Good afternoon, and welcome to the Carvana first quarter 2018 earnings conference call. All participants will be in listen only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star then one on your telephone keypad. To withdraw your question, please press star then two. Please note this event is being recorded. I would now like to turn the conference over to Michael Levin, Vice President of Investor Relations. Please go ahead.
Thank you, Austin. Good afternoon, ladies and gentlemen, and thank you for joining us on Carvana's first quarter 2018 earnings conference call. Please note that this call will be simultaneously webcast on the investor relations sections of the company's corporate website at investors.carvana.com. The first quarter shareholder letter is also posted on the IR website. The schedule of investor conferences we will be attending over the next several weeks can be found in the events section of our IR site, and we look forward to seeing all of you out on the road. Joining me on the call today are Ernie Garcia, Chief Executive Officer, and Mark Jenkins, Chief Financial Officer.
Before we start, I would like to remind you that the following discussion contains forward-looking statements within the meaning of the federal securities laws, including but not limited to Carvana's market opportunities and future financial results that involve risks and uncertainties that may cause actual results to differ materially from those discussed here. A detailed discussion of the material factors that cause actual results to differ from forward-looking statements can be found in the Risk Factors section of Carvana's most recent Form 10-K and Form 10-Q. The forward-looking statements and risks in this conference call are based on current expectations as of today, and Carvana assumes no obligation to update or revise them, whether as a result of new developments or otherwise. Now with that said, I'd like to turn the call over to Ernie Garcia. Ernie?
Thank you, Mike, and thanks, everyone, for joining our call. We're very proud of our strong first quarter, which was a great start to the year. We exceeded our guidance across the board and are raising our estimates for the year. This outperformance was driven by strong demand for our offering and strong execution, which allowed us to continue servicing that growing demand while delivering exceptional experiences to our customers. Q1 was the first time where our quarterly unit sales grew by over 10,000 units year-over-year. Our triple-digit growth rate has been the fastest among the public auto retailers since we launched, for the first time this quarter, we also had the fastest organic growth in absolute unit terms among all the public automotive retailers.
This is a significant achievement that speaks clearly to the quality of experiences we are delivering to our customers, to the level of demand that exists for those experiences, to the scalability of our business model, and to the quality of our execution to date. We are proud of this achievement and excited about what it suggests for the future. We had a strong quarter in GPU, comfortably surpassing our guidance. Mark will give more detail on what drove those gains, but they are broad-based and have us feeling very confident about our midterm goal of $3,000. We continue to show significant progress in EBITDA margin in the first quarter, driven by unit growth, improvement in total GPU, and the leveraging expenses that is inherent in the business. The path to profitability is clear. Q1 was a great quarter for market openings as well.
We opened 12 markets in the quarter and increased our population coverage to just over 45%. We also opened a vending machine in Tampa and opened another this morning in Charlotte. Our market ops expansion, real estate, and logistics teams continue to deliver, and we are grateful for everything they do. In April, we completed our acquisition of Car360. Car360 blew us away with both the quality of their technology and, even more importantly, with the quality of the team they assembled to build it. They bring to Carvana an expertise in computer vision and 3D reconstruction that we didn't previously have and pair that expertise with an enthusiasm and creativity that complements our own and will continue to push us to put highly functional and beautiful products in front of our customers.
We believe the combination of this expertise with our proprietary photo booth will create an experience unlike any other. We'll be rolling out some of the initial integrations prior to year-end and plan to relentlessly continue improving our merchandising technology from there, always aiming to increase the power and improve the usability of our interface to allow our customers to explore the cars they're interested in and to get them to smile while they're doing it. We've now been a public company for just over a year, and in that short time, we've made a tremendous amount of progress. One year ago, we were coming off a year of about 18,000 units. This year, we expect to sell roughly five times that amount. A year ago, we were coming off a year of $1,023 in total GPU. This quarter, we expect to cross 2,000.
A year ago, we were coming off a year of negative 23.2% EBITDA margin. This quarter, we expect to be between negative 8.5% and 11% EBITDA margin. Taking an even larger view, we crossed over the fifth anniversary of our first market launch in the first quarter. Five years ago, we were a small team with the dream that we could change the way people buy cars by being thoughtful about what the customer needed and wanted and leveraging modern technology to reimagine the entire process of buying a car. We wanted to change the way cars were produced, so we built a system to do it all online and in a centralized way. Excuse me, procured, so we built a system to do it all online and in a centralized way.
We wanted to change the way cars were merchandised, so we built a photo booth that we paired with a seven-day return policy that we believed could replace the test drive. We wanted to take advantage of the fact that eliminating test drives would allow us to centralize our inventory, so we built reconditioning and certification centers at a scale that had never been seen before. We wanted to give our customers everywhere access to all those centralized cars, so we built a logistics network from scratch to maximize our customers' selection. We wanted to change the way customers finance cars, so we built a simpler, more intuitive system that allowed our customers to complete the process in minutes. We wanted to give our customers memorable experiences, so we built a team of exceptional people and united them with a culture of customer centricity.
We wanted to do all that and prove that the unit economics could be compelling. In summary, we wanted to build a paradigm-shifting company. That was just five years ago. Over the last five years, this team has systematically turned that dream into a reality. We aren't done dreaming or building. We aren't even close. We look forward to the next five years and believe we can accomplish even more over that period than we did over the first five. Mark?
Thanks, Ernie. Thanks everyone for joining us today. Unless otherwise noted, all comparisons in what follows will be on a year-over-year basis. Q1 was the strongest quarter in our company's history, and we're excited about what this quarter means for our growth trajectory. Retail units sold totaled 18,464 in Q1, an increase of 122%. Total revenue was $360.4 million, an increase of 127%. Total gross profit in the quarter grew by 251% to $34.2 million, and total GPU in Q1 was $1,854, an increase of $685. We continue to see broad-based improvements in GPU, including a reduction in average days to sale to 70 days in Q1 from 93 days a year ago, as well as gains in reconditioning, wholesale, financing, and new products. EBITDA margin was negative 12.4% in Q1, an improvement of more than 300 basis points from the prior quarter and more than 900 basis points year-over-year.
In addition to GPU gains, we are showing significant operating leverage while also rapidly expanding our geographic footprint. We opened 12 markets in Q1, our largest number ever, and have opened six more so far this quarter. We now expect to open 35 to 40 markets in 2018, bringing our end-of-year totals to 79 to 84 markets. There are more than 200 metropolitan areas in the U.S. with a population greater than 200,000 people, and these markets collectively make up about 80% of the U.S. population. We believe our online sales model is well suited to serving customers in markets of all sizes, and that these statistics provide a useful framework for our expansion opportunities over time. For the full year, we are raising our outlook to 90,000 to 94,000 units and $1.75 billion to $1.85 billion in total revenue.
We are also raising our GPU outlook for the year to $1,975-$2,175 based on the strong execution we are seeing across all parts of the transaction. In the second quarter, we anticipate continued gains across our key financial metrics. We expect retail units sold of 20,000 to 22,000, an increase of 87%-106%. We expect total GPU to be 2,000 to 2,200, an increase of 500 to 700 year-over-year as we march toward our midterm goal of 3,000. We expect EBITDA margin to be between -11% and -8.5%, reflecting continued operating leverage as we grow. You should use approximately 143 million weighted average shares on a fully exchanged basis in Q2 and approximately 145 million for Q3 and Q4. We ended Q1 with $121.5 million in cash and equivalents.
We also ended the quarter with $46.2 million in capacity under our master sale leaseback agreement and more than $40 million in real estate assets on our balance sheet eligible to be sold under this agreement. On April 30th, we completed a follow-on public offering of 6.6 million shares of Class A common stock, yielding proceeds after underwriting discounts of approximately $173 million. We believe this additional capital provides a significant amount of operating flexibility as we continue to execute our plan. As we look forward to the remainder of 2018, we are excited about continued progress toward our financial goals. We intend to continue to expand our footprint, rapidly grow sales, increase GPU, and demonstrate operating leverage. Thanks for your attention. We'll now take questions.
At this time, if you would like to ask a question, please press star then one on your touch-tone phone. If you are using a speakerphone, please pick up your handset before pressing the keys. To withdraw your question, please press star then two. Our first question comes from David Lim with Wells Fargo. Please go ahead.
Hi. Good afternoon, everyone. The question that I have is, can you explain or can you dimensionalize why was wholesale so strong from a GPU perspective? Then can you bridge the retail GPU, the 902 versus the 555? How much of it came from the days turn inventory improvement versus reconditioning process and other efficiencies?
Sure, David. Happy to hit that. We had our strongest wholesale quarter in our company's history in Q1. I think we averaged $579 per wholesale unit sold. When you apply that to retail units, it was approximately $73 per retail unit sold of wholesale gross profit. Those are the strongest numbers that we've had to date. I think there's a number of things that are driving that. First, I think our wholesale team did a tremendous job executing and selling these units. I think we're starting to build up a brand in the wholesale market, much like we've done on the consumer side of the market, where dealers know we're a place that you can come to buy high-quality wholesale cars.
I think secondly, we talked about this before, but we acquired Carlypso in mid-2017 and placed one of the Carlypso founders over the product side of buying cars from customers, increasing our management focus on that area, and also have enhanced our technology for valuing cars that we buy from customers. Now, having said all that, we still view ourselves in the very early innings of building out a business of wholesaling cars or buying cars from customers. I think we're cautiously optimistic about how that business is going, but it's still very early and we are pleased with the results in the quarter, and we'll look to continue to build on those but are in the early innings there.
As it relates to retail gross profit, the second part of your question, we stepped up by about $137 in retail gross profit quarter-over-quarter. The biggest contributor to that sequential gain was the absence of any promotional activity in Q1. As you may recall, in Q4, we do our annual Cyber Monday promotion, which has an impact on retail GPU. In this particular Q4, we also did some hurricane-related promotion activity. That was the biggest driver of the sequential increase. I think, as you may recall on the previous call, we thought Q1 retail GPU might be a little softer than we otherwise would have expected, given the patterns that we were seeing in the wholesale market and the associated depreciation rates at the end of Q4.
We were generally pleased with the number that we put up, which was part of our GPU beat, was exceeding our own expectations there on retail. We were generally pleased with that outcome, given what we were seeing in the wholesale market dynamics.
Hey, David. This is Ernie. All I would add is, I think the path to 3,000 is clearer than it's ever been. As we said in the shareholder letter, we expect to cross over $2,000 this quarter. We expect to be at a similar turn time to where we were in the first quarter. It gets really easy to extrapolate up to that $3,000 number by just imagining that we continue to push down that term over time. We continue to get the benefits of scale from better leveraging the logistics network and the Inspection Centers. From there, it's a hop, skip, and a jump to $3,000, and we think we can get there with products that we're already offering to customers. We're feeling really, really good about that path.
Just to follow up on that 902 versus the 555. Was half of that coming from essentially the inventory turnovers, improvement in inventory turn, or was it a little bit more significant than that?
Oh, yeah, sorry, David. To be clear, I was talking about sequential changes in retail GPU, just to be clear.
Right.
On a year-over-year basis, you're certainly right. We decreased average days to sale by 23 days year-over-year. That's a large part of the bridge between Q1 2017 retail GPU and Q1 2018. We also made gains in reconditioning and inventory management over that period. I think we talked about those over the course of our calls in 2017, some of the initiatives that we executed there. Average days to sale coming down was the big driver, and then gains on the cost side, particularly as it relates to reconditioning, was the secondary driver.
Got you. My final question is, can you also bridge the $879 and the $596? Is it just more add-ons, or are you getting a little bit more on the Ally side? Thank you.
Sure. Adding GAP waiver coverage in mid-2017 was certainly part of that bridge year-over-year. Another part of that bridge year-over-year was increases in our finance gross profit per unit. That came from, in part, a softer than normal quarter on finance gross profit in Q1 2017, as we were transitioning over to Ally for the first time. We made some gains there, just by lapping that softer than usual quarter. We've also made gains over time in optimizing our scoring and pricing technology, and that has led to finance gross profit relative to a year earlier as well.
Great. Thank you, and congratulations.
Thanks, David.
Our next question comes from Alvin Concepcion with Citi. Please go ahead.
Hi, this is Nicholas Jones on for Alvin. I guess first question would be, and I know it's probably still a little early, is there any update on repeat customers in your early markets? I know you commented on lapping your fifth anniversary, I think, in Atlanta. Any additional color on progress in sourcing your own vehicles? Will you need to build out a larger team to improve on that and match what CarMax is doing?
Yeah. On your first question, I think to your point, we just crossed over our fifth year in operation in Atlanta. The average consumer purchase cycle for buying a car is about once every five years. We're just heading into the meat of repeat purchases for that cohort, which was on the order of 100 cars in that first year in Atlanta. The data is still relatively thin. That said, we do have early data from customers that are buying a second car or third car much more quickly than average. I would say that the early data there looks very strong, and we're very excited about it, but it's still noisy enough and the counts are small enough to where I don't think we're comfortable quantifying that. I would say we're optimistic about that and feel good.
I think if you look at the cover of our shareholder letter, for example, there's a picture up there of a family that's bought six cars from us, and I think that's not exactly repeat customers, but it gives you a sense where when we touch someone in the family, we tend to sell cars to everyone in the family. I think the early indications there are good, but nothing to statistically nail that down just yet. As far as progress in sourcing our vehicles goes, I think I'd go back to Mark's answer as it related to the wholesale gains. This was our record quarter in average wholesale margin, and every bit as importantly, our record quarter in terms of cars that we bought from customers. I think we are making a lot of progress there that looks really good.
It is off a relatively small base, I don't know that we're ready to give visibility into what our expectations are for the future there yet. I would say the last several months have been very encouraging. The team's doing a great job. They're testing a number of things and working on just improving conversion of the funnel. Also starting to test getting more people into the top of that funnel. I think that looks good, but it's too early to quantify that as well.
Great. Thank you.
Thank you.
The next question comes from Nat Schindler with Bank of America Merrill Lynch. Please go ahead.
Yeah, hi. Is there any way you can comment at all on the average credit score of your customers that you're sending on to Ally on your financing side? Also, are rising rates having an effect, allowing for potentially more money to be from Ally for loans that you're passing them?
On the first question, I think what I would say is, in every demographic dimension, credit, age, income, gender mix, et cetera, we've seen a pretty standard mix of buyers that are buying cars from us. I don't think there's anything significant to call out there, and there hasn't been any meaningful migration in the kind of population that we're servicing. I think that continues to play out as expected. As it relates to rising rates, I think rising rates do impact us. I would say that the impact is moderately negative. When rates rise, it makes any given receivable that we originated worth less. We can take those rising rates and put them into our pricing system and pass them on as long as that's what the rest of the market is doing.
The way that we actually implement that is instead of just passing those rates straight through, we basically watch to see what's happening to consumer and street price sensitivity to make sure that we can kind of see in customer responses that the rest of the market is passing those through, then we will pass them through. That kind of creates a little bit of a natural lag there that creates something of a headwind. We don't think it's super meaningful, but it is a negative all else considered.
Okay, one more follow-up that's completely unrelated. I was wondering about your GPUs. Have you done any change to what you think your basic discounting is versus comparable used car sales?
No. That's been consistent. Nothing specific to call out there either.
Wonderful. Thank you.
Thank you.
Our next question comes from Colin Sebastian with Robert Baird. Please go ahead.
Great, thanks, and nice quarter, guys. Ernie, maybe just a couple for you to start. Could you perhaps expand on more details around the pinch points that you mentioned have emerged over the past several months? You also talked about improving conversion rates in the funnel on website traffic. Is that something you've already seen, even with the national ad campaign presumably bringing more out-of-market traffic? Is that more of an expectation you have as you add more functionality to the website?
Yeah. Well, first of all, I think, Colin, you're the first one to ever give us credit for a good quarter at the beginning of your question, so we got to thank you for that. That's a moment for celebration. I think on your first question, I think we view this as a really high-class problem. I think the first thing I would say is for the year, we're raising our estimates for total revenue and units sold. I think that what's kind of driving these pinch points is we saw roughly a 50% jump in sales roughly 45 days ago when tax refunds hit. We obviously have a plan going into tax season, and we're prepared for the expected volume increases. 50% at this scale, that increases sales on a monthly run rate basis by something on the order of 2,500 units a month.
That's basically like a $600 million a year business that shows up out of nowhere. Whenever that volume shows up that quickly, I think you're always going to find some holes in your plan. I think it's going very well, but the form that those pinch points take is just we'll see a little extra demand versus what we are ready to handle that kind of strains certain logistics routes. While most of the logistics system has excess capacity, there are certainly routes inside the system that are overwhelmed right now and that are extending out delivery times, and we tend to see those delivery times being extended out, leading to decreased conversion rates. Similarly, we now have 62 markets open.
When you have 62 markets, sometimes your staffing doesn't quite keep up with the pace of demand in some of those markets, and that can also lead to pushing out delivery times, which also can constrain demand a little bit. I think those are what are driving it. Now, as it relates to the demand that we're seeing, I would say we're seeing very high quality demand and are extremely excited about it. Basically, have seen that all this year, starting in January and flowing all the way through. We're definitely seeing much higher growth rates at the top of the funnel than we're able to convert into sales today because of some of these pinch points, but we're working very quickly to resolve them. That demand is what has us feeling so good about the year.
Okay. Then maybe Mark, regarding the advertising expense, I was wondering if you could break that out between the portions that are brand or national ad spend versus more direct response or local spend, and how you expect that to trend over time.
Sure. At a high level, our marketing mix in terms of brand versus more direct has not changed too much. I think we've hovered around 50% TV as a fraction of the total spend for quite a while now. That's still roughly the case. Within that TV bucket, we are transitioning more toward national advertising as we expand our geographic footprint. That's one of the places where we see a lot of positive feedback in opening more markets. One of the nice things about opening markets is the more markets we have, the more efficient it becomes for us to utilize national TV spend as opposed to more local brand channels. That has good positive impacts overall. I think the results that we're seeing early on from our marketing spend are positive. Our 2018 cohort has the lowest first quarter CAC that we have on record.
We're excited about that.
I do think as we look forward over time, we'll continue to expand our national TV advertising as a share of spend. Again, that just comes with opening more markets and expanding our service to a wider set of customers nationwide.
All right, thank you.
Thank you.
Our next question comes from Sharon Zackfia with William Blair. Please go ahead.
Hi, good afternoon. I guess, Ernie, you were talking pretty quickly about those pinch points that you were seeing in satisfying demand. Am I understanding correctly? Is it you don't have adequate amount of haulers to get the cars to people in the time that you'd like to? Or could you go over that again? I guess I'm asking that in relation to logistics costs because they kind of went up on a per car basis for the first time, I think, at least as far back as my model goes. I'm wondering if that's kind of tied together.
Yeah. First of all, I guess just to yes is the answer of sort of, I would say. If you think about we have 62 markets now, there are many different logistics legs that are connecting all those markets together. We've got legs connecting to different inspection centers, and we've got legs connecting out to all the different markets. When we're shipping cars to customers across the country, oftentimes cars go through many different legs. All you need is a pinch point in any of those legs, and it can result in delivery times being extended out for some customers in markets that rely on those legs. That's the form that's taking. Now, we also invested significantly in the logistics network in preparation for all of this growth, that's a big part of what you're seeing.
Across logistics network in general, we've got excess capacity, but we certainly have legs that are kind of constrained today.
Yeah. Sharon, one other point that I would make there is when we're growing at these very high rates, there's a bit of a lag between when we have to make expenditures on things like the logistics network and when we realize the large sales gains. In a quarter like Q1, we're building out the logistics network, getting ready for that 50% jump in demand that Ernie described. We're doing that well ahead of time, that sort of lag effect can give rise to some temporary fluctuations in something like logistics expense as well.
Okay, thank you. A follow-up. Mark, could you give us an idea of what kind of the range of CapEx might look like for 2018?
Sure. We historically have not given CapEx guidance. Part of the reason for that is one of the more unpredictable parts of our business is predicting when exactly we're going to launch vending machines. Vending machines are far and away the largest CapEx component for us. I think if you think about the key drivers of CapEx, again, we've talked about it quite a bit, it is those vending machines as well as our hauler network. Those together make up the meaningful majority of our total CapEx. In terms of specific guidance, we haven't done that because of the uncertainty around when exactly and how many exactly vending machines we're going to open. We've had a relatively consistent pattern over time. I wouldn't expect anything too crazily out of the ordinary.
Okay. Thank you.
Our next question comes from Ron Josey with JMP Securities. Please go ahead.
Great. Thanks for taking the question, guys. Just maybe two quick ones. On the day to sale outstanding, I just wanted to drill down a little bit further there, just given the importance to gross profit. So 1Q's DSO, 70, the second quarter in a row in 70s, and that's despite the inventory ramp into the 1Q seasonality. So Mark, is it fair to think basically going forward, we're most likely to see DSOs continuing to come down rather than revisiting the 90-100-plus DSOs that we saw last year? Ernie, just as a follow-up to your comment on the pinch points and there's certain legs that are constrained today, totally understand that, when you're launching new markets, are you doing so to avoid those overstressed legs?
Can you just talk about the trade-off between perhaps slowing down market rollouts, excuse me, to basically get those legs up and running versus accelerating them as we're doing now? Thank you.
Sure, Ron, to hit the point on average days to sale. We've obviously made a tremendous amount of progress on bringing down average days to sale. We were at 97 days as recently as Q3 2017, and have pulled that down to 70 days in Q1. We do expect that to fall a bit further in Q2. Given all the progress we made, we're looking for the rest of the year to keep that in a relatively stable and reasonable range from Q2 on the rest of the year. Part of that is driven by our inventory plan.
I think when you think about our plans for inventory over the course of the rest of this year, we expect inventory to remain relatively flat within a range up or down for the next couple quarters, and then grow toward the end of the year as we have done in the past in order to satisfy first half 2019 demand.
Yeah. Just to get into your second question. I do think so most of the stress in the system is basically moving cars out to the east, where we have our oldest and most mature markets. Many of the markets that we're opening are in parts of the network where we are less stressed. We do take that into account. We also expect to fix these pinch points relatively quickly over a matter of months. I wouldn't say it's a huge consideration. I also would say that the pinch points emerge less as a direct result, at least, of market openings. It's more as a result of just kind of the large discontinuous growth that we see as we head into kind of late February when we see tax refunds drop. There's just a lot of demand that shows up there.
We do our best to plan for that demand ahead of time. As I said, you have 62 markets, and you're predicting demand at that level of granularity, and then you've got many logistics routes. If you're off a little bit here or there, some pinch points are going to emerge, and then we just got to work to quickly remedy that.
That's great. Thanks, guys. Super helpful.
Our next question comes from Seth Basham with Wedbush Securities. Please go ahead.
Thanks a lot, and good afternoon.
Hey, Seth.
My first question's on the cadence of the quarter. Clearly, you guys did a great job beating your top-line expectations. Could you describe what happened from a monthly cadence perspective? What drove the upside in March, specifically?
Sure. Well, we'll give some color on that. I guess what I would say is, I think on our last call, we talked about the expected increase in demand that we would see in late February, and we kind of called out being able to handle that operationally as probably the biggest variable in our guidance. I think we did a good job handling that in March. We definitely believe that there was more demand in March than we were able to convert. I think that's much less true of January and February, because that kind of comes at a point when we're more at a steady state, and we're able to handle that level of sales.
I think, just the pinch points that emerge in March as you see those higher volumes are kind of rolling over into April and May, and we're expecting to resolve those relatively quickly, which is why we feel so good about the full year.
Got it. That's helpful color. Secondly, as we think about your vehicle acquisitions, how are those trending from a cost standpoint? Are you guys making progress on lowering acquisition costs of vehicles you're acquiring from auctions? What is that mix between vehicles acquired from auctions and other sources, and how has that trended?
Sure. I think I would describe broadly our vehicle acquisitions from auctions as being approximately even or stable. I think we're constantly looking to make improvements there. We're constantly improving our technology. We really like what we're doing there overall. It's been relatively stable over the last short time period. I think in terms of mix, that's been fairly stable as well. One thing that we've talked a lot about, particularly as it relates to wholesale on this call, is the opportunity in sourcing more cars directly from customers. I think we've seen some nice progress there in terms of the volume of cars that we're taking in from customers. That's showing up to a degree in our wholesale volumes. We're also feeling good about the margins on those cars that we're taking in.
As I mentioned earlier, we're cautiously optimistic about our abilities to source more cars from customers. We're in the very early innings of that today.
Okay, that's helpful. Lastly from me is thinking about one of your oldest markets, Atlanta. I know you don't like to talk about individual markets in too much detail, but how should we as investors think about the EBITDA margins you're earning right there as a proxy for the future when other markets mature?
I think the most straightforward way to think about market contribution is to use the cap data that we provided in our most recent 10-K, then you see our results and our guidance on GPU. One key element when thinking about market economics is the fact that our company GPU is a pretty good proxy for the average market because we do have centralized inventory, we have a centralized transaction platform, and all markets are drawing upon that centralized inventory and using that centralized transaction platform. One easy way to think about contribution is, we had a $440 CAC in Atlanta in Q4. We have an $1,850 GPU this quarter. That's the starting point. Now, there are other expenses in the market. We have to do last mile delivery.
We've got advocates in the market who are driving haulers out to the customer's door. We're spending some gas to get those haulers out to the door. That's a relatively inexpensive cost compared to the cost of operating a brick-and-mortar dealership. Those costs are relatively small, but those are there as well, in addition to the two big line items that you have.
One just I think useful and interesting data point there as well is in March, we had in terms of sales per market operations employee nationwide, we were up to 32 sales per market operations employee, which was up 40% year-over-year. You can see there's a lot of leverage showing up in those additional line items as well.
Thank you, guys.
Thank you.
Our next question will come from Steve Dyer with Craig-Hallum Capital Group. Please go ahead.
Good afternoon, guys. This is Ryan on for Steve.
Ryan.
Just one question for us. You guys have talked a lot about operating leverage. When I look at your 2018 guidance for revenue and GPU, both went up, implying better gross profit. However, EBITDA margin guidance implies a bigger loss. Can you help me reconcile the higher gross profit and what appears to be even higher incremental OpEx? Thanks.
Sure, yeah. I think we're obviously very pleased with the operating leverage that we've shown in the business. I think in Q1, we improved EBITDA margin loss by 900 basis points. That included a reduction of approximately $1,000 per unit in SG&A. That reduction in SG&A was split across two big buckets, compensation and benefits, and then other overhead costs. We do expect to see future leverage there as well. As it relates to the rest of the year, I think we're feeling good about all of our guidance. I think we saw very strong results on GPU in Q1. We're feeling very good about our above $2,000 GPU guidance for Q2. We're feeling good about the year generally.
Our next question comes from Sameet Sinha with B. Riley FBR. Please go ahead.
Yes. Thank you. I guess staying on the topic of operating leverage, the ad spend per unit kind of stayed pretty much flat year-over-year. I guess, Mike, to your point, you're talking about how you're prepping or investing in advance of your seasonally strong quarters. Can you touch on that and also talk about ad spend, and the kind of factor you use to decide on ad spend? Is it number of markets, your population, number of expected units? Can you help us think about that, especially as we try to model this out? My second question for you, Ernie, is you spoke about the pinch points. I guess one of the other pinch points I'm thinking about is just in terms of people searching on your site and not finding the right inventory. Can you quantify that?
Is there a way to figure out how many people leave or do not purchase just because they couldn't get something in the right age or right mileage or configuration, and just leave, and how you use those data points to kind of configure your future purchases. Thank you.
Sameet, on advertising leverage, let me hit that first. I think the really important point to make there is we think about advertising leverage first and foremost at the cohort level. We provided some data in the most recent 10-K that shows how advertising expense per unit evolves over time within a cohort, which you could also think about as within a market. What we've seen there is very steep declines in advertising expense per unit as markets age. AdEx per unit starts high when we first launch a market and begin our brand advertising and begin raising awareness. As that awareness grows, we see steep declines in advertising expense per unit, with Atlanta being the most mature market all the way down to $440 per unit in Q4.
I think when you're then kind of stepping up to the company level, what you see at the company level is a mix of more mature markets that have gone through that steep decline in advertising expense per unit as they've aged, paired with new markets that tend to start at a higher advertising expense per unit. It's that mix that generates the company overall number, which of course when you're opening a lot of markets, can be relatively stable based on mix effects. The other thing that I would add there is that although new markets are more expensive from an advertising expense per unit than more mature markets, we've seen very positive trends there in the first quarter advertising costs of new markets. 2017 was lower than any of the cohorts before it, as we presented in the 10-K, 2018 was lower still.
Yeah, I would just add to that, I think national marketing is an exciting tool here because I think when you think about local marketing, if you double markets, what you basically have to do is you double your marketing spend, and then you expect to see similar responses in the new markets to what you saw in the old markets. When you move to national marketing, if you double markets, you then have this choice. I want to double marketing, which then kind of doubles marketing everywhere. Or do I want to hold marketing flat, which then will lead to the same level of exposure in all the new markets, but will decrease my overall spend per sale.
I think that sort of a network effect is not unlike the inventory network effect that exists in our business as a result of our ability to access national markets. I think that's exciting, and I think that we'll continue to watch that and look at the sensitivity to national TV marketing over time and make the right decisions as we continue to learn more. That's the same choice that we've got in inventory, and I think you asked this question about inventory. There's no doubt that having the car a customer wants absolutely is a primary driver of converting any given customer. Having either a larger inventory or a more intelligently distributed inventory leads to more sales.
I think as it relates to more intelligently distributed inventory, we're leveraging all kinds of data from many different data sources, some of them external data sources. Many of them data sources of just all of our customers clicking around our website every single day that are informing which cars we want to buy, where we want to buy them, how much we want to pay for them. I do believe that we're exceptional at that. As it relates to the number of cars that we put on the site, that's this question of, at any given time, are you going to leverage growth to push down turn time by holding inventory flat, or are you going to leverage growth to be a reason to grow inventory and increase conversion everywhere?
Over the last 12 months, for the most part, on average, we've roughly held inventory flatter and we've pushed turn time down significantly. That's always a choice that we get to make. I think it won't be long until we get to a place where our turn time is inside the range that we want it, and we'll start to just grow inventory continuously with demand, which I think will be an exciting time for conversion.
Sameet, just to hit your second question very quickly, the biggest determinant of our marketing spend is the population that we're marketing to, which is obviously also heavily correlated with the number of markets that we're in.
Our next question comes from Gary Prestopino with Barrington Research. Please go ahead.
Hi. Most have been answered, but do you have the average monthly unique visitors for the quarter?
It was 1.6 million, I think in the shareholder letter, we also said that we crossed over 2 million in April. That was up from 1 million nine months prior.
Okay. Then, just in terms of inventory, Ernie, how much did you have to flex up your inventory here in Q1 to get into the spring selling season? Did you keep it flat with Q4?
No. We definitely flexed up inventory in Q1 versus Q4. I think that's something that we've done historically. I think that is commensurate with the increases in sales that we've historically seen when tax season hits. We did grow inventory in Q1. We did that while lowering average days of sale, that just corresponds to the big increase in demand that we see starting late in Q1.
Can you give us an idea of what your inventory is going into the quarter here?
Sure. We ended Q1 with just under $300 million in inventory on the balance sheet.
Okay. I'll get from units. As you grow, have you considered taking advantage of the leading auction companies to do some reconditioning for you? Especially in areas of the country where you might not have a presence with an IRC?
Yeah. We're doing that all internally. We think that that's important to control the quality of the car. We've got a really strong process at our inspection centers where we can take a car in, stock it in. We can go through and make sure we've got all the right data about the car so we make sure we can merchandise it properly. We put it through our reconditioning standards so we can make sure that we certify it to a level that we believe is a high enough level to put in front of our customers. We're able to photograph the car, as well, for merchandising it. We think that's all important for us to control, and so we're controlling that all today.
Okay. Thank you.
Thank you.
Thank you.
Our last question for today will come from James Albertine with Consumer Edge. Please go ahead.
Well, very good. Thanks for squeezing me in, gentlemen, and congratulations on a solid first quarter.
Thank you.
Wanted to ask, and as well, congrats on the follow-on offering. Wanted to focus on capital allocation here a little bit, if I may. You did an acquisition here of Carlypso. Curious how you think that's going to flow through, how we should sort of gauge the necessity in terms of timing, and then the progress you're making in terms of integrating those benefits, where we're going to see that manifest itself and how soon, perhaps. As well, as you're balancing growing into new markets, you're talking a lot about pinch points. You're talking a lot about excess demand for where you have supply to fulfill. We're not seeing vending machines carry into every market you're opening, and you're obviously leveraging your national branding to some degree there, too.
I'm just wondering, given it's one of your highest CapEx uses right now, does the ROIC that you're calculating on your vending machine additions necessitate further vending machine investment at this point? Or is it better served to sort of utilize it elsewhere, given the successes you're having in large swaths of the country without a vending machine?
Okay, let's start with Car360 there. You also mentioned Carlypso, so I'll briefly touch on that. Carlypso was the first acquisition that we did that was earlier in the year in 2017. A lot of the benefits of that are already flowing through. As Mark brought up earlier in the call, one of the founders of Carlypso is now running our efforts to buy cars directly from customers and doing a great job. We're feeling really good about where we are there and how we've been able to integrate that. We're definitely still working on many things that we expect to be rolling out over time, but in general, that's going very well, and we're very happy with kind of that partnership.
As it relates to Car360 built really pretty exceptional technology. Where you'd expect to see that is just in enhancements to our vehicle player, where customers can go on our website, click on a car, and then explore that car. I'll just say, when I got the demo from them the first time, I laughed out loud because I thought it was so incredible, and my mind was spinning a million miles a second trying to figure out how they were doing what they were doing. We're really excited to integrate that technology and give that experience to our customers. That will take time because there are many different levels of integration that we plan to do and many interesting functionalities that we expect to roll out over time.
I think we expect the first iterations and integrations to start to roll out toward the end of the year. We'll be working as quickly as possible to continue to roll out more of the incredible technologies they've built. I think we're very excited about that. I'll let Mark come back around to talk about how that flows through the financials. As it relates to vending machines, you're right, we definitely do not have vending machines in all of our markets. We're very pleased with the growth that we're seeing across the company and in both markets with and without vending machines. We do believe that we have excess demand today, and I think that's very exciting. We also believe that vending machines continue to be very high ROI.
We feel like we can measure that ROI pretty surgically, because we put the vending machines on the ground, we see the changes to sales, we see the operational benefits that we get. We can quantify those. Put those up against kind of the cost to finance those vending machines through sale leaseback arrangements that don't even end up tying up equity capital. The math remains very clear and the response remains very consistent across many of those different markets. I would say, we're feeling great about the core offering. We're excited about the benefits of national marketing as we continue to leverage that more fully. We continue to believe very strongly in the power of vending machines, and we'll keep investing in those.
As it relates to Car360 and the financials, in terms of revenue versus profit, OpEx, other than depreciation and amortization, the impacts will be immaterial. There will be some additional depreciation and amortization associated with the acquisition, and we'll provide additional details on that in Q2.
Okay, great. Listen, I really appreciate the thoughtful answer. To the degree that you've shared data in the past on your market share and trajectory there, to the degree you're willing to share it on the ROI side, I think that could go a long way, but I understand it's a trade secret. I do appreciate the comments, and thanks again, and best of luck.
Thank you.
Thank you.
This will conclude our question and answer session. I would like to turn the conference back over to management for any closing remarks.
Thank you everyone for joining the call. We had a great quarter. Much more importantly, the pieces are in place for a bright future full of great quarters. Thanks everyone on Team Carvana for everything you do every day. All companies are nothing more than a collection of people working toward a common goal, and I'm proud to be part of this collection of people, to share with you all the mission to change the way people buy cars and to be succeeding in fulfilling that goal to the degree we are. Thank you.
The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.