So I think if you add up CarMax through all parts of their transaction, they're closer to 4,000 in total GPU. And as we said, our midterm goal is to get to about 3,000. So I think it's a reasonable question to ask kind of what's the gap there. That gap, part of that is made up by pricing differences. I think that's really the majority of the gap in our long term plan. And then I think there are a couple of different areas where we think that over time will be relatively better or maybe not quite as relatively strong as they are. So I think they are uniquely strong in wholesale and have built up that business over a period of time. And I'm not sure there's anybody that does it better than they do. And we don't anticipate getting to the same wholesale penetration or profit per unit that they get to. I think that is something that we will drive to over time, but that's not something that we're expecting to achieve to get to our $3,000 target. I think somewhere where we think we debatably can do better is in financing. While they do have a captive finance company and that enables kind of monetizing finance receivables in a more complete way on those receivables that you keep, we've built a pretty unique finance platform that allows us to, while not taking credit risk, monetize finance originations across the entire credit spectrum, not just at the top end of the credit spectrum, while passing that credit risk on to third parties. And we believe over time that could end up being more efficient across the sum of all customers. So roughly that gives you a sense of kind of the different pieces. I don't think we're going to break them down further than that. Great. Thank you guys. Thank you. Our next question is from Colin Sebastian with Robert Baird. Please go ahead. Thanks and congratulations on the quarter. Outside of the high level cohort analysis, curious if you're seeing essentially the same buying behavior credit profile in the new markets? And if that's true, then is there any reason why you could not perhaps accelerate the pace of market rollouts? And then secondly, and I apologize if you addressed this already, but with the tax refund delays, if you could quantify what impact that will have on Q2 in terms of the strength that you're anticipating in that quarter? Thank you. Sure. So I'll start with that one as well. So in terms of buying behavior, I think when we launched this business, we expected to sell to millennials that were buying electric cars. And I think what we found since the very beginning is that that was not true. We're selling across the age demographics, across income demographics, gender mix looks like the rest of the industry. So, we're selling to a very broad based customer. I think the reason we've arrived at for that is that younger people are more comfortable adopting new technologies and so that gives us an advantage there. Older people have probably bought cars multiple times in their life and are maybe more open minded to an alternative. And so I think we've got an angle to pitch there. So we have not seen that change across markets. We've seen very steady results with the exception of the general impact where new markets are ramping a little bit faster than old markets, but across demographics and everything else, nothing notable has been changing there. And we view that as very positive because we think it means that we're not kind of quickly penetrating a small niche in the market. We believe that we're kind of growing into the belly of the market and that means that there's a lot of headroom for us to continue to grow into. So that would be kind of the answer there. As far as accelerating markets, I think the pace that we're kind of counting on today it's a pretty significant pace. Opening 16 to 18 markets is a real undertaking. Now it is very capital light. And so just to remind you what we need to do to open a market, generally speaking, we need to find real estate to stage deliveries from. We have to buy a couple of single car haulers that cost about $75,000 each. We need to buy a multi car hauler that will connect to our logistics network. In general, those costs about $250,000 each. We drop in our expansion advocate team that go in and hire new employees and train them up on all of our processes. And then we turn on marketing from home office, which again, just to reiterate the point, as we roll into national TV marketing, that will get more and more efficient over time. The economics of that are pretty good because the other thing that happens when you open a new market, that market immediately benefits from the higher GPUs that we have today across the company, whereas Atlanta had to deal with the smaller GPUs, but the same general expenses early in its life. So I think the economics suggest we should grow as quickly as we possibly can. I think we are also seeing significant market share gains across all of the different markets that we're already in. And we want to make sure that we're rolling out at a pace that's sustainable and that we're able to maintain the quality of experience that we're delivering to customers across the entire transaction, buying the right cars, reconditioning cars and getting them into the right kind of shape that customers expect when buying a car online, delivering a website that functions very simply and is easy to use, and then kind of handling any phone calls and questions that come in from that customer and then that last mile delivery to make sure that we give the customer an incredible experience they're going to want to tell their friends about. And so, I think the real thing that will slow us down there, I think, is going to be just making sure that we're opening at a pace that is responsible. I think the economics, both from an earnings and from a cash flow perspective, are pretty positive pretty fast. And so our goal of 2016 to 18 is something that we feel good about this year and then we'll continue to reevaluate that as we move forward. And then Collyn, I'll hit your question about the impact of tax season on Q1. So we estimate that the impact of the delayed tax season was about 10 percentage points of year over year growth. So that equates to 400 to 500 units of units in Q1. Thank you. Our next question is from Sharon Zackfia with William Blair. Please go ahead. Hi, good afternoon. Most of my questions were answered, but just to I guess follow-up on the tax refund dynamic. In the past, I know this is a little unusual, but in the past, if you've had a tax refund delay, have you normally seen those sales come back pretty quickly? Or did you see a stronger second half of the first quarter versus first half, for example? And then secondarily, on the kind of dipping your toes into national advertising, are you seeing an increase in the percent of sales outside of markets that you ship for free? Great. Okay. So let's start with the first one. So in terms of the delay to tax refunds, I think it was unique this year and it was unlike anything we've seen before. So there was a lot change that resulted from I think there being a significant amount of fraud related to tax refunds in the past that caused the government to delay sending those refunds out. What that resulted in is a couple of week delay. It also resulted in a compression. So we saw more of that money come in more quickly, which was kind of a new operating challenge that we hadn't dealt with historically. I think our view is that we would expect that to continue to unfold similarly in the future. And so, we will be planning our inventory plans and our operating plans and our staffing plans, etcetera, around that going forward. But I do think it surprised us a little bit this year as it was different than we've seen in the past. We have seen historically some variability in the timing of those tax returns, but it's generally been much more muted and kind of the wave looks similar, but the timing that it starts has been a little bit different. And roughly speaking, I would say that trend has been slowly moving out over the last 5 or 6 or maybe even longer years. It's just kind of been slowly moving more and more into the back half of the first quarter. So, I think as it relates to the future, we're going to model in the future similar to how it was this year, because we feel like that created a concrete time that the government needs to release those refunds and we expect to see them similarly going forward. And then your second question was on the impact of national TV. So I would emphasize we're dipping our toes in. I think that was a good way to put it. And starting to get to the phase where it's cost effective to turn on national advertising at low intensity levels. And I think we'll increase that over time as we continue to add markets. In terms of the impact on out of market sales, we are seeing some impact there, which is not surprising. But I think the bigger impact that we're even more excited about is one thing that we've noticed in our markets is that some previous awareness on the part of customers, typically due to proximity to other markets where Carvana has been operative, those seem to be an accelerator of market growth. And so I think one of the things that we're most excited about with the limited national TV that we're doing today and the increased amount of national TV that we'll do over time is that it very well could have the impact of accelerating market curves relative to what we've seen. So I think that's where that's one thing that has us excited. Okay, great. Thank you. Our next question is from Dan Salmon with BMO Capital Markets. Please go ahead. Hey, good afternoon, everyone. Could you guys maybe give us a quick update on your mobile efforts? When do you maybe expect to Sure. So, I'll first respond to our Sure. So, I'll first respond to our mobile efforts in general. So, I think the first priority as it relates to mobile was making our website very, very easy to use on a mobile device and we made that choice for two reasons. 1, if you make it work in a browser, it works across all devices. And then 2, consumers for the most part are buying cars once every several years. And so it's not necessarily a repetitive transaction. We're having a direct link on your homepage to the website is necessarily super valuable. And so we focused first on the website. I think we've had really significant and strong progress there. We rolled out in October of last year, a new mobile optimized purchase process. Prior to that, it was possible for customers to complete the entire purchase on their phone, but it was not the most straightforward exercise. What we've seen since then is we've gone from virtually 0% of customers that went entirely through the purchase process on their phone to about 15% of customers that are going entirely through the entire purchase process on their phone. We see more than that cross device shopping. And I think that that's the most important thing. We want to make sure that we do as we build tools that enable customers to act in whatever ways customers want to act. So I don't think a key metric for us is necessarily trying to drive the mobile sales rate up. I think a key metric for us is making sure that we build tools that enable customers to do whatever it is that they want to do. Now we also are working on a post sale app. A post sale app we feel is fairly valuable because it is something that the customer will use multiple times. There are many checkpoints, check-in points that we can have with that customer that we believe to be valuable to both us and them. And that's something that we are developing today and testing today and something that we are likely to roll out in a more fulsome way through the end of this year. But that's something that we're working very closely on. Great. And then maybe just one follow-up as you build the brand nationally Ernie, what are the key metrics that you're looking at to track the progress there? And I'm thinking things like unaided versus aided brand awareness, net promoter scores, things like that. Yes. So let's talk about the drivers of that. I think if we go all the way back to kind of square 1 here and say, why should a customer buy a car from Carvana? I think if we look at the fundamental underlying facts, 18% of U. S. Car buyers are test driving 0 or 1 car prior to buying a car. About 33% of customers are test driving only 1 car. If we can deliver on our promise, which is give customers a bigger selection of cars than they can get in any other single place, deliver it to their door as soon as the next day, sell it to them for $1500 less than they're likely to pay at a dealership and then give them a 7 day return policy. Rationally, customers should transact with us. And so I think the next question is, how do we do a good job telling that story and how do we get customers comfortable wanting to buy a car in that way? And I think that's all about first getting the awareness at all. People have to know that you exist. I think secondly, people need to understand the way that the business model functions. They need to get that they can go online and buy a car. And that truly means an e commerce experience that feels like what you would get if you went to amazon.com, not like what you would get if you went to a dealership website. And I think that conveying that story is not super straightforward. I think that's something that takes time and that we have to work hard at. I think the next thing is, okay, now I understand what you do and I kind of get the value proposition, but do I really trust the car quality that's going to show up to my door and do I really trust that the 7 day return policy is going to be as easy as you say it is. And that's all about credibility. And I think that takes even more time than explaining our business model. So, I think those are the series of goals that we need to achieve. And I think national advertising is a big part of seeding that awareness and then starting to tell the story of what we do. I think some of the brand assets like the vending machine are also a very important part of telling that story as well as establishing credibility because when you drive past the big, blowing car vending machine, it gives you a sense that there's a real company underneath it. And then I think credibility is simply the function of us delivering very high quality customer experiences over and over and over again, because I think we're reliant on to continue to grow and to have kind of penetration levels in markets that are higher than we've ever seen before. I don't think it's about waiting for macro trends where customers all of a sudden decide they're ready to buy cars online. I think it's about normalizing the behavior. I think if your friends and neighbors bought a car from Carvana and had a great experience and you go on the website and realize the price is lower and there's a great selection, I think that people are going to do that. So, I think why I would spell it out. I think there's a lot of stuff for us to do. National advertising is a big, big part of it. And then just building a brand generally has a lot of different components. And most importantly, always is just delivering on customer experiences one at a time that are really, really strong. Our next question is from Ron Josey with JMP Securities. Please go ahead. Great. Thanks for taking the question. Just, two quick ones here. On Atlanta, I think you said it was good to see 1,000 cars per month, I think being sold. I just wonder, can you compare that to Nashville and perhaps other markets somehow knowing of course each market is different in terms of size, tenure and penetration rates. But then as it comes back to inventory, which is one of the biggest questions we consistently get, if one market is seeing 1,000 cars per month sold and with around 7,000 cars currently in stock, can you just help us understand the dynamics between constantly acquiring cars based on demand while growing inventory levels at like at a slower pace than growth? And then for Mark, just on the cash balance, with 1Q cash ending as it did, of course, IPO certainly changes things significantly. But is there a minimum threshold that you need or you need to work with? Thank you. Great. A lot of good questions in there. So, Atlanta compared to Nashville, I mean, the way that I think we think about this internally and the way that I think we would suggest that you think about it as well is the way that we outlined it in the S-one. And so, we think about this as market share and then we think about it as population size inside of a market. So, Atlanta is about 1.5 times larger than the average of all of the markets that we have launched. And so, it requires 65 the market share to get to the same unit sales. I think that's a good way to kind of think about how markets are reacting relative to one another and what we should expect out of markets as they're achieving similar market share growth, because we do see very kind of similar and consistent growth in market shares across markets. As far as the way that markets are growing, I do think we want to stay away from giving too much color on that at a market by market level. But again, I would just reiterate that it's been very consistent across time and that we feel good about our ability to forecast that as a result of that consistency. Acquiring cars at an accelerating rate relative to sales or I guess acquiring cars at a decelerating rate relative to sales, I don't think that there's a lot that really changes there. We're buying cars. Historically, we basically had to buy enough cars to satisfy all the sales that we were getting. And then we had to buy more cars to make sure that we are growing our inventory. And we've sort of taken that second part of vehicle acquisition out for the time being while we kind of crank down average days to sale. And so effectively what it means is just buy fewer cars and replace the cars that are being sold, which in many ways actually relieve some operational burdens because it also reduces the number of cars that we have to park in any given place. So we need less land to sell the same number of cars. So I don't think that there's a lot to that. There are a number of little things that pop up as you're trying to manage through that, but I don't think there's any real showstoppers there. And then I'll take a first swing at the cash question as well. What I'll say on that is, when we size the IPO, we are very thoughtful about that and we had 2 considerations. One consideration is we wanted to make sure that we got enough cash to manage through our plan. And we did that and those expectations have not changed. The other consideration is we did not want to take more dilution than we absolutely needed to because we felt really good about the company and the business that we're building and didn't feel like that was a point in time when we wanted to invest in building an enormous capital cushion. So, I think the question is a completely fair one. I think the capital cushion we believe is there and we believe that will demonstrate itself over time. If you take a look at kind of our EBITDA margin, I think it gives you a really good sense of how you can project where cash is going to go over time. We basically have kind of 2 impacts, 2 different forces that are kind of showing themselves today. 1 of those forces is rapidly improving EBITDA margins and then one of those forces is very strong growth. As you look through the rest of this year, I think you should look at those forces and assume that they're going to roughly kind of balance out and that's effectively what's in our guidance. And that means you should see improving EBITDA margins, but relatively flat EBITDA losses. As we head into 2018, we expect the EBITDA margin impact to kind of overwhelm the growth impact and we expect that to dramatically reduce our cash consumption at that point. So that's how we're thinking about it. And Ron, the only thing I'd add there is, I do think it's useful for us because we do rely on working capital financing to look both at cash and liquidity. And so we definitely think about we certainly think about cash very, very carefully, but also take into consideration liquidity available on debt facilities. Very good point. Thank you. And our last question will come from Steve Dyer with Craig Hallum. Please go ahead. Thanks. Good afternoon. A couple here that you haven't gotten to. Can you remind us what percentage of your sales you originate the loans for and maybe where you'd like to see that over time that seems like a lever? Sure. So approximately 70% of our customers choose Carvana Financing. That's a number that we're very proud of. And I think a number that's enabled by the technology that we built that provides a very seamless experience to customers who could very easily fill out a credit application form on their phone or desktop and then select from a wide variety of financing options that we make available in a very transparent way. I think the when we think about our gross profit growth over time, we think there are many drivers that are much bigger than changes or improvements in the attachment rate of financing. Again, at 70%, we obviously have a very high take up of that. And so we're not modeling in meaningful increases there. Okay. And then just your GPU guidance both for the year as well as kind of the midterm goal of getting that to 3,000 ish or so. How would we think about bucketing that just amongst retail, financing and other and then I guess wholesale will be probably de minimis? Yes. Hey, Steve. So this is Ernie. I would say, I think about that as existing in 2 buckets. As discussed earlier, I think the first bucket is the effects that are driven directly by scale and that is about half the walk from where we are today to where we want to be. And then, I think the other half are the effects driven by product enhancements and pricing optimizations, which is basically how we've gotten all of our growth historically. So, if you go back to 2014, our average total GPU across transaction was about $200 negative. If you go to 2015, it was about $200 positive. In 2016, it was about $1,000 positive. So I think we've demonstrated the ability to build out product that pushes average GPU across the transaction up. And that's the 2nd bucket. So again, just to reiterate, half the walk comes from scale, sell more cars more quickly, drive down turn times, have less depreciation, customer gets the exact same deal, but it results in a higher margin on the car, more throughput through the inspection centers, lowers cost of goods sold. Again, customer gets the exact same deal and it results in higher GPU and then the other half are all the product build out. So we're going to And then Steve, to map that back into the specific lines on the income statement that you asked about, that half that's due to scale, that's primarily related to the vehicle. And then the other half that already described is primarily related to the other parts of the revenue streams, including wholesale and other attachment products. Okay. That's very helpful. Just a couple more quick ones. Just again, sort of circling back to the cash and capital intensity conversation. How should we think about, I guess, number of vending machines a year? You kind of guided to 16 to 18 new markets. Vending machines have certainly lagged that for good reason. What how should we think about the rollout of the actual physical vending machines going forward? Yes. So what I would say on that is, we think the vending machines are a great investment. As I said earlier, the response that we've seen has been very strong and has been multiples times better than we would need to have a solid ROI on that incremental investment. So, we want to roll out vending machines as quickly as we can. What I will also say is it is by far and away the most operationally and capital intensive thing that we do. Acquiring a vending machine site means going out and finding a location, negotiating a deal with the landowner, going through entitlements with the city and getting approvals and then physically building the space itself. That's a long chain that takes time. So we've got a real estate team that's working very hard to get as many vending machines up as they possibly can, but that is something that I think you should expect to lag market openings for the next couple of years. And over time, we're very motivated to accelerate that, but I do think that that is the slowest part of our rollout strategy. Got it. And then just a clarification lastly for me, I think you had indicated your guidance for the year sort of assumed that new markets, the markets, the new cohorts sort of grow and gain share at a rate similar to some of the early ones, but yet you're seeing some of the more recent ones trending ahead. Is that fair? Did I hear that right? That is fair. Just to make sure I'm very clear on that, I think we've seen a clear trend in the way that markets ramp. And for the most part, that trend has been that as we've matured as a company, new markets ramp faster. I think there is a small degree of that built into our expectations. And I think that what's not built into our expectations is an acceleration of that effect that could be driven by more brand awareness because of the IPO or national advertising or anything else. We do not expect that to show up in a dramatic way over the next 6 months. We do think that it is upside potential over a longer period of time, but it's not modeled in or expected today. All right, great. Thanks very much. Thank you very much. This concludes our question and answer session. I would like to turn the conference back over to management for any closing remarks. Yes, this is Ernie. Thank you very much to everyone for taking the time to join our call. We hope you'll stay tuned as we continue to execute on our plan and expand the availability of Carvana's new way of buying cars nationwide. To everyone on team Carvana out there, thank you for everything you've done to get us here. Achieving these goals relies on nothing more than us executing on our plan. Let's do it and let's have fun along the way. Appreciate it guys. Thank you very much. The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.