What the unit actually cost, once the freight is in it
Landed cost is not one calculation. It is one addition and one allocation, and the allocation is where the same $1,800 of shipment costs produces three different unit costs for one SKU. Each is arithmetically correct, and none of them is right for the whole $1,800 — because that $1,800 is five costs that do not move with the same thing, and neither does the number you divide it by. The two SKUs below are invented so the arithmetic can be checked.
- Landed cost per unit = the goods, plus every cost of getting them here, divided across the units. The hard part is the dividing, not the adding.
- Split the same $1,800 of shipment costs three ways — by unit, by value, by volume — and one SKU lands at $2.36, $2.72 or $3.50 a unit. That is a 48% spread with no error anywhere in it.
- None of the three is the answer, because the $1,800 is not one cost. It is five: $1,000 of freight and drayage, $200 of fixed entry and brokerage, $250 of duty, $50 of insurance and $300 of per-piece prep. Send each of the five where it belongs, ruling once on the fixed fee that has no driver of its own, and the same SKU lands at $2.75.
- Match each cost to what actually drives it: volume or chargeable weight for ocean freight, the basis actually used to price the move for drayage and local delivery, value for ad-valorem duty and for insurance priced on insured value, unit count for per-piece handling.
- The denominator splits the same way. The fixed part — in this worked example, the $1,200 of freight, drayage and per-entry fees — has to go over the units you received: $1.20 a unit, not the $1.00 you get from units ordered. The remaining $600 — duty, insurance and per-piece prep — mostly follows the goods that actually arrived instead, though insurance is the one to check: a premium with a minimum may not move at all, and prep falls only if it is charged on units actually handled/received.
Ask what a SKU costs and the fastest answer is the supplier’s price. It is the number on the invoice, it is the number in the cost field, and it is the number the margin calculations run on.
It is also, for anyone importing, meaningfully wrong. The goods did not teleport. Somebody paid for a container, a customs entry, a broker, duty, drayage and a pallet of stretch wrap, and all of that is as much a cost of having the unit on your shelf as the unit itself.
Landed cost is what you get when you put those back in. The addition is easy. What follows is the part that is not.
Step 1 — Collect everything that belongs to the shipment
One scoping note first. This is operational landed cost: the per-unit figure you buy, price and reorder on. How much of it is capitalised into inventory in your accounts is a separate question, governed by your accounting policy and settled with your accountant.
Where to draw the line: start with costs directly required to acquire and bring the goods to their intended location, then apply your accounting policy consistently.
That usually sweeps in more than people expect:
- The goods — the supplier invoice, net of any credits.
- Freight — ocean or air, plus the trucking at both ends. Drayage from port to warehouse is freight, even though it arrives on a separate invoice weeks later.
- Duty and customs charges — whatever your goods attract, plus the entry fee itself.
- Broker and handling fees — customs brokerage, documentation, port charges. Demurrage, if you incurred it, is economically a cost of that shipment; whether it is capitalised into inventory or expensed depends on your accounting policy.
- Insurance on the shipment.
- Inbound prep — labelling, polybagging, repacking, palletising, whether done by you or billed by a third party.
Two costs sit on the boundary and are worth ruling on once. Storage at your own warehouse after receipt is not landed cost — the goods have landed; that is a holding cost. Payment-processing or financing charges on the purchase are usually treated separately too, because they scale with how you paid rather than with the shipment.
The practical failure here is not disagreement about the boundary. It is that some of these invoices arrive after the goods do, so the shipment gets costed on the freight quote and never revisited when the broker’s bill lands weeks later.
The landed-cost error that is hard to catch is not arithmetic. It is an invoice that arrived after somebody had already decided the shipment was finished.
Step 2 — Allocate, and know what each method distorts
Here is one shipment with two very different SKUs. Goods total $10,000; the shipment costs on top of them — freight, duty, brokerage, insurance and prep — total $1,800:
| SKU | Units | Unit price | Line value | Volume each | Total volume |
|---|---|---|---|---|---|
| A — small accessory | 1,000 | $2.00 | $2,000 | 0.002 m³ | 2.0 m³ |
| B — bulky main item | 200 | $40.00 | $8,000 | 0.015 m³ | 3.0 m³ |
| Total | 1,200 | — | $10,000 | — | 5.0 m³ |
Open that $1,800 up before dividing it. It is not one cost. It is five, and they do not move with the same thing:
| Shipment cost | Amount | What actually drives it |
|---|---|---|
| Ocean freight and drayage | $1,000 | Ocean freight: volume or chargeable weight; drayage/local delivery: allocate by the basis actually used to price the move. |
| Customs entry, documentation and fixed brokerage | $200 | One shipment, one entry |
| Duty charged on declared value | $250 | What the goods are worth |
| Insurance | $50 | Insured value, subject to minimums |
| Prep and labelling at $0.25 a unit | $300 | Pieces someone has to touch |
| Total | $1,800 | — |
That composition decides both halves of what follows: how the $1,800 splits between the two SKUs, and, in step 3, what it gets divided by. Ignore it, run one basis over the whole $1,800, and here is what happens.
Three bases are compared against that $1,800 below. These are three possible allocation bases, not three equally correct ways to allocate freight. The right basis depends on what created the cost.
Method 1 — equal per unit
Divide the $1,800 by the 1,200 units: $1.50 per unit, applied to everything.
| SKU | Goods | Allocated | Landed per unit | Increase |
|---|---|---|---|---|
| A | $2.00 | $1.50 | $3.50 | +75.0% |
| B | $40.00 | $1.50 | $41.50 | +3.8% |
Simple, and brutal on cheap items. A $2 accessory absorbing $1.50 of shipment cost has had its cost raised by three quarters, while the $40 item barely notices. Whether that is a distortion or the truth depends on whether unit count is what drove the $1,800 — and on the table above, unit count drove $300 of it, the prep, and none of the other $1,500.
Method 2 — by value
A is 20% of the goods value, B is 80%. So A takes $360 and B takes $1,440.
| SKU | Share | Allocated | Per unit | Landed per unit | Increase |
|---|---|---|---|---|---|
| A | 20% | $360 | $0.36 | $2.36 | +18.0% |
| B | 80% | $1,440 | $7.20 | $47.20 | +18.0% |
When the pool is spread strictly in proportion to the same goods value you are measuring the increase against, value allocation raises every SKU by the same percentage — 18% on both lines here — which is why it looks so fair. It fits costs that do scale with value, such as duty charged as a percentage of declared value, and insurance to the extent the premium tracks insured value. That is $300 of this shipment’s $1,800. It does not fit a container, which does not care what is inside it.
Method 3 — by volume
A occupies 2.0 of the 5.0 m³, B occupies 3.0. So A takes 40% and B takes 60%.
| SKU | Share | Allocated | Per unit | Landed per unit | Increase |
|---|---|---|---|---|---|
| A | 40% | $720 | $0.72 | $2.72 | +36.0% |
| B | 60% | $1,080 | $5.40 | $45.40 | +13.5% |
Volume fits the largest single line in the pool, the $1,000 of freight and drayage — the ocean freight by nature, the drayage only if the move was priced that way. Here it is being applied to all $1,800 of it, including the $300 of prep that a cubic metre has nothing to do with.
The three side by side
| Method | SKU A landed | SKU B landed | What it assumes drove the cost |
|---|---|---|---|
| Equal per unit | $3.50 | $41.50 | Handling each piece |
| By value | $2.36 | $47.20 | What the goods are worth |
| By volume | $2.72 | $45.40 | Space on the boat |
SKU A ranges from $2.36 to $3.50 — a 48% spread — and not one of those numbers is a mistake. If A sells for $4.99, the choice of method is the difference between a 53% gross margin and a 30% one. On that SKU, the allocation method is the profitability analysis.
And not one of the three is what this shipment actually cost, because each of them runs a single basis over a pool of five costs that do not share one. The rule below produces a fourth figure, and it is the only one with a reason underneath it.
So which one
Match each cost to its own driver rather than picking one method for everything:
| Cost | Allocate by | Because |
|---|---|---|
| Ocean or air freight | Volume or weight | You are buying space, not value |
| Drayage and local delivery | The basis actually used to price the move | It may be priced by weight, container, trip, pallet count or a flat charge |
| Duty charged on value | Value | It is literally a percentage of value |
| Brokerage charged as a percentage | Value | The fee itself scales with declared value |
| Insurance | Often allocated by insured value | Premiums can also carry minimums and fixed charges |
| Customs entry and other fixed per-entry fees | A convention you choose and keep | Fixed per shipment, so no driver picks itself |
| Prep, labelling, polybagging | Unit count | Someone touches each piece |
The same shipment, each cost by its own driver
Run that rule over the five components and almost nothing is left to taste. Freight — and, in this example, the drayage priced with it — follows the 40/60 volume split, duty and insurance follow the 20/80 value split, prep follows the pieces at $0.25 each. Only the $200 of fixed entry and brokerage needs a ruling, because nothing about a per-entry fee picks a driver for you — declared value is the convention used here, chosen once and kept. Equal-per-line or equal-per-unit can also be defensible for truly fixed fees; consistency matters more than pretending the fee had a natural driver.
| Component | Allocated by | SKU A | SKU B |
|---|---|---|---|
| Freight and drayage — $1,000 | Volume, 40/60 | $400 | $600 |
| Entry and fixed brokerage — $200 | Convention: declared value | $40 | $160 |
| Duty — $250 | Declared value, 20/80 | $50 | $200 |
| Insurance — $50 | Insured value, 20/80 | $10 | $40 |
| Prep — $300 | Unit count at $0.25 | $250 | $50 |
| Total allocated | — | $750 | $1,050 |
| Per unit | — | $0.75 | $5.25 |
| Landed per unit | — | $2.75 | $45.25 |
A lands at $2.75, B at $45.25, and the $1,800 is still fully allocated. That $2.75 is not the $2.36 the value basis gave, not the $3.50 the unit basis gave, and not the $2.72 from volume either. It sits near the volume answer because freight and drayage are the largest line in the pool, and it misses it because they are not the only line. At $4.99 that is a 45% gross margin, between the 53% and the 30% above.
What it costs to produce is one extra column on the shipment’s cost sheet, naming which bucket each invoice fell into. The arithmetic is the same arithmetic as the three methods above; what is different is that the pool was broken up before any of it was applied.
And then the honest caveat, because precision has a maintenance cost. If your shipments are made of items broadly similar in size and value, all four figures land close together and equal-per-unit is a perfectly good answer that costs almost nothing to keep up. The spread above only opened up because the shipment mixed a $2 accessory with a $40 item. The more your catalogue mixes cheap-and-small with expensive-and-bulky, the more the method matters — and if it never mixes, stop worrying about it.
In SKU Compass, the purchase order’s landed-cost section takes two figures: a Total Fees (split across units) amount and a Per-Unit Added Fee. Enter a total and the section prints its own working — “Total fee ÷ units = per unit” — which is method 1 above.
Step 3 — Divide the fixed part by what arrived, not by what you ordered
This is the step that gets skipped, and it connects directly to short shipments.
Freight is not refundable by the unit — and that is a statement about freight, not about the $1,800. The composition splits the denominator exactly as it split the numerator. Of the $1,800, $1,200 is ocean freight, drayage, customs entry and fixed brokerage. In this worked example, treat the $1,200 as fixed shipment-level charges that did not fall when 200 units failed to ship. It is still $1,200, now spread across 1,000 units. On your own shipments, check the invoice first: drayage in particular can change with actual handling, route, pallet count or detention.
| Basis | Units | The fixed $1,200 spread across them | Effect |
|---|---|---|---|
| Units ordered | 1,200 | $1.00 / unit | Understates the fixed part by $0.20 |
| Units received | 1,000 | $1.20 / unit | What the fixed part actually cost |
$1.00 against a true $1.20 is the fixed part understated by a sixth — 16.7% — on every unit that did arrive, and it stays that way unless somebody re-costs the shipment after it closes. Note that it is measured against the $1.20, not against the $1.00: reading it the other way around turns the same error into a flattering 20%.
Most of the other $600 does not behave that way — and the part that might is worth knowing about rather than assuming. That is why one denominator cannot be run over the whole pool:
| The rest of the pool | As ordered | What happens when 200 units never ship |
|---|---|---|
| Duty charged on declared value | $250 | Follows what was actually imported and declared, so it may be lower |
| Insurance | $50 | May follow insured value down — or not, where the premium carries minimums or fixed charges |
| Prep at $0.25 a unit | $300 | If prep is charged only on units actually handled/received, it falls to $250 |
Goods cost sits outside that pool and behaves like the $600 rather than the $1,200: the economic cost should ultimately follow what was actually supplied and owed, including any credits for short shipment. So match the denominator to the cost. Fixed shipment costs such as ocean freight and per-entry fees do need spreading over the units actually received. One denominator does not mechanically apply to every landed-cost component, and the re-spreading above is a correction to the fixed part only.
In SKU Compass, that basis is an explicit choice: the landed-cost table offers Received and Ordered, with Received selected by default, and its columns run Base COG, + Fees and New COG against the Current cost.
Then put the number where decisions are made
A landed cost that lives in a spreadsheet nobody opens has changed nothing. The point of the exercise is that the corrected figure becomes the cost your margin reports, your reorder economics and your pricing decisions actually use.
Which raises the one thing to be careful about: a landed cost is true of a shipment, and the cost field it feeds is usually true of a SKU. If freight rates have moved since the last container, pushing the new figure forward changes the cost basis of stock that arrived under the old rate. Decide whether the field is meant to represent historical landed cost or current replacement cost; do not let one silently substitute for the other. Either can be the right answer, but it is a decision, and it should be made deliberately rather than discovered later in a margin report nobody can reconcile.
Questions people ask
What is landed cost per unit?
The total cost of getting one sellable unit into your warehouse: the supplier’s price plus that unit’s share of freight, duty, brokerage, insurance and inbound prep. It is the number that tells you what a unit really cost you, as opposed to what the supplier charged for it.
What costs should be included in landed cost?
Start with costs directly required to acquire and bring the goods to their intended location, then apply your accounting policy consistently: goods, freight at both ends, duty and customs entry, brokerage and port charges, insurance, and inbound prep such as labelling or repacking. Storage after the goods have landed is a holding cost rather than a landed cost, and financing charges are usually kept separate because they scale with how you paid.
How do I allocate freight across different products?
Break the shipment’s costs into components first, because the drivers differ inside that one total, then allocate each component by whatever drives it. Ocean freight follows volume or chargeable weight, because you are buying space; drayage and local delivery follow the basis actually used to price the move. Duty charged on value follows value, and insurance is often allocated by insured value. Per-piece prep follows unit count. Splitting one lump sum evenly across units is quick and fine when your items are similar in size and value, but it overcharges small cheap items badly when a shipment mixes them with large expensive ones.
Should landed cost be based on units ordered or units received?
Received, for the part of the pool that is genuinely fixed. Charges that are fixed for the shipment do not shrink when it arrives short, so dividing them by the units you ordered spreads them over units that never arrived and understates every unit that did. Check the invoice before treating a charge as fixed: drayage in particular can change with actual handling, route, pallet count or detention. In the worked example, treating $1,200 as fixed shipment cost with 1,000 of 1,200 units delivered, the honest figure is $1.20 a unit, not $1.00. The rest of the pool is different: duty assessed on what was actually imported and declared may be lower, per-piece prep falls with the pieces if it is charged only on units actually handled or received, and insurance may follow the insured value down or may not, where the premium carries a minimum. Check that one rather than assume it.
Is landed cost the same as cost of goods sold?
They are closely related but not identical. Landed cost is what a unit cost to get into your warehouse; cost of goods sold is what you recognise when it sells, under whatever costing method you use. Landed cost is normally the input to that, which is why getting it wrong quietly moves your reported margin.
How often should I recalculate landed cost?
Calculate provisionally at receipt, then finalise and true it up when the late invoices arrive. The common failure is costing a shipment on the freight quote and never revisiting it when the broker’s bill arrives weeks later. Set a point in your process where a shipment is finally costed, and do not let it be the day the goods arrive.
