Every Freight File Tells You Its Margin Too Late to Do Anything

A forwarding desk quotes a shipment, moves it, closes the file and reports a margin. The number is produced after everything that could change it has already happened. It is accurate, it is auditable, and it is useless — because the point at which it becomes known is the point at which nothing can be done about it.

Four things move that number between the quote and the close, and every one of them is knowable earlier than it is known. Chargeable weight computed on a different basis from the one the carrier will use. Duty, excise and VAT assessed on the wrong base. Master-to-house cost apportionment by a rule that does not sum back to the master exactly, so the difference lands wherever the rounding falls. And a supplier invoice above the quoted cost, discovered when it arrives rather than when it exceeds a tolerance. A file costed continuously — with an exception raised the moment any of the four moves — converts all of this from reporting into management.

Each of the four is worked through below, in the order they arrive: the one mistaken for arithmetic, the one that belongs to somebody else's accounting, the one that hides for years, and the one that turns up in accounts payable. In every case the question is the same — what could have been known earlier, and by whom.

Chargeable weight is a contract term, not a fact

This is the first of the four, and the one most often treated as arithmetic that cannot be wrong. It is also the reason a forwarder who stores what he ships and delivers what he stores ends up reconciling three systems at month end rather than reading one number.

Freight is charged on the greater of actual weight and a volumetric or density-based figure, and the arithmetic differs by mode, by carrier and by contract. Air uses a volumetric calculation with a divisor named in the agreement. Less-than-container sea uses a revenue tonne — the greater of weight and volume on the basis that trade names. Road, courier and express each have their own conventions, and a single carrier may apply different ones to different lanes.

None of that is difficult. What makes it expensive is that the quote and the cost are frequently computed on different bases by different people, and the discrepancy only appears when the carrier's invoice does. A desk quoting from a rate spreadsheet uses whichever rule that spreadsheet was built around; the carrier charges what its contract says. The gap is the margin, and it is invisible until it is spent.

Two disciplines close it. The chargeable weight must be computed by the system from measured dimensions and weights, on the basis the specific lane and carrier contract name, and stored against the shipment rather than recalculated ad hoc. And those dimensions have to be real — which is one of the practical arguments for the handling unit being a record that carries its own weight and dimensions instead of a figure someone typed into a booking form.

Duty on the wrong base

Customs charges are assessed on a value built to a defined rule, and the rule is not "the invoice". Which elements of freight and insurance are included, how the incoterm shifts the boundary, what an exemption or a free-zone movement does, whether excise applies before or alongside the tax calculation — each of these changes the assessed amount, and getting one wrong changes the file's cost without changing anything visible on the paperwork.

This is the seam where forwarding meets the importer's own accounting, and where both sides tend to assume the other has it right. The mechanics belong to their own articles rather than to this one: how duty and import VAT are actually assessed on a UAE import sets out the base, and the importer's own problem of getting those figures into an inventory valuation is a separate one again.

The discipline for a forwarder is narrower and simpler: the declaration is a record with a computed base, and it should not be releasable without the payment reference that proves the charge was settled. A declaration released on trust is a cost that will arrive later attached to a file that has already been reported.

The apportionment that does not sum

This is the one that hides for years, because every individual number looks reasonable.

A consolidation has a master cost — the container, the flight, the trucking of the whole — and several house shipments that must each carry a share. Apportionment is by weight, by volume, by chargeable weight, by value, or by a hybrid. Whichever rule is chosen, one property is not optional: the shares must sum back to the master cost exactly. No rounding leak, no residue absorbed silently into the last house on the list, no shortfall left on the master to be written off at period end.

In practice this is where implementations fail quietly. A percentage apportionment computed to two decimals across seven houses will not sum to the master, and the difference has to go somewhere. If nobody decided where, the code decided, and the file that happens to be processed last is carrying other people's costs. That file's margin is wrong, the others are wrong in the opposite direction, and the total is right — which is precisely why nobody catches it. Everything reconciles at the level anyone checks.

The test is trivial and almost never run: take a consolidation, sum the apportioned costs of the houses, and compare to the master to the fils. Do it for a consolidation with an awkward number of houses and an awkward weight distribution. If the answer is not exact, ask where the residue went, and expect the person answering not to know.

The supplier invoice nobody was waiting for

The last mover is the simplest. A supplier — a carrier, a haulier, a customs broker, an agent at the other end — invoices above what was quoted. In most operations this is discovered in accounts payable, weeks after the file closed, by someone who has no way to tell whether it is a genuine overrun, a misapplied rate or an accessorial that was legitimately incurred.

Two mechanisms fix most of it. Accrue the expected cost when the service is booked, so an unmatched invoice is visible as a difference rather than as a new fact. And raise an exception when an invoice exceeds the accrual by more than a tolerance, while the file is still open and somebody remembers the shipment. A tolerance that is too tight produces noise nobody reads; one that is too loose produces nothing. It should be set per charge type, and reviewed once a quarter against what it actually caught.

A related discipline belongs here: a charge should be priced on the basis its charge code names. A documentation fee is a flat fee. If it is configured so that it scales with tonnage because everything on the file scales with tonnage, it is wrong on every heavy shipment and nobody will ever look, because the total still looks like a freight invoice.

Before any of the four: the quote itself

The four movers above all act on a file that has already been priced. The quote carries two failure modes of its own, and both are worth closing first because they are the cheapest interventions available.

An expired quotation converted into a job is a shipment moving at a rate that no longer exists. The cost side has moved; the revenue side has not. This should be refused outright rather than warned about, because the person converting it is under time pressure by definition.

A lost quote closed with no reason makes the conversion ratio a number with a dishonest denominator. That is worse than not measuring conversion at all: a metric everyone quotes and nobody can defend is how a commercial team ends up optimising something that is not happening. Both refusals sit in Logix for the same reason — they are the cheapest possible interventions, and they both act at the only moment the outcome is still open.

What to do on Monday

You do not need a systems programme to find out how bad this is. Take five closed files from the last quarter, ideally including one consolidation and one air shipment, and do this by hand:

1. Recompute the chargeable weight from the measured dimensions on the basis the carrier contract names. Compare to what was quoted and to what was invoiced. Three numbers; note every case where they differ.
2. Sum the apportioned costs of a consolidation and compare to the master. Exactly.
3. Line up supplier invoices against the quoted costs, and mark every one that came in high. Ask, for each, when it could first have been known.
4. Check the charge codes on one complex file: is every charge computed on the basis its name implies?
5. Then compute the margin as it would have looked at three points — at quote, at departure, and at close.

That fifth step is the whole exercise. The distance between the first number and the last one is what your forwarding desk is currently managing blind, and it is the business case for costing the file while it is open. Once the figure exists, the systems conversation is short, and it is a much better conversation to have than the one where a vendor tells you what the figure usually is.

This is the same lesson the warehouse side of the business learns about its own invoice — that a number reconstructed after the fact is a number that arrived too late to change anything, which is the argument set out in why a 3PL bill has to be computed from the operation's own events. A forwarder who also stores and delivers is running both problems at once, which is exactly the case where one implementation across the three operations stops being a preference and starts being the only way the month-end numbers agree.