Measure first, and measure the freight, not the carton
Length × width × height in inches, divided by 1728, gives cubic feet. Use the outer envelope: the pallet deck, the wrap, the corner that leans past the stringer. A pallet measured at 48 × 40 × 45 is 50.00 ft³. The same pallet with four inches of unlisted height is 54.44 ft³, and the density you report drops by roughly eight percent without a single pound changing.
Weight means the weight that gets scaled, pallet included. Divide pounds by cubic feet, carry the number to two decimals, and stop rounding early — 14.98 and 15.02 sit on opposite sides of a band edge, and that is the entire dispute.
Eight pounds per cubic foot is the first line most pallets sit near: a pallet at 8.2 pcf bills at 100 and the same pallet at 7.9 pcf bills at 125, and the only thing that moved is a couple of inches of stretch wrap nobody wrote down.
Ten is the second: 10.1 pcf bills at 92.5 and 9.9 pcf bills at 100, which is a difference the carrier tape can find on its own.
Twelve separates 92.5 from 85, and it is the line where counting the pallet deck in the height instead of in the length is enough by itself.
Fifteen is the expensive one — 14.98 bills at 85 and 15.02 bills at 70 — which is why the arithmetic gets carried to two decimals before anyone rounds.
The complete thirteen-band density chart, 400 down to 50, is on the freight class chart page.
Then check the item. Some commodities carry density sub-bands inside their own NMFC entry. Some are classed by what they are, not by how dense they are. A perfect density calculation on an item that is classed by name is a wrong answer delivered confidently.
Where the money actually moves
Almost nobody loses money at the calculator. They lose it at the dock, when the carrier reweighs or reclasses and the invoice arrives with a number that is not the number on the BOL. Two camps argue about this constantly, and they are not arguing about the same thing.
The first camp treats it as an evidence and ownership problem: "Reclasses and reweighs get most of the attention, but I've seen liftgate, limited-access, appointment, oversize…" — the point being that the class fight is one line item among several, and all of them fail the same way. Put more precisely: "Reclasses and excessive carrier charges are different—they become an evidence-and-ownership problem after …"
The second camp treats it as extraction: "All the reclasses lately and some carriers excessively way overcharging is ge…" The question underneath it is blunt: "So it's arbitrary based on mood and how much more revenue they need?"
Neither camp has persuaded the other. And between them sits a question that got asked and never got answered: "so a time stamped PDF of the density and class wouldn't help with the audits and reclassifications?" The only reply it drew was a shrug — "No demand, stop trying to supply."
The answer that both camps skip
A timestamped PDF of your density and class does help, and it is not enough on its own. That is the part the argument keeps sliding past, because both sides want a single object to be either useless or decisive, and it is neither.
What a timestamped record actually proves is what you declared, at a moment, from inputs you chose. It does not prove what the carrier measured. Those are two different claims, and an audit compares them. So the record only becomes evidence when it carries the inputs, not just the outputs: the three dimensions to the half inch, the weight and where that weight came from, who took the measurement, and the minute it was taken. A PDF that says "Class 85" and nothing else is a conclusion with no supporting work, and it will lose to a reweigh ticket every time.
The second half is harder and cheaper: the record has to travel with the shipment. If the numbers live in your inbox and the carrier's numbers live on their scale ticket, the dispute is not about freight, it is about which document is authoritative — and you have already conceded that by not putting yours on the BOL. Send the density sheet with the load, reference it on the BOL, and the argument narrows to a real question: is the gap between the two cubes explained by a measurement choice — pallet, overhang, wrap — or is it a different number entirely?
That distinction is what you are actually buying. Not vindication. A narrower argument.
The gaps that turn into class disputes
Multi-pallet shipments are where the average lies. One pcf for the whole load hides the pallet that drags it: five pallets at 12 pcf and one at 5 pcf is not a 10.8 pcf shipment, it is a shipment with one pallet that will get reclassed. Compute per pallet, then decide whether to bill the average or split the line.
Height is the quiet one. Two inches of unlisted pallet height on a 45-inch load changes the cube by about four percent, which is enough to move a band when you are sitting on an edge.
And keep the accessorials separate from the class argument. Liftgate, limited access, appointment windows, oversize — those are their own line items with their own triggers, and bundling them into a "they overcharged me" complaint is how a winnable reclass gets thrown out with the rest. The same goes for the cash side: "The lumper charges are out of control, some of those warehouses basically hold your freight hostage until you pay a few hundred in cash" — real, expensive, and a completely different fight from density.
What to keep
Measure the outer envelope. Divide by 1728. Divide weight by cube. Two decimals. Check the item. Write down the inputs and the minute, put the sheet on the BOL, and keep the accessorials in their own column. The class you report is a claim; the record is what makes the claim arguable.