Why Convoy Failed: What Sank the Digital Freight Broker

Convoy failed because it was a freight broker that was funded and valued like a software company. Its app automated much of a broker's own work, but a broker's largest cost is the payment to the trucking company, which the spot market sets for every broker at once, and Convoy never had enough freight on any lane to change it. When the 2022–25 freight recession and the end of cheap venture money arrived together, it ran out of time and stopped operating on 19 October 2023.

The end came with almost no warning. On Wednesday 18 October 2023 the shippers who used Convoy found that their loads had been cancelled. The next day the Seattle company ceased operations. Eighteen months earlier it had closed a $260 million financing package, and deal-data summaries put its valuation at about $3.8 billion (reported; that figure has not been found in a primary filing). The chief executive's memo blamed a "massive freight recession" and a contraction in the capital markets (reported paraphrase). Both were real. Neither explains why a company with more than $1 billion behind it had no buffer when they came.

What was Convoy?

Convoy was founded in Seattle in 2015 by Dan Lewis (CEO) and Grant Goodale (CTO), both former Amazon employees (reported). It was a digital freight broker: it took full-truckload freight from shippers, priced it automatically, and offered it to small trucking companies through a smartphone app. The idea rested on a real feature of US trucking. The market is enormous, it is fragmented into hundreds of thousands of carriers, most running fewer than ten trucks, and much of it still ran on phone calls, email and spreadsheets.

Convoy's products went beyond matching. It offered automated reloads and shipment batching, chaining loads so that a carrier's truck ran full more of the time. It built a drop-trailer pool, Convoy Go, so carriers could hook a preloaded trailer and leave instead of waiting at a dock. It paid carriers fast. And it framed the mission around waste: trucks running empty, and broker margins that technology could squeeze out.

That pitch attracted investors from far outside freight, including Jeff Bezos and Bill Gates (reported), and set expectations of software-style growth and margins. Financially, though, everything Convoy sold was brokerage. Its revenue was the spread between what the shipper paid and what the carrier was paid, the same line as a phone-based broker's, priced by the same market. If you want the mechanics in detail, our guide to how freight brokers make money walks through a load from quote to payment.

How much money did Convoy raise?

The timing of the last round matters. It was announced on 21 April 2022, weeks after the Federal Reserve began raising interest rates in March. It was close to the last moment that kind of round was possible for a loss-making logistics start-up.

Why did Convoy fail?

Four problems compounded. Each would have been survivable alone. Together they left a company that had to close the gap between its price and its cost through operations, at the exact moment it could no longer cover that gap with investors' money.

The software attacked the smaller cost

A broker that charges a shipper $2,000 for a load and pays the carrier $1,700 keeps $300. The carrier payment is typically about 80–90% of a broker's gross revenue. The remaining spread pays for the broker's own people, systems, sales, claims and the cost of paying carriers before shippers pay. Software can cut that second bucket by letting each employee handle more loads. It cannot do much about the first, because the carrier's price is set by the spot market, which every broker sees on the same load boards.

Convoy's automation did reduce the broker's own cost per load. But that was the smaller cost. Its claim to reduce the larger one rested on something else.

Density never arrived

Batching and reloads lower the carrier's cost only if the broker has many loads, on the same lanes, at the same times. A carrier will accept a lower rate on one load if the next one is waiting at the destination and the truck does not drive home empty. That requires lane density. Convoy's share of any given lane was small, so the theory rarely worked at scale.

The bull case was a sequence. First, subsidize both sides to build a network of shippers and carriers. Second, once enough loads flowed through the same lanes, use batching and drop trailers to raise each truck's loaded miles. Third, earn a spread as good as incumbents' at a lower price to shippers, and take share. Each step depended on the one before, and the first required years of funded losses. The 2019 and 2022 rounds paid for step one. Step two was never proven at scale.

It grew fastest when margins were thinnest

A broker's margin moves against the spot market. When spot truckload rates spike above the contract rates a broker has agreed with shippers, the broker must pay carriers more than it budgeted, and margin per load compresses. When spot falls below contract, margin widens until shippers rebid. Convoy grew fastest in 2020–21, when spot rates were high and the spread was thinnest, and much of that volume reportedly came at thin or negative margins.

The stylized broker in the book's Appendix A shows why that matters. It is an invented company, not Convoy's numbers:

The broker that buys growth by cutting its margin to 5% loses money on every load even with more productive staff, and ties up far more cash, because it still pays carriers in days and collects from shippers in weeks. That is the pattern behind several venture-funded digital brokers.

The cycle and the capital market turned together

From mid-2022 spot rates fell below contract, shippers rebid their lanes at lower prices, and volumes fell. The large incumbents, C.H. Robinson, TQL and RXO among them, cut prices to hold share. The trucking market entered what the industry called the longest freight recession on record, lasting into 2025. At the same time, rising rates ended the funding environment of 2020–21 for every loss-making logistics start-up. Convoy cut staff in three rounds from June 2022 to February 2023, closed its Atlanta office and cut again in 2023 (reported). Talks to sell or merge failed, and it stopped operating rather than file a conventional reorganization.

There was also a quieter mistake. Convoy Go put capital into trailers. It solved a real problem, unpaid hours at docks, but it added fixed cost to an asset-light business whose revenue was cyclical.

What happened to Convoy's technology?

The technology outlived the company. On 1 November 2023 Flexport announced it had acquired Convoy's technology and some key staff, not the company or its liabilities, for an undisclosed sum (one secondary source says about $16 million, which is unconfirmed). Flexport relaunched it as the Convoy Platform in early 2024 and announced on 28 July 2025 that it would sell the platform to DAT Freight & Analytics, the load-board company. Flexport later described Convoy's network as more than 400,000 truck drivers and 80,000 carriers (company claim). The story of Flexport's own reset is in what happened to Flexport.

That outcome is the clearest evidence of what Convoy was worth. The software had value. The brokerage business built on it did not.

Who paid for Convoy's failure?

The shippers' reaction says the most. They moved their freight elsewhere the day Convoy stopped. A service customers can replace overnight has no switching cost, and a business without switching costs cannot charge for its convenience.

What did Convoy's rivals do differently?

The incumbents had what Convoy was trying to buy: freight volume across thousands of lanes, long shipper relationships and balance sheets earned in earlier cycles. When they adopted the same automation, they applied it to volume they already had. C.H. Robinson is the clearest case: in the third quarter of 2026 its adjusted gross profit rose about 14.5% year on year while its average headcount fell about 10.4% (company 8-K, 28 October 2026). Landstar survived the same recession with a model in which independent agents and owner-operators carry most of the fixed cost. Uber Freight survived by buying Transplace in 2021 for about $2.25 billion and leaning on managed-transportation contracts and its parent's balance sheet.

Technology was available to everyone. Density was not. That asymmetry decided the contest. The wider pattern, from Convoy to quick-commerce start-ups, is set out in the industry study of supply-chain failures and busts and in our list of the biggest supply chain bankruptcies.

Convoy and the flow equation

The Flow Equation argues that a supply-chain business creates durable value only when four factors are true at once: a cost to serve below what customers pay across the cycle, density, reliability, and capital matched to the cycle. The factors multiply, so a zero in any one makes the product zero.

Convoy had one strong input, venture capital, which is not one of the four factors at all. It had no structural cost-to-serve advantage, it never reached density, and its capital was matched to the next funding round rather than to the freight cycle. Its reliability was adequate but easily replaced. The book's Chapter 18 on capital cycles opens with Convoy for this reason: the money arrived at the top of the cycle and was spent in the trough. For how the larger industry earns and loses money, see our guide to the logistics industry business model, and for the broker economics in full, the industry study of freight brokerage and digital freight platforms.

The takeaway

Convoy did not fail because its software was bad or its founders careless. It failed because venture money bought volume, and volume bought below cost does not lower anyone's cost. Its automation cut the broker's smaller expense, the larger one stayed with the spot market, and the lane density that might have changed that never came. When the freight cycle and the capital market turned in the same year, there was nothing underneath. Chapter 18 of The Flow Equation works through why capital raised at the top of a cycle so often ends this way.