Carvana Case Study
by @scuttleblurb x @AndrewRangeley

this was excellent - a thoughtful and balanced thesis, Includes: overview of the problem $CVNA solves, how they delight customers, their competitive advantages, unit economics & the short thesis

The problem: Buying a used car is a subpar experience where there are sharky salesmen whose commission is tied selling us high-priced cars. The lots are small, with limited choice and there’s a lot of paperwork.

Version 1.0 of rethinking the car buying experience: CarMax. The idea was applying best concepts of retail to car buying. They had sophisticated inventory management, RFID tags on all cars to tag how quickly certain models were turning. They were shifting from a variable cost

model to a fixed cost model, relying more on data. Then in 2012, there were a few online first, asset-lite players. They didn’t own the logistics, but Carvana emerged as the most vertically integrated and owned the logistics and reconditioning centers – as well as the financing.

CarMax still sells more cars in 1 year than Carvana has sold in its lifetime, but has half the EV. It’s still a very fragmented market, and CarMax, the largest national player, has less than 2% market share. CVNA has 0.5% share nationally, roughly.

David is in the process of getting a van right now from CarMax. He had to pay the shipping costs to get the car model shipped to him, which required a 15 minute conversation. Then he has to go to the lot, fill out paperwork, test out the car. It’s not a big deal, but with CVNA,

you can order online, the car is shipped to your door, and you have 7 days to test out the car and return if not satisfied. The user experience is quite different.

CVNA’s online UX is also quite different to Vroom and Shift - their website offer more functionalities

(ex: car defects, full angles shown). Scale economies in UX can possibly compound over time.

If CarMax wanted to replicate CVNA, they’d have to reconfigure their supply chains, since they have 200 stores and are selling 6000 cars/store – so still a pretty local model.

CVNA’s inventory is nationally pooled. Each Carvana Inspection and Refurbishing Center (IRC) is huge and can sell 50 000 cars. Then they have the last mile logistics around them. So CVNA cuts out the local selling points.

CarMax is trying to have an omni-channel experience, launching in 2018. Traction hasn’t been great, and only 10% of cars are being sold through home delivery/curb-side.

By investing heavily in the IRCs & logistics, Carvana minimized the distance between inventory pools and the

consumer – so sales are converted faster and consumers get their cars shipped more quickly. And so CVNA has high NPS scores (similar to Apple’s) & there's word of mouth. This draws more sales, so fixed costs are leveraged, some into price and some into IRCs and so on.

Carvana is advertising, penetrating new markets and gaining more reach. Gross profit/unit has trended higher because 1) they cross-sell a lot of ancillary revenues (ex: insurance, financing) (cc @RyanLPape , an Xpel x Carvana partnership would be value-adding for both

parties, 2) get leverage on the IRCs they open, 3) source more inventory from customers than through auctions. They’ve gone from less than 20% customer trade-ins to 40% now.

Compared to sourcing inventory through auctions, it’s better by 500-1000$/unit to source through

customers. There’s also a feedback loop between inventory and leverage from advertising.

As they’re going through more product, they’re getting leverage through operating expenses (advertising, vending machines, etc.)

Gross profit/unit (GPUs) has climbed from 177$ in 2015 to 2323$ now. And SG&A/unit has fallen from 4500$/unit to 3700$/unit. Eventually those 2 lines will converge and gross profit will overtake S,G&A.

Comparing GPU at Carmax vs Carvana: 2400$ vs 2300$. Carmax has a lot more

wholesale exposure though. But there is a big disparity between retail GPUs: Carmax is at 2200$ and CVNA at 1400$. For wholesale, Carmax is doing 1000$ and CVNA, 385$. So the bull case lies in the closing of this disparity in gross profits/unit.

Carvana has a lot of leverage to go – its IRCs are only 50% utilized so there’s a long way to fill out just current capacity. By end of 2020, their IRCs should be enough to sell 600k cars, even though they currently only sell 200k. Carvana has been around for 27 years and does 8%

EBITDA margins. David expects Carvana to do 11% margins (mgmt. expects between 8 and 14%) so to some degree, margins will be higher. They don’t need to hire more salespeople, each car lot is selling 8x more cars than a store unit at Carmax.

David’s bigger interrogation is around the growth rate. If they were to grow units by 40% each year over 7 years, getting to 5% market share and put 11% EBITDA margins and a 30x multiple, you could compound at high-teens (this was in August). But in their oldest market (Atlanta),

where they have 2% market share, sales growth is only 18%. And in the markets where they opened between 2013 and 2016, the growth rate is only 50%. But Carmax in the late 70s was growing at 70% a year and less than a decade later, down in the low teens. So could there be

constraints (ex: sourcing inventory or growth rate could fall a lot more sharply)? Keep in mind they’re currently growing in the triple digits.

But CEO said last year: there were 40M cars sold last year in America and CVNA could get to 5% market share, but it’s such a pain to

sell cars now that the average car only trades hands every 6.5 years. If they get that down to 6 years, the market’s 15% bigger and Carvana could sell 2-3M cars. (that’s the hyper bull case). But there’s some questions around the sustainable growth rates.

The short thesis around CVNA: GPUs mostly comes from financing, but this is too static and it's really part of the bull case. There’s also a lot of equity offerings to finance growth and reshore the BS. But it's likely necessary, and returns on that dilution have been good.

end