Twenty-four screens stood between a Carvana customer and car insurance. That was an early version of the checkout Root and the online car seller were building together. Consider the timing: someone had already chosen a vehicle and arranged to buy it. Now came another application. Another small obstacle course. The teams spent months reducing that journey to three steps. The product launched in July 2022. By April 2026, the partnership had sold more than 200,000 policies.
- The bet: use driving behavior to help price personal auto insurance.
- The reckoning: heavy losses forced spending cuts and a sharper focus on underwriting.
- The move to copy: offer insurance where a customer is already buying a car.
That little checkout contains much of Root’s story. The company began by asking whether a smartphone could help insurers understand drivers better. It subsequently had to answer a less glamorous question: could that understanding support a durable insurance business? Getting from the first question to the second cost real money, real jobs, and several years.
The phone in your pocket becomes a witness
Alex Timm brought insurance experience from Nationwide. Dan Manges brought engineering experience, including a stint as Braintree’s founding chief technology officer. They founded Root in 2015. Their premise was straightforward: an insurer should pay more attention to the way a person drives. A phone, already riding along on every journey, could supply evidence that an application form could not.
Root’s app measures signals such as braking, turning, and driving times. Its original test-drive experience observed ordinary trips over several weeks, then used the results to help determine eligibility and price. The appealing customer was the careful driver who suspected a conventional quote was missing something. Here was a chance to let actual behavior enter the conversation.
There are limits to that idea. Root also considers other predictive factors; a good driving score does not promise an offer or the cheapest rate. Its current disclosures say telematics is not used in California or Maryland. And the buying experience has evolved: Root’s October 2026 Wyoming launch explicitly lets eligible customers purchase immediately without a test drive. The founding insight survives inside a broader product.
The bill arrives before the breakthrough
Root sells insurance. Its subsidiaries take on claims obligations, collect premiums, and pay covered losses. Revenue also includes fees and investment income. This puts it in the same economic contest as established carriers, however different the interface looks. Pricing must leave enough money for accidents, claims handling, customer acquisition, and the machinery needed to keep operating.
Capital initially gave Root room to experiment. A $350 million Series E was followed by its 2020 public offering and a separate $500 million private placement. But capital did not settle the underwriting problem. Root reported a $521.1 million net loss in 2021. Its later annual report identified customer accident loss ratios as the principal driver of its historical losses.
Bars show absolute magnitude. Orange denotes losses; yellow denotes profits. Full fiscal years, not adjusted EBITDA.
In 2022, the response became painful. Root cut approximately 330 employees early in the year, then approximately 160 later. Its filings describe inflation, rising loss costs, reduced marketing, and a decision to concentrate resources on pricing, underwriting, and embedded products. The reversal was measurable: 2024 brought its first profitable full year, with $30.9 million in net income.
The change suggests a useful distinction for anyone building a business around prediction. A model may find an interesting pattern long before the company can profit from it. Root had to improve the financial result surrounding its predictions. By 2025, revenue reached $1.517 billion and net income reached $40.3 million. Those numbers describe progress, with a still modest profit margin.
The newer results extend that record: Root reported $61.3 million in net income for the first six months of 2026. A profitable period cannot establish that every future policy will be priced correctly. It does make the turnaround more substantial than a single good quarter.
Find the moment, then remove the form
Carvana offered a different route to a customer. A person purchasing a car has an immediate reason to arrange coverage. Root could appear inside that transaction, using a technical integration to reduce repetitive work. Carvana also committed approximately $126 million in primary capital when the companies announced their expanded relationship in 2021. This distribution experiment came with a substantial financial commitment.
“Adding car insurance should be as easy as any other aspect of online shopping.”
Alex Timm · Carvana product launch, July 2022
The copyable lesson is the sequence: locate an existing decision, connect the systems, then cut the steps. Root and Carvana’s early 24-screen flow shows that signing a partnership does not automatically produce simplicity. The integration requires work. In September 2026, the companies extended their agreement through at least August 2028. Root has also announced partnerships with Hyundai Capital America and The Zebra, expanding the places a driver can encounter its coverage.
Root’s stated culture fits the work: curiosity, uncomfortable conversations, and judging ideas on their merits. Its careers page puts data and customer obsession up front. Those are management aspirations, rather than proof of everyday life inside the company. Still, they describe the kind of organization this product demands: engineers, actuaries, and insurance operators willing to keep revising the same problem.
A useful app still needs a useful policy
For customers, the offer is practical. Root provides familiar personal auto coverages, including liability, collision, and comprehensive. Drivers can change policy details, retrieve insurance cards, and start claims from their phones. The company says filing a claim can take about three minutes; settling one is a different process. Renters insurance is also available in selected states.
Root’s alternatives include GEICO, Progressive, State Farm, and other auto insurers. Its distinction lies in how it combines driving data, its own insurance operations, and digital distribution. Competitors can also use telematics. Root’s advantage therefore depends on execution: pricing risk accurately, acquiring customers economically, and keeping the experience convenient after the purchase.
It is a poor fit for someone outside its service area, unwilling to share driving data where required, or seeking commercial vehicle coverage. Savings depend on the individual quote. Comparing equivalent limits and deductibles matters more than an advertised discount. By October 21, 2026, Root had reached 38 states. The company’s ambition remains large. The instructive detail is smaller: a customer once faced 24 screens, and a team kept removing them.