Customer Concentration Risk: What a Freight Brokerage Loses When Its Biggest Customer Leaves
Revenue share is the wrong way to size customer concentration. Measure the top account's share of true profit, plus the carriers, lanes, and people that leave with it.
Mithrilis Team
14 min read
Last updated: July 2, 2026.
When a brokerage loses its largest customer, the damage is rarely the number on the org chart. The account that looked like 22 percent of revenue turns out to have been carrying closer to 40 percent of the profit, three of the carriers the desk leans on every week, and the two backhaul lanes that balanced the network. So how do freight brokers measure customer concentration risk in a way that reflects real exposure rather than a revenue percentage on a slide? Most cannot, because the true picture is split across the TMS, the accounting system, and the accessorial records, and no single screen adds it up. This guide lays out a sharper method: size the top account by its share of true profit after detention, accessorials, and slow-pay, then add the operational dependencies that walk out the door with it. Get that right and you can put a defensible number on both revenue-at-risk and profit-at-risk, and stress test the book before a customer, rather than the market, forces the question.
TL;DR
A brokerage's real concentration exposure is not its top customer's share of revenue. It is that customer's share of true profit after fuel, detention, accessorials, and slow-pay, plus the carriers, lanes, and staff relationships that leave alongside the freight. Reported revenue concentration routinely understates or overstates the real risk, because true margin per customer is scattered across the TMS, accounting, and accessorial systems and never reconciled against one record. Connect those systems into a single shipment record and the concentration question gets a defensible answer: true profit per customer, a revenue-at-risk and profit-at-risk figure you can act on, and a 60-day stress test, with every number traceable back to the source row it came from.
Key takeaways
- Revenue share is the wrong denominator. A top account can be 22 percent of revenue but 40 percent of true profit, or the reverse, and the gap only shows once trailing costs are attributed per customer.
- Broker margins are thin enough that attribution changes the ranking. A FreightWaves analysis of a representative mid-market brokerage put revenue per load near $1,912 at a 9.91 percent gross margin, roughly $189 per load, so one unbilled accessorial can flip a customer's contribution.
- The freight brokerage market is fragmented, with more than 15,000 registered brokerages and only about 80 posting $100 million or more in revenue, which is why a single anchor customer is existential for most shops.
- Concentration risk is operational, not just financial. The carriers, lanes, and rep relationships tied to the departing customer are part of the exposure and rarely appear in any concentration number.
- Measure two figures, not one: revenue-at-risk (gross billings that disappear) and profit-at-risk (true margin dollars that disappear, minus what is recoverable versus stranded).
- Asset carriers carry the same shape of risk through dedicated lanes and anchor shippers, and feel it harder because owned trucks and domiciled drivers are committed to that customer's freight.
Revenue share is the wrong number to watch#
Almost every brokerage that tracks concentration at all tracks it the same way: the top customer's billings divided by total billings. It is easy to pull, and it is the number a lender or an acquirer asks for first. It is also the number most likely to mislead you, because revenue and profit do not sit in the same place on a freight book.
Two customers can each run a million dollars of billings a year and contribute nothing alike to the bottom line. One tenders clean, drop-and-hook, short-dwell freight on lanes where you have committed carriers and pays in 30 days. The other loads live at a facility that routinely holds trucks four hours past the appointment, generates a lumper fee on half its loads, disputes the occasional claim, and pays in 62. On a revenue chart they are twins. On a true-margin basis one is your best account and the other may be losing money on every third load. Rank your book by revenue and you are ranking it by the one dimension that hides which customers actually carry the business.
That gap matters more the thinner the spread gets, and broker spreads are thin. A FreightWaves analysis of a representative mid-market brokerage put average revenue per load at $1,912 and gross margin at 9.91 percent, about $189 of gross margin per load, against a break-even nearer 11.3 percent once payroll, technology, and financing are covered. At $189 of cushion, a single unbilled detention charge or a fuel surcharge that lagged the market does not dent the margin on a load, it erases it. TIA's own quarterly benchmark, reported by FreightWaves, has recorded average truckload broker gross margin as low as 11.4 percent on an invoice near $2,000 per load, with the largest brokers running thinner than smaller shops. The margin you are dividing up per customer is small and uneven to begin with, so when the per-load buffer is that thin, the trailing costs attributed to each customer decide the ranking, and revenue share cannot see any of them.
Nor is this a large-broker problem you can ignore as a small shop. The market is fragmented: FreightWaves has noted more than 15,000 registered freight brokerages in the United States, with only about 80 of them posting $100 million or more in gross revenue. For the thousands of brokerages well below that line, a single anchor customer is not a concentration statistic, it is the difference between making payroll and not. The same article flags highly concentrated customer bases as one of the constraints that stall brokerage growth in the first place. And brokers sit in the middle of a very large market, U.S. business logistics costs reached $2.4 trillion in 2025, about 7.8 percent of GDP by the CSCMP State of Logistics Report, while capturing only a thin slice of each load, which is exactly why losing one anchor account lands so hard.
What "true profit per customer" actually includes#
Reported margin per customer is the buy-sell spread your TMS shows at booking, summed across that customer's loads. True profit per customer is that spread after every cost that arrives later lands against the right account. The two numbers diverge because the later costs live in different systems and settle on different clocks.
| Cost layer | Where it lands | Why it hides in the revenue number |
|---|---|---|
| Fuel surcharge lag | TMS or accounting, days later | The surcharge billed to the shipper can trail the price you paid to cover the load |
| Detention and dwell | Accessorial or facility records | Often unbilled or written off, so it never reduces the customer's reported margin |
| Lumper and other accessorials | Separate accessorial ledger | Recorded away from the load, so per-customer totals are rarely rolled up |
| Claims and OS&D | Claims system, weeks out | Adjudicates long after the load closed, so booking-day margin looks intact |
| Slow-pay carrying cost | Accounting aging report | A 62-day payer ties up working capital that a 30-day payer does not |
Each row is a real cost of serving that customer, and each one sits in a system that does not automatically reconcile against the shipment. Stack them and the customer that looked like a solid 12 percent account can settle at 6, while the demanding-but-clean account you were tempted to fire holds its margin. This is the same true-margin problem we walk through for spot freight in where reported margin and true margin diverge on spot versus contract, applied here to the customer axis instead of the lane axis. Detention in particular is worth its own recovery discipline, which we cover in recovering detention and demurrage as a broker.
Reported customer margin versus true customer profit
Reported customer margin is the sum of booking-day buy-sell spreads on that account. True customer profit is that sum minus the fuel lag, detention, accessorials, claims, and slow-pay carrying cost that land afterward. In a soft market with fat spreads the two sit close. In a thin-spread market they separate fast, and the account that ranks first on revenue may not rank first, or even in the top five, on the profit that actually funds the business.
Revenue-at-risk and profit-at-risk are two numbers, not one#
Once true profit per customer exists, concentration risk resolves into two figures that a revenue percentage collapses into one.
Revenue-at-risk is the gross billing tied to the account that disappears the day it leaves. It is the number lenders and buyers care about, and it drives the top-line hole you have to refill. Profit-at-risk is the true margin dollars that disappear, which is the number that decides whether the business survives the quarter. The two can point in very different directions. An account that is 25 percent of revenue but 10 percent of true profit is a smaller shock to the bottom line than its headline suggests. An account that is 15 percent of revenue but 30 percent of true profit is a far larger one. You cannot tell which case you are in without attributing the trailing costs.
Separate what disappears from what is stranded
Not every dollar tied to a departing customer is lost at the same speed. Direct load margin goes immediately. Carriers you sourced for that account may follow the freight, or may stay if you can re-lane them. Overhead, rep salaries, and technology are stranded costs that persist for months after the revenue stops. Profit-at-risk is sharpest when you split the immediate loss from the recoverable and the stranded, because the recoverable portion is where retention and re-selling effort actually pays off.
The point of two numbers is that they drive different decisions. Revenue-at-risk tells you how urgently to diversify the book. Profit-at-risk tells you how hard to fight to keep this specific account, and what you can afford to concede to do it. A brokerage that only watches revenue share will over-defend a large, low-profit customer and under-defend a smaller one that quietly funds the desk.
The dependencies that leave with the customer#
Financial concentration is only half of it. The other half is operational, and it never shows up in a revenue ratio.
Start with carriers. Anchor customers often come with a small set of carriers you recruited and tuned specifically for their lanes. Lose the customer and the freight those carriers relied on goes with it, so the relationships thin out just when a tightening market makes capacity hardest to replace. Then the lanes: a big account frequently anchors one leg of a balanced network, and its backhauls are what made adjacent lanes coverable at a workable rate. Pull it and lanes that had nothing to do with that customer get more expensive to cover. Finally the people. A single rep or a small pod often holds the day-to-day relationship, the facility contacts, and the tribal knowledge of how that shipper actually operates. If that account is most of their book, the departure is a retention risk on your own staff, not only a revenue event.
The dependency map is where the surprise lives
Brokerages that get blindsided by a customer loss almost always sized the account on revenue alone and never mapped what depended on it. The recoverable question, which carriers can be re-laned, which lanes go underwater without the backhaul, which reps are over-exposed, is answerable in advance. It is only a surprise if nobody connected the customer to the carriers, lanes, and people attached to it before the notice arrived.
Why the real number is scattered, and why that hides the risk#
The reason most brokerages fall back on revenue share is not laziness, it is that the honest number is genuinely hard to assemble. True profit per customer requires the TMS load record, the accounting system's payments and aging, the accessorial and detention ledgers, and the claims file to agree on a single shipment and roll up per customer. Those systems were bought at different times, speak different formats, and reconcile, if at all, at month-end. By the time the books close, the customer you were worried about is a line in a spreadsheet, not a decision you can still act on.
This is a data problem, not a spreadsheet problem, and it is exactly what a connected operational record solves. When the TMS, accounting, fuel, accessorial, and claims data are unified against one shipment record, true profit per customer updates as each cost lands instead of at quarter-end. You do not have to rip out the TMS to get there, which is the whole point of unifying your systems without a rip and replace: the systems stay, the record on top of them becomes coherent. That is the thesis behind the Mithrilis platform, intelligence from connected data rather than automation of a single workflow. It does not fire the customer or cut the rate for you. It surfaces the concentration pattern no single tool can show, with every figure traceable to the source row it came from, so you can verify a profit-at-risk number before you take it to a QBR or a lender. A concentration figure you cannot trace is just another number to argue about after the account is already gone.
A 60-day concentration stress test#
Sizing the risk is not a one-time audit. It is a standing test you can run whenever an anchor account looks shaky or a renewal approaches. The horizon is 60 days because that is roughly how long the top-line hole takes to bite once the freight stops and the last invoices settle.
Rebuild true profit per customer for the trailing year
Attribute fuel lag, detention, accessorials, claims, and slow-pay carrying cost back to each customer across the last twelve months, off one connected record rather than five disconnected exports. This is the SEE step: show me the profit I am actually earning per account, not the spread I booked.
Rank by profit share and find the gap to revenue rank
Sort customers by true profit share, then place their revenue rank beside it. The accounts where the two ranks diverge most are your real concentration story. A customer that is number one on revenue and number six on profit is a very different exposure than one that leads both.
Map the dependencies behind the top accounts
For each top-profit account, list the carriers sourced mainly for it, the lanes its backhauls make coverable, and the reps whose book it dominates. This is the benchmark step: compare each account's dependency footprint against the rest of your own network so you can see which losses cascade and which stay contained.
Model the 60-day hole and pre-stage the response
Project the cash and margin gap if the account leaves, split into immediate, recoverable, and stranded. Decide in advance what you would reprice, which carriers you would re-lane, and how you would rebalance the exposed reps. Humans make those money-moving and relationship calls; the connected data just makes sure they are working from the real number.
Run continuously, this is also a WATCH function. A connected record can flag an anchor account drifting before the loss notice arrives: tender volume trending down, quotes getting rejected more often, pay dragging from 45 days toward 60. Those are the early signals that a concentrated customer is loosening, and they are worth catching while there is still time to act, the same early-signal logic that keeps a single late load from becoming a chain of missed appointments in stopping the exception cascade.
Same risk, different shape for carriers#
The concentration trap is not a broker-only problem, and asset carriers should not read it as one. For a freight broker, the exposure runs through cover risk and thin spreads: lose the anchor shipper and you lose the freight that justified your carrier relationships and balanced your lanes, while the trailing accessorials decide how much profit actually walked out. The broker's edge is knowing, before a renewal, what each customer returned on a true-profit basis, not what the TMS spread claimed.
For an asset carrier, the same concentration shows up as dedicated lanes and anchor-customer commitments, and it bites harder. The carrier owns the trucks and employs the drivers, so equipment and roster get committed to a specific customer's freight and domiciled for its lanes. Lose that customer and the trucks do not disappear, the fixed cost does not disappear, and the drivers hired for those runs need freight the carrier may not have. Margin sensitivity is higher when the assets are yours, which makes true revenue per customer, linehaul minus the detention absorbed, the deadhead created, and the fuel the surcharge did not recover, the number that tells a carrier which anchor accounts are worth defending and which dependence to unwind while it still has the choice.
See your true profit-at-risk on your own book#
Customer concentration is real, it is usually mismeasured, and the fix is not a better spreadsheet. It is a connected record that turns revenue share into true profit per customer, revenue-at-risk into a paired profit-at-risk number, and a vague worry about the biggest account into a 60-day stress test you can run on demand.
The fastest way to see it is on your own numbers. Mithrilis connects the systems the risk hides in, your TMS, accounting, fuel, accessorial, and claims data, into one shipment record, reconciles them continuously instead of monthly, and keeps every adjustment traceable to its source. From there, Atlas answers concentration questions in plain English, ranks customers by true profit rather than billings, and flags an anchor account drifting while there is still time to respond. Every answer shows its work, because you should be able to verify every result, a principle we wrote into our manifesto. Request a demo and we will show you which customer your business actually depends on, and by how much.
Related Mithrilis capabilities
The Mithrilis platform
How connected data becomes verifiable true-profit intelligence per customer.
For freight brokers
True profit per customer, per lane, and per carrier instead of a revenue share.
For asset carriers
True revenue per anchor customer across dispatch, fuel, and detention.
Spot-load margin leakage
Where reported margin and true margin diverge on freight.
Frequently asked questions
The common method, dividing the top customer's billings by total billings, measures the wrong thing. A sharper measure is the top account's share of true profit after fuel lag, detention, accessorials, claims, and slow-pay carrying cost, plus the carriers, lanes, and staff relationships that leave with it. That requires attributing trailing costs per customer off a connected record, then reporting both a revenue-at-risk and a profit-at-risk figure rather than a single revenue percentage.
Because revenue and profit do not sit in the same place on a freight book. Two customers with identical billings can contribute very differently once detention, unbilled accessorials, and slow-pay are attributed to each. With broker gross margins often near 10 percent, roughly $189 on a $1,912 load in one FreightWaves analysis, a single unbilled charge can flip a customer's contribution, so a revenue ranking hides which accounts actually fund the business.
Revenue-at-risk is the gross billing that disappears the day the customer leaves, which drives the top-line hole and matters to lenders and buyers. Profit-at-risk is the true margin dollars that disappear, which decides whether the business survives the quarter. An account can be a large share of revenue but a small share of profit, or the reverse, so the two numbers drive different decisions about how hard to defend the account and how urgently to diversify.
Carriers you recruited and tuned for that customer's lanes, whose freight goes with the account. Lanes whose backhauls that customer anchored, which get more expensive to cover once it is gone. And the reps or pod who hold the relationship and facility knowledge, who become a retention risk if the departing account was most of their book. None of these show up in a revenue ratio, which is why the dependency map has to be built separately.
Run a 60-day test: rebuild true profit per customer for the trailing year off a connected record, rank customers by profit share and compare that to revenue rank, map the carriers, lanes, and reps each top account depends on, then model the cash and margin gap if the account leaves, split into immediate, recoverable, and stranded. Run continuously it also becomes an early-warning system, flagging tender volume, quote rejection, and payment drift before a loss notice arrives.
Yes, and often more sharply than brokers. Asset carriers concentrate through dedicated lanes and anchor shippers, and because they own the trucks and employ the drivers, equipment and roster get committed to a specific customer's freight and domiciled for its lanes. Losing that customer strands fixed cost that does not disappear with the revenue, so true revenue per customer, linehaul minus absorbed detention, deadhead, and unrecovered fuel, is the number that tells a carrier which anchor accounts to defend.
Topics
Keep reading
Beyond On-Time Percentage: What a Real Carrier Scorecard Measures
Read→On-time percentage grades the past. A real carrier scorecard weighs tender acceptance, dwell, claims, POD timeliness, and performance drift, per lane and per facility.
The Quotes You Shouldn't Be Winning: Margin-Aware Pricing for Freight Brokers
Read→Win rate and quote speed are vanity metrics. Margin-aware quoting prices each load against what the lane, customer, and carrier actually returned on a true-margin basis, and walks the quotes you would regret winning.
The Operational Ontology: Why the Data Model Is the Real Moat in AI-Native Freight
Read→Freight AI in 2026 sells faster execution on single-system data. The durable advantage is the operational ontology: one verifiable model of customer, carrier, load, and facility across every connected system.