Good afternoon and welcome to Carvana's First Quarter 2017 Earnings Conference Call. All participants will be in listen-only mode. [Operator Instructions] After today's presentation there will be an opportunity to ask questions. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Alex van Leeuwen, Investor Relations. Please go ahead.

Thank you Austin, good afternoon ladies and gentlemen and thank you for joining us on the Carvana's first quarter 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 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 in future financial results that involves risk 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 prospects related to its initial public offering filed pursuant to rule 424-B under the Securities Act with the SEC on April 28, 2017. 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. 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 are very excited to share strong Q1 results with you today. Hopefully you’ve all had a chance to read our shareholder letter, which is available on our Investor Relations site. We decided to issue a shareholder letter prior to our calls in an effort to give you all the information you need and 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 have 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 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 it to truly address the customer experience we had to change the economics of the business and culturally orient ourselves toward taking pride and 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 have made the process so simple that our customers can buy car from Carvana on their phone or computer in less than 15 minutes.…

Thank you, Ernie and thank you all for joining us today. Full details on our first quarter financial results are available in our shareholder letter. We 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’m very excited about our financial model, 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% to $159.1 million. While we are extremely pleased with our growth in the quarter, we believe that our growth would have been even greater where if not for the delay in IRS tax refunds, which affected consumer spending patterns. In the second quarter, 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 $9.7 million, an increase of 146%. We improved total gross profit per unit by $123 to a total of $1,169 in the first quarter. We see this increasing another $200 to $250 sequentially in the second quarter. We are also committed to improving EBITDA margin. EBITDA margin was negative 21.6% in Q1, a significant improvement from the prior quarter. We see EBITDA margin improving sequentially to negative 18% to 18.5% in the…

We will now begin the question-and-answer session. [Operator Instructions] 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 have outlined could possibly get better and maybe some of the assumptions that are underpinning that, maybe some color on what you are expecting for maybe market share in total for you guys?

Yes David, this is Ernie, I will take first swing 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 market 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 important driver, we are basically looking at the impacts 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 in the new models in the new markets that you have 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 a 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 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 have now opened three additional vending machines aside from our first vending machine in Nashville, and we have 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 of 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 kind of more seeded awareness in the market prior to us arriving there, something that will impact that going forward is we just started to test national 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. Byproduct of that aside from reducing incremental marketing cost 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 preexisting markets. So that would be how I respond to that question.

And thank you. My final question is SG&A per unit retail looks like there were some movement in that number this quarter versus last year. I was wondering if you could give some details, it looks like the number may have gone up on a year-over-year basis by just the shape. 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 million in revenue up to $365 million in revenue, and becoming a – in 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 lot of infrastructure and corporate headcount associated with preparing to be a public company. We saw that take the form of increased SG&A per unit in 2016 in particular in the second half of 2016 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&A per unit going forward overtime.

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 too. 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 from Bank of America Merrill Lynch. Please go ahead.

Yes, hi Ernie, hi Mark. Congratulations on your first public quarter. Your GPU, you certainly had 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 retail units -- retail GPU as oppose 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've 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 1169 that we printed in Q1. Those gains are across all parts of the transaction. We don't planned to give a detailed breakdown of what we expect on each element of transaction, but I can't 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 vehicles will be in spite of slightly rising average day 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 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 two 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 straight forward. 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 two 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 off-loading a bunch of recalls in the fourth quarter. So we don’t expects those two impacts to persist.

Okay. And going to that timing element, how long you expect to be keeping cars, are you building that -- getting that scale advantage now. You’re marketing basically 7,500 plus cars available at an any given time, if you’ve doubled your sales you’d obviously have and keep that same roughly you would have your amount of time you are 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 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 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 is reason to believe we should be able to outperform that because we are 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 is certainly upside. And the way the math works is exactly as you outlined it, if we can sell twice many cars at the same inventory roughly speaking you have your turn times.

Thank you guys, good job on the first quarter.

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 the national campaign into your full year guidance, since you will be going in a bit warmer than what you have experience in the past? And then 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 two lines instead of one, and kind of how that plays out? Thank you.

Great so I’ll start this is Ernie, I will 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 oppose to buy in each of our markets and add that all up. And so we started to test that. The way the 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 is being something that is sort of continuously rolled out over the next 12 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 for existing markets, that this could cause new markets to ramp faster than they historically have that’s not something that we are banking on and 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.

And then Mark, I’ll hit the point on IRC utilization, so our three 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 83 -- 134 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.

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 pieces that are on the reconditioning on prep side and what are the pieces on the days to sale and kind of market side?

Yes, this Ernie so I’ll take that. So, days to sale, I think you can think about it’s broken up into those two components so basically three. So the first component is getting the car to the reconditioning center to get it refurbish. 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 comment 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 are up on the website in about 65 to 76 days give or take and that give you a sense of what the other two components are. And that's what we think we are 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 in those two 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 transport cars and we take that into your account when we buy them. So, we effectively are less likely to win cars that are further away because we are bidding less on them and therefore kind of the cost of carrying that car during that extra transport period is already taken into account. Once the car arrives at inspections center and is prepared for sale that is differently 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 gain that we’ll have there are small relative to the gains that we’ll get from just pushing more customers on to the website across the same fixed supply.

Got it, okay. And what -- and maybe just in Nashville at this point, in that market what's kind of the mix of delivery versus pickup that you are 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?

I mean it’s very high level, I think in the future this is probably is something 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 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 that it will be significantly different.

Great. And maybe just kind of very high level here, during the road show you had 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 are thinking about in terms of GPUs for total you had about 3,000 versus them maybe close to 4,000 and that's just coming from the SG&A difference that you are applying to gross profit. So, without opening a capital finance company you are 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, 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 mid-term 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 and our long-term plan. And then I think there are a couple of different areas where we think that overtime we 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 long period of time, and I'm not sure there is anybody that does it anybody 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 overtime, 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 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 to the top end of the credit spectrum, while passing that credit risk on to third parties. And we believe overtime that could end up being more efficient across some of all customers. So roughly that gives you a sense of kind of the different pieces, and 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 [ph] new market. And if that's true, then is there any reason why it could not perhaps accelerate the pace of market velocity [ph]? 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 are anticipating in that quarter?

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 buy 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 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 is 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 is a pretty significant pace, opening 16 to 18 markets is a real undertaking. Now it is very capital like and so just to remind you what we need to do to open a market, generally speaking we need to find…

I think Colin, I will 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 a follow-up on the tax refund dynamic. In the past I know this is a little unusual, but in the past you have had 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 have seen before. So there was a law 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. With that resulted in is a couple weeks 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 haven’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, et cetera around that going forward. But I do think it surprised us a little bit this year as it was different than we have 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 look 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 five or six 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 are going to model it in the future similar to how it was this year because we feel like that created a concrete time that kind of 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 are dipping our toes in I think that was a good way to put it. And we are starting to get to the phase where it’s cost effective to turn on national advertising at low intensity levels and I think we will increase that overtime 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 have 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 does seem to be an accelerator of our market growth. And so I think one of the things that we’re most excited about with the limited national TV that we are doing today and the increased amount of national TV that we will do overtime is that it very well could have the impact of accelerating market curves, relative to what we have seen. So I think that’s where -- that’s one thing that has us excited.

Great, thank you.

Our next question is from Dan Salmon with BMO Capital Markets. Please go ahead.

Hey, good afternoon everyone. Could you just maybe give us a quick update on your mobile efforts? When do you maybe expect to efficiently launch the mobile app and what impact do you think it could have on the business near-term once it does launch?

Sure. 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, one if you make it work in a browser it works across all devices. And then two, consumers for the most part are buying cars once every several years. And so it's not necessarily 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've rolled out in October of last year a new kind of 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 zero percent 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 across device shopping. And I think 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. And 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 rollout in a more fulsome way through the end of this year, but that’s something that we are 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 are looking at to track the progress there and I am thinking things like unaided versus aided, a brand awareness, net promoter score things like that.

Yes so let’s talk about the drivers of that I think if we go all the way back to square one 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. buyers are test driving zero or one car prior to buying a car, about 33% of customers are test driving only one 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 delivered to their as soon as the next day sell it to them for $1,500 less than they likely to pay at a dealership and then give them a seven 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 in e-commerce experience it feels like what you would get if we went to amazon.com not like what you would get if you went to dealership website. And I think conveying that story is not super straightforward I think that’s something that takes time 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 it’s going to show…

That’s great color. Thank you, Ernie.

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 to 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 that like at a slower pace than growth. And then for Mark, just on a 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, 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 I think we would suggest that you think about it as well, is the way that we outlined it in the S1. And so we think about this as market share and then we think about 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% of the market share to get to the same unit sales. And I think that's a good way to think about how markets are reacting relative to one and other and what we should expect out of markets as they are achieving similar market share growth. Because we do see very kind of similar and consistent growth in market shares across market. As far as the way that markets are growing, I do think we want to stay awake 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 times 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 is a lot that really changes there. We're buying cars historically we basically have to buy enough cars to satisfy all the sales that they we were getting. And then we have to buy more cars to make sure that we're growing our inventory and we sort of taken that second part of vehicle acquisition out for the…

And Ron the only thing I would 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 at 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 gone into, can you remind us what percentage of your sales you originate the loans for and maybe where you would like to see that overtime 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 buy the technology that we build 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 on a very transparent way. I think the -- when we think about our gross profit growth overtime, we think there are many drivers that are much bigger than changes or improvements and the attachment rate of financing again at 70% we obviously have a very high take up of that. And so we are not modeling in meaningful increases there.

Okay. And then just your GPU guidance both for the year as well as kind of the mid-term goal of getting to that 3,000-ish or so, how do we think about bucketing that just amongst retail financing and other than I guess wholesale will be probably de minimus?

Yes, hey Steve so this is Ernie. I would say I think about that as existing in two 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 have gotten all of our growth historically. So if you go back to 2014 our average total GPU across transaction was about $200 negative, and if you go to 2015 it was about $200 positive, in 2016 it was about $1,000 positive. So I think we have demonstrated the ability to build out products that pushes average GPU across the transaction up and that’s the second bucket. So again just to reiterate half the walk comes from scale, sell more cars more quickly driven 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 to 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 outs. So we’re…

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 Ernie described is primarily related to the other parts of the revenue streams including wholesale and other attachment products.

Okay, it’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 you have kind of guided to 16 to 18 new markets vending machines have certainly lag that for good reason. 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 have seen has been very strong and has been multiple times better than we would need to kind have a solid ROI on that incremental investment. So we want to rollout the vending machines as quickly as we can. What I will also say is that 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 land owner, 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 overtime 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 assume that new markets, the market the new cohorts sort of grow and gain share at a rate similar to some of the early once, but yet you are seeing some of the more recent once trending ahead, is that fair did I hear that right?

That is fair and just to make sure I’m very clear on that. I think we've seen a clear trend in the way that market 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 this was not built into our expectations is an acceleration in that effect that could be driven by more brand awareness because of the IPO or national advertising or anything else. We do not expect to that to show up in a dramatic way over the next six months. We do think that there is upside potential over a longer period of time. But it’s not modeled in or expected today.

Alright, 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 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 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 you guys. Thank you very much.

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