There is no shortage of material explaining freight class. Eighteen classes, 50 through 500, driven by density, stowability, handling and liability. Measure the pallet, divide by 1,728, divide the weight by the cubic feet, look up the band.
That’s all correct, and it isn’t where the money goes.
The money goes on classes nobody calculated. They were inherited — set once by somebody who is no longer there, for a product that has since changed, and carried forward on every BOL since because it’s what was on the last one.
The arithmetic is the easy part
Density is length times width times height, in inches, divided by 1,728 to get cubic feet, with the shipment weight divided by that. Include the pallet. Include anything hanging over the edge of it.
Do that honestly and you get a number, and for a density-based item that number maps to a class. Fine. (Our calculator does exactly that if you want to check one now.)
The reason this gets written about constantly and still goes wrong constantly is that almost nobody does it per shipment. They use the class on the paperwork, because the paperwork has always said that.
How a class goes stale
The product changed. Packaging got lighter, the item got redesigned, the case pack changed from twelve to eighteen. Density moved. Class didn’t.
The pallet configuration changed. Same product, stacked differently, or double-stacked where it used to ship single. Different cubic feet, different density, different class.
Somebody guessed once. A class was needed on a Friday, an approximation went on the BOL, nothing went wrong for a while, and it became the standard.
The commodity is density-based and nobody noticed. Plenty of NMFC items don’t carry a fixed class at all — the class moves with density, in bands. Shipping the same product at different densities means different classes, correctly. A single stored class is wrong most of the time by construction.
An FAK exists but isn’t applied. Freight-all-kinds arrangements let mixed freight rate at a negotiated class. If the agreement has one and the BOL doesn’t reflect it, you pay the higher standard class and the audit passes, because the invoice matches the paperwork.
Why the broker eats it
The correction lands on whoever is between the shipper and the carrier.
The carrier reweighs and reclasses at the terminal, rebills the difference, and their evidence is a scale and a dimensioner. The shipper gave you a class in good faith and considers the matter closed. You quoted your customer off that class and already invoiced them.
Going back for more money on a load that delivered fine is a conversation most brokers have once. Then they stop having it and absorb the difference instead, which is how a slow margin leak becomes permanent.
What actually prevents it
Calculate from dimensions, don’t accept a class. If you capture length, width, height and weight, you can derive density and suggest a class. That flips the default from “trust the paperwork” to “verify, then quote.”
Flag disagreement rather than silently overriding. If the shipper says 70 and the density says 125, that’s a conversation before pickup, not a surprise afterwards. Sometimes they’re right — the item has a fixed NMFC class that isn’t density-based. Sometimes they’re wrong. Either way you want to know before you quote.
Check the big repeaters first. You do not need to re-verify everything. In most books a handful of commodities make up most of the volume. Verifying five of them catches nearly all of the exposure.
Apply the FAK you already have. If you negotiated one, make sure it’s on the paperwork. This is free money left behind more often than it should be.
What to do about a class you know is wrong
If verification says the long-standing class is wrong, the conversation with the shipper is easier than it feels — provided you bring the measurement rather than the opinion.
Dimensions, weight, the calculated density, the band it lands in. That’s a fact about their freight, not an accusation. Most shippers would rather find out from you than from a rebill, because the correction was coming either way and it lands on their invoice in the end.
The ones who insist on the old class anyway have told you something useful about how the rest of the account will go.
The wider point
Freight class is the clearest example of a pattern that runs through LTL: the expensive errors aren’t the ones where somebody did the arithmetic wrong. They’re the ones where nobody did the arithmetic at all, because a number was already sitting there and it looked authoritative.
An invoice audit won’t catch it — the invoice matches the BOL, so the check passes. It only surfaces when the carrier measures, and by then it’s a rebill.
The short version: the class on the paperwork is a claim, usually inherited, frequently stale. Deriving it from dimensions at quote time turns a rebill six weeks out into a question you asked before pickup.
That’s one of the quote-time checks behind LTL for truckload brokers, and the same reason auditing in transit beats auditing the invoice — the pallet still exists while it’s moving.
So here’s the question worth sitting with: on your highest-volume commodity, when did anyone last actually measure it?