Carvana: the online used-car company that nearly went bankrupt before it recovered

Where used-car retail was in 2012

Buying a used car in 2012 was one of the most-disliked retail experiences in any consumer category. The showroom was opaque, the pricing involved hours of negotiation with a salesperson on commission, the financing was structured around dealer-friendly markups, and the inspection process was something you mostly had to trust. Most consumers tolerated the experience because they had to — there was no functioning alternative.

Carvana’s thesis was that the entire used-car purchase could move online if the operational infrastructure existed to support it: inventory acquisition, vehicle reconditioning, financing, delivery, and returns. Ernie Garcia III had family-business background in auto finance and knew the operational economics. The bet was that consumers would pay a small price premium for a transparent, no-haggle experience — and that the cost structure of online retail (no commission-based sales force, fewer physical locations) would absorb that premium and still be profitable.

The build-out

Carvana launched in 2012 with a simple website and a few vehicles in Phoenix. The first decade of work was almost entirely operational: building reconditioning centers, hiring buyers to acquire inventory at auction, financing the inventory holdings, building delivery logistics across the country, and figuring out how to do title transfers in 50 states with different rules. None of that was visible to customers, but all of it was the actual business.

The brand layer that became visible was the “car vending machine” — glass towers in major US cities where customers could pick up their car by inserting a giant Carvana token and watching the car descend through the tower. The vending machines weren’t practically necessary (most cars were delivered to driveways), but they were brand assets that generated press, social-media content, and free awareness no advertising budget could buy.

The 2022 crisis

Carvana’s stock hit a peak of about $370 in mid-2021 as pandemic-era used-car demand surged and the company expanded aggressively. The expansion required holding more inventory, which required more financing, which became expensive fast when interest rates rose in 2022. The same time, operational issues (delivery delays, title-transfer problems, customer-service capacity) created reputational drag in a number of states.

Through 2022, Carvana’s stock dropped about 99% from its peak. The company was widely expected to go bankrupt. The combination of rising financing costs on auto loans, falling used-car prices, inventory write-downs, and operational issues at scale produced a quarter-after-quarter losses cycle that looked terminal. By the end of 2022, the company was negotiating with creditors and seemed unlikely to survive 2023 intact.

The recovery

Carvana didn’t go bankrupt. Through 2023 the company restructured debt, cut operating costs, reduced inventory holdings, paused expansion, and focused on operational discipline. Each quarter through 2023 and into 2024 produced incremental operational improvements. By 2024, the company was profitable on an adjusted EBITDA basis and the stock had recovered substantially (though still well below the 2021 peak).

The recovery is real, but the underlying lesson is unchanged. DTC in high-AOV, low-frequency, regulated categories like used cars is structurally harder than DTC in subscription or repeat-purchase categories. The launch playbook worked. The operational and financial discipline required to sustain the model through interest-rate cycles, used-car-price swings, and competitive pressure is much harder. Carvana is still in business and still growing, but the story is a cautionary one about the gap between launch innovation and operational durability.

How RGM thinks about DTC in high-AOV categories

When clients ask about applying DTC to high-AOV categories (cars, furniture, appliances, mattresses), the Carvana case is useful as both an example and a warning. The category-disruption opportunity is real — the existing retail experiences are bad enough that a transparent, online-first alternative can take meaningful share. The operational complexity is also real — financing, inventory, fulfillment, and returns are much harder at $20K+ AOV than at $50 AOV.

The honest framework: model the post-launch unit economics under a recession-and-rising-rates scenario from day one. Plan for the chapter where competitor density arrives and CACs spike. Build the operational discipline before you need it, not in response to a crisis. Carvana got the launch right and the scale-up phase wrong, then re-learned operational discipline in public. We tell clients that the recovery is harder than the launch and that the operational backbone needs to be the priority before the brand layer.

Frequently asked questions

Related