Image: Truck Hauling Continuous tracks @ Canoe Creek Service Plaza by DanTD, licensed under BY 4.0.
Your audit compared the fuel line on last week's invoice to the fuel line on the rate confirmation, found no variance, and released the payment. What it actually confirmed is that two documents agree, which is not the same as the charge being right.
Fuel surcharge arrives as one charge code. Behind that code sit four different calculations that happen to share a name. A pegged per-mile formula, a banded matrix, a percentage of line haul, and a deferral to whatever the carrier published that week all bill identically, and an engine that treats them as one thing will wave three of them through untested.
Four structures, one charge code
The first structure is arithmetic you can reproduce. Take the diesel index, subtract the contract peg, divide by an assumed miles per gallon, multiply by billable miles. FreightWaves' explainer works it through at $4.15 diesel against a $1.25 base at 6.5 MPG, which yields roughly $0.4462 per mile, or $535.44 across 1,200 miles. Three of the four inputs there live in the contract and appear nowhere on the invoice.
The second is a step function. Carriers publish mileage tariffs that map diesel price bands to a per-mile amount, so the charge does not move continuously with the index. It jumps at a band edge. Two readings a cent apart can fall either side of a boundary and bill at different rates, and nothing on the invoice shows which band the carrier used.
The third abandons miles entirely. LTL and parcel surcharges are a percentage applied to the line haul charge. TForce Freight's published class-based LTL surcharge was 57.1% effective 7 October 2026, applied to line haul charges as provided in item 1 of its rules tariff, with density-based shipments routed to a separate item altogether. One carrier, one week, two schedules.
The fourth is not a calculation at all. The rate confirmation says fuel is per contract, or per the carrier's published schedule, and leaves the figure blank. The number materialises on the invoice weeks later. There is nothing to match against, so matching reports a pass.
Which week's index did the carrier use
Every structure above except the last one takes a reading from the same place, and that reading is a weekly snapshot rather than a market price. The EIA collects retail diesel prices every Monday from a sample of 590 outlets, excluding Alaska and Hawaii, and the figures reflect the pump price as of 8:00 a.m. local time on Monday. Publication follows at around 10:00 a.m. ET on Tuesday, or Wednesday when a government holiday intervenes.
So a surcharge keyed to the index is keyed to a specific Monday. Which Monday is a contract term, and carriers do not all pick the same one. OilPriceAPI's tracker lists Old Dominion at a 52.3% surcharge effective 14 October 2026 against a DOE index of $6.199 per gallon. That $6.199 is the reading for the week ending 5 October. A fresher number existed, $6.179 for the week ending 12 October, published on the Tuesday before that surcharge took effect, and it played no part in setting it.
A week of drift is not a rounding error when the index is moving. EIA's national average ran $6.529 for the week ending 21 September, then $6.382, then $6.199, then $6.179. Hold the peg at $1.25 and the assumption at 6.0 MPG, and the first and last of those readings differ by about 5.8 cents per mile. On a 1,200 mile lane that is $70 on a single load, from nothing but a disagreement over which Monday governed.
The MPG assumption moves more money than that. Truckload schedules commonly sit between 5.5 and 6.5 MPG, and at the week ending 12 October reading those two figures produce per-mile surcharges roughly 13.8 cents apart, or about $165 over the same 1,200 miles. Neither number is wrong. Only one of them is the one you negotiated.
A percentage surcharge inherits every error above it
Per-mile structures fail on their own inputs. Percentage structures fail on someone else's. When the surcharge is a percentage of line haul, the base it multiplies has already been through a base tariff, a discount, and whatever classification the carrier applied, and an error in any of those is quietly amplified by the fuel line rather than isolated to the charge that caused it.
Take a $4,000 gross line haul under a 70% discount. Apply 52.3% to the $1,200 net and the surcharge is $627.60. Apply it before the discount and it is $2,092. Both calculations use the correct percentage. Only the first follows the tariff, and the invoice total is the only place the difference shows up.
The spread between carriers is real too. TForce published 57.1% for the week beginning 7 October while the Old Dominion figure for the following week was 52.3%, against the same national index. That gap is a schedule decision, not a fuel market, so benchmarking one carrier's percentage against another's proves nothing about either.
Why a deferred fuel clause defeats matching
Naive audit logic needs two numbers. Give it one and it has no opinion. A rate confirmation that defers fuel to a contract matrix supplies a pointer, and the matcher resolves pointers by passing the line.
Resolving it properly means holding four things the rate confirmation does not carry: the schedule the contract names, the index week that governs the ship date, the mileage basis the tariff specifies, and the band or formula that the resolved index reading selects. Miss any one and the recomputed figure is wrong in a way that looks like carrier error.
Mileage basis deserves particular attention. A surcharge priced against practical miles, loaded miles, or a household goods guide mileage will produce three different totals on the same physical trip, and the difference compounds with every cent the index sits above the peg.
| Step | Rate match | Recomputation |
|---|---|---|
| What it compares | Invoice line to rate confirmation line | Invoice line to the figure the contract produces |
| Peg | Not read | Pulled from the contract and applied |
| MPG | Not read | The contracted assumption, not the carrier's |
| Index week | Not read | Resolved from ship date against the schedule's effective dates |
| Mileage basis | Not read | Practical, loaded or guide miles as the tariff specifies |
| Percentage base | Not read | Net line haul after discount, accessorials excluded |
| Deferred clause | Passes by default | Resolved, then computed |
| What it catches | A typo | An input the carrier substituted |
How Penny prices the line before the carrier does
Penny stores the fuel clause rather than the last amount billed. For each load she identifies which of the four structures the contract uses, resolves the index week the ship date falls under, applies the peg and MPG the contract fixes, and derives the surcharge independently before comparing it to the invoice.
Variances then arrive attached to a term. The index week was off by one. The MPG assumption was 6.5 in the contract and 5.5 on the bill. The percentage was applied to gross line haul instead of net. A carrier can settle a dispute that names the substituted input in one email, which is the same argument our post on index-linked surcharges makes about ocean bunker and emissions charges.
When the contract fixes none of those terms, the finding belongs upstream rather than in a dispute queue. Miles can pin a peg and an MPG assumption at renewal for far less than Penny will ever recover arguing them after the fact.
Where to start
Pull ten invoices from one carrier and one month, then write down four things for each: the structure the contract names, the index week it specifies, the peg and MPG or the percentage base, and the mileage basis. Recompute all ten from those terms alone.
If the contract does not supply all four for any of them, you have your answer without doing the arithmetic. ATRI's 2026 report puts the 2025 industry average cost of operating a truck at $2.336 per mile, against $1.854 per mile once fuel is stripped out. A charge that size should not be the one your audit takes on trust.
Book a walkthrough to see Penny recompute fuel surcharge from your own contracts and the published diesel index behind them.