Freight
How many units fit in a container, and whether you should fill it
Dividing a container's cubic metres by a carton's is the number every forwarder page gives you, and it is a number you will never load. The useful questions come after it: what the box really takes, and whether the empty space is worth filling.
Marcus VolstedCo-founder, Restocio
Published 10 October 202620 min read
A 40 foot high cube container holds about 76 cubic metres and a little under 29 tonnes of cargo on the carrier's own sheet. How many cartons you actually fit is set by how your carton divides into the inside length, width and height of the box, and that answer sits well under the volume division: in the worked example below it is 600 cartons loaded one way round, against 791 by volume. Whether you should then buy enough to fill the space is a separate question, and it is the one this page is really about.
This page picks up where FCL vs LCL, and the CBM where it flips stops. That one decides whether a shipment is big enough to book a whole box instead of paying by the cubic metre, and it closes by saying it has nothing to say about whether you should be buying a container's worth of a product at all. This one does: the cartons-per-box arithmetic first, because everything else is built on it, then what to do with the space once the box is booked.
What the carriers say a container holds
The figures below are Maersk's, from its own dry container specification sheet, read on 29 September 2026. Hapag-Lloyd's container specification booklet, dated March 2016 and read the same day, gives the same shape of number with slightly different values from one series of boxes to the next, which matters more than it sounds.
| Container | Inside length × width × height | Door opening, width × height | Capacity | Maximum payload |
|---|---|---|---|---|
| 20 foot standard | 5,896 × 2,350 × 2,393 mm | 2,350 × 2,274 mm | 33 m³ | 28,200 kg |
| 40 foot standard | 12,032 × 2,350 × 2,393 mm | 2,340 × 2,274 mm | 67 m³ | 28,800 kg |
| 40 foot high cube | 12,032 × 2,350 × 2,697 mm | 2,340 × 2,577 mm | 76 m³ | 28,620 kg |
Two things on that table are worth noticing before you use it. A 40 foot box is not twice a 20 foot box in weight. All three carry roughly the same payload, so a dense product fills a 20 foot container's weight allowance long before it fills its floor, and a 40 foot box buys you space, not tonnage. And the high cube adds 30 centimetres of height and nothing else. Whether those 30 centimetres are worth anything depends entirely on whether your carton stacks into them, which is the next section.
Then there is what none of these numbers are: measurements of the box you will get. Hapag-Lloyd puts it plainly in its booklet: the inside dimensions “are nominal figures. Because of production tolerances, a difference in measurement is possible”. Its own 40 foot high cube series run from 3,830 to 4,300 kg of tare weight, so the payload runs from 28,670 down to 28,200 kg depending on which box turns up. Maersk's two pages do not even agree with each other: the specification sheet says 28,200 kg for a 20 foot standard box and its cargo weight FAQ says 28,300. None of that is carelessness. It is what nominal means, and it is why loading plans are made to the carton and not to the litre.
The weight figure has a second catch. It is what the box is rated to carry, not what a truck may carry on your roads. Maersk's FAQ says “specific weight limits can vary by country” and to contact the local office, and Hapag-Lloyd says the same of road and rail limits. Ask your forwarder what a loaded box may weigh at destination before you plan to the payload, because a container that is legal to ship and illegal to drive to your warehouse is a problem you find out about at the port.
Why the volume division is a number you never load
Take a carton of 60 by 40 by 40 centimetres, which is 0.096 cubic metres, holding 24 units. Seventy-six cubic metres divided by 0.096 is 791 cartons, and that is the figure a container calculator will print. Now load it.
The inside of the high cube is 12,032 mm long. Twenty cartons at 600 mm use 12,000 of it. It is 2,350 mm wide, and five cartons at 400 mm use 2,000, leaving a 350 mm strip that a 400 mm carton cannot enter. It is 2,697 mm high, and six layers at 400 mm use 2,400, leaving 297 mm of air above the top layer. Twenty by five by six is 600 cartons, 57.6 cubic metres, about three quarters of the nominal 76. The other quarter is not lost through bad packing. It is the shape of the carton against the shape of the box, and while the cartons stay flat no amount of care recovers the 297 mm of headroom, because nothing in this carton is shorter than 400 mm.
Turn some cartons and you do a little better. One carton laid across at 600 mm beside four at 400 mm uses 2,200 of the 2,350 mm width instead of 2,000, and repeated down the length that pattern gives 110 cartons a layer instead of 100, so 660 in the box. Stand one layer on end, 600 mm tall on a 400 by 400 mm footprint, and five flat layers plus that one use 2,600 of the 2,697 mm: 150 cartons in the standing layer, 700 in the box. That is the kind of thing a good packer at the factory works out, and it is worth asking for their loading plan rather than assuming yours. Even so, 700 is about the ceiling for this carton in this box, and 791 was never on offer.
The same carton, in the three boxes, loaded the plain way:
| Container | By volume | Loaded one way round | Nominal volume actually used | Units, at 24 a carton |
|---|---|---|---|---|
| 20 foot standard | 343 cartons | 9 × 5 × 5 = 225 cartons | 65% | 5,400 |
| 40 foot standard | 697 cartons | 20 × 5 × 5 = 500 cartons | 72% | 12,000 |
| 40 foot high cube | 791 cartons | 20 × 5 × 6 = 600 cartons | 76% | 14,400 |
Read the 20 foot row slowly. Its inside height is 2,393 mm, and a sixth layer of 400 mm cartons needs 2,400. Seven millimetres cost a whole layer, 45 cartons on top of the 225. The 40 foot standard box, at the same 2,393 mm, loses the same layer, and there it is 100 cartons. A carton two centimetres shorter would have put 270 cartons in that box instead of 225. If you have any say over the carton, this is where a conversation with the supplier pays for itself, and it is a conversation most importers never have because the calculator said 343.
Pallets take more. A pallet has its own footprint and its own height, and cartons that overhang or under-fill it waste space twice, so a palletised load fits fewer cartons than a floor load of the same product. Shapiro, a US freight forwarder, adds the practical limit on top of the geometric one: it recommends customers stop loading a 20 foot container at about 27 of its 33 cubic metres “to allow for air space for unloading”, read on 29 September 2026. Between the carton's shape and the room to get it out again, the honest planning figure for a box is a fair way under its nameplate.
Cube out or weigh out
A container is full when either limit is reached, and which one comes first is a property of your product, not of the box.
| Light product | Dense product | |
|---|---|---|
| Unit weight | 0.9 kg | 2.5 kg |
| Carton, 24 units plus packaging | 23 kg | 61.5 kg |
| Cartons the 40 foot high cube takes by space, loaded one way round | 600 | 600 |
| Weight of those cartons | 13,800 kg | 36,900 kg |
| Cartons the 28,620 kg payload allows | 1,244 | 465 |
| The box is full at | 600 cartons, 14,400 units. It cubes out | 465 cartons, 11,160 units. It weighs out, with about a quarter of the space empty |
The dense product weighs out at 465 cartons with a quarter of the space still empty, and that empty space is not available to anyone: adding another carton makes the box overweight. This is the case where a high cube is money wasted, since the extra height is air you are already not allowed to fill, and where the road limit above can bite before even the container's own rating does. Which case you are in decides which number to plan against, and the weight per carton is the one figure on this whole page you have to get from your own product rather than from a carrier's sheet. The same weight-or-volume logic sets how a freight bill is split across the units afterwards, worked through in landed cost for Shopify stores, and this page does not repeat it.
Should you order enough to fill a container?
Here is the shift in the question. Everything above answers how much fits. Nothing above says how much you should buy, and the two get confused because the first number is so easy to produce. Fourteen thousand four hundred units is what the box holds. What you need is your daily sales times the days the order has to cover, minus what you hold and what is already on the way, and how many days an order should cover is its own page. If the two numbers happen to match, good. When they do not, buying to the box is buying stock because it fits, and stock bought because it fits is a minimum order quantity you set for yourself.
That framing is useful because the arithmetic already exists. Minimum order quantities, and what to do when the MOQ is more than you need works out what accepting a floor really costs: the surplus units above what you need, divided by daily sales, is extra days of cover, and that cover is compared against the product's season and against the cash it ties up. A container fill-up goes through exactly that test. The only difference is that the person imposing the minimum is you, and the discount on offer is the freight the extra units ride for free.
Free is the word to be careful with. Once a box is booked, the marginal freight on another carton is zero right up to the point where the box is full, and that is a real saving. It is also the only saving. The extra units still cost their purchase price today and still sit on a shelf until they sell, with whatever chance the product has of going out of fashion first. How much cash the next container ties up counts the first of those; the other two are yours to judge. What fills a container profitably is demand that was going to be there anyway, pulled forward by a few weeks. What fills it expensively is demand you hope will turn up.
Where the economic order quantity fits, and where a box breaks it
There is a formula for exactly this trade, and it is older than the shipping container. In 1913 a production engineer named Ford W. Harris published How many parts to make at once in Factory, The Magazine of Management, reprinted in Operations Research in 1990 and read in that reprint on 29 September 2026. His opening line is the whole idea: “Interest on capital tied up in wages, material and overhead sets a maximum limit to the quantity of parts which can be profitably manufactured at one time; 'set-up' costs on the job fix the minimum.” Swap set-up cost for the cost of placing and shipping an order and you have the economic order quantity, EOQ, as inventory textbooks teach it.
EOQ = √(2 × annual demand × cost per order ÷ annual holding cost per unit)
Harris wrote it with a 10 per cent carrying charge baked into the constant. This is the same formula with that charge as a variable.
Run it on a product selling 25 a day, 9,125 a year, bought at 8.00 and carried at an illustrative 25 per cent a year, so 2.00 per unit per year to hold. Say the fixed cost of an order, the admin plus the origin and destination handling on an LCL shipment, is 500. The formula gives √(2 × 9,125 × 500 ÷ 2) = 2,136 units, about 85 days of sales, 89 cartons, 8.5 cubic metres. That is an LCL shipment, well under the range where a full box starts to win, and for this product the formula is doing its job.
Now look at what the formula assumed to get there. The 500 is a cost per order that does not change with the quantity: Harris divides the set-up cost by the lot size to get the set-up cost per unit, which only works if the set-up cost is the same whatever the lot. An LCL shipment more or less honours that. The per-cubic-metre rate is proportional to the quantity, so it drops out of the decision, and the handling charge is the fixed part. A container does not honour it at all. Below the crossover the fixed cost is small and the freight is variable; at the crossover the fixed cost jumps to the price of a box and the variable part falls to nothing; at the next box it jumps again. The cost per order is a staircase, and the formula has no way to say so.
| What EOQ assumes | LCL, paid by the cubic metre | FCL, paid by the box | |
|---|---|---|---|
| Fixed cost per order | One number, the same whatever the quantity | Close enough: the handling charges | The whole box price, whether it sails full or a third empty |
| Freight per extra unit | Constant, so it drops out of the decision | Roughly constant per cubic metre | Zero until the box is full, then the price of a second box |
| Cost curve | Smooth, one minimum | Smooth enough | A sawtooth: cost per unit falls to each box boundary and jumps at it |
| Where the answer lands | Anywhere | Anywhere | Ideally just under a boundary, never just over one |
Feed the same product a flat box price instead, the 1,900 for a 20 foot container that the FCL vs LCL worked example uses, and the formula obediently answers √(2 × 9,125 × 1,900 ÷ 2) = 4,164 units, 167 days of sales, 16.7 cubic metres. Two self-consistent answers to one question, because the cost per order depends on which answer you picked, which is the circularity a container introduces and a spreadsheet EOQ never notices. Worse, 4,164 units is 174 cartons in a 20 foot box that takes 225, so 51 cartons of space are still empty, and the formula cannot tell you that those 51 cartons now ride free.
Two forwarders say the same thing from the other side. iContainers, in a page updated on 18 September 2026 and read on 29 September, now writes that “there is no fixed volume at which FCL always becomes cheaper than LCL. The crossover depends on the route, shipping date, container rate, LCL tariff, local charges and delivery arrangements”, and warns that a rule like “switch to FCL at 15 CBM” can miss destination fees. Shapiro still gives its rule of thumb, “about 15 cubic meters”, as the point to start considering a full container. What matters for this page is that the fixed cost EOQ wants as an input is the very thing the forwarders will not commit to.
Harris would not have minded. He wrote that no formula “should be depended upon entirely”, that using it “as a check, is at least warranted”, and that in his examples the loss from too small a lot was larger than the loss from too large a one. On a container that last point cuts both ways: rounding down to stay under a box boundary costs you the per-cubic-metre rate on the overflow, and rounding up past it costs you a whole second box. Use the formula to find the neighbourhood. Use the box to decide which side of the boundary to land on.
Fill the space or leave it empty
The box is booked. The order that justified it is 4,164 units, 174 cartons, and there is room for 51 more cartons, about 1,240 units, at no extra freight. Fill it? Put the year's costs side by side for the 25 a day product, using the same illustrative rates as above, so the three plans can be compared on one line each.
| Plan | Order size | Shipments a year | Freight a year | Holding a year | Total a year |
|---|---|---|---|---|---|
| LCL at the EOQ | 2,136 units, 8.5 CBM | 4.3 | 5,786 | 2,136 | 7,922 |
| A 20 foot box at the EOQ, 51 cartons left empty | 4,164 units, 16.7 CBM | 2.2 | 4,164 | 4,164 | 8,328 |
| The same box, filled to its 225 cartons | 5,400 units, 21.6 CBM | 1.7 | 3,211 | 5,400 | 8,611 |
Filling the last 51 cartons saves about 950 a year in box bookings and costs about 1,240 a year in stock sitting on the shelf, so the full box is still the most expensive of the three plans for this product, if only by about 300 a year, and the shared container is the cheapest. That is before counting the cash: 1,236 extra units at 8.00 is about 9,900 out of the bank on order day, for stock that does not start selling until the sixth month, and before any chance that the product stops selling at all. A full box is not the cheapest box. It is the cheapest freight per unit, which is a different sentence.
Change the demand and the answer flips. Quadruple the sales to 100 a day, 36,500 a year, and the LCL plan costs about 23,100 a year, at 17 cubic metres a shipment that would be past the crossover in any case, while a 20 foot box filled to its 225 cartons every 54 days costs about 18,200. At that volume the box wins, and it wins full, because the units filling it were going to be bought within two weeks anyway. That is the whole rule in one line: a box pays when the units that fill it are pulled forward by a few weeks, and it costs when they are pulled forward by months. On these rates the line sits at about five weeks, which is why the seven-week fill above loses, and only just, and it moves with what it costs you to hold stock: a cheaper carrying charge pushes it out, a dearer one pulls it in. Which of those you are looking at is your daily sales rate set against the box, and nothing on a carrier's sheet knows your sales rate.
What to fill it with, if you fill it
The fill decision is rarely about one product. A container from one supplier carries a dozen lines, and the question is which of them should take the spare cartons. Ranked by how many weeks of demand each choice pulls forward:
| Fill the space with | What it pulls forward | Verdict |
|---|---|---|
| Products from the same supplier whose order deadline falls inside the next cycle anyway | A few weeks of demand you were about to order | Best use of the space. The next order shrinks by the same amount and the freight on those units is genuinely saved |
| More of the product already on the box | Seven weeks, on the example above, and it still loses by about 300 a year | Only if the units would otherwise have shipped LCL later. If the next box sails full regardless, the saving is zero and only the holding cost is real |
| Whatever is cheapest per unit, or was on offer | Demand you have not measured | The per-order minimum problem in a different coat, and the usual source of dead stock |
The first row is the one to work from, and it needs a list you may already have: every product from that supplier, ordered by how soon its order has to go in. Pull from the top until the box is full or the next deadline is far enough out that pulling it forward starts to cost real months. The second row has a trap in it that is easy to miss. Topping up product A on this box only saves money if it removes freight from a future shipment. When you ship a full box every cycle whatever happens, moving units from the next box to this one changes nothing but the date the money leaves.
How this lands in Restocio
Restocio treats a container as sea freight paid per box. Under Settings, Freight & Margin, you type what one container costs you door to door and the weight it typically carries, and every unit you put in the container is priced by its share of that weight. In the order dialog the container is a column per line, next to boat, train, plane and truck. Container units come out of the boat quantity first, and if you type more than the recommendation the extra is added to the order, which is how a deliberate top-up is placed. The summary line then says how many boxes that weight fills, rounded up, and what the freight comes to. Recommended quantities round up to whole cartons when a product has a pack size, and the Replenishment list sorts by days of stock left, soonest first, which is the list the first row of the table above asks for.
Now the limitation, and it is the largest one on this page. Restocio has no cubic metre field. It knows a product's weight and nothing about its size, so the “about two containers” it prints is a weight estimate: total kilos divided by the kilos you said a box carries. For a dense product that is close to right, because the box weighs out. For a light, bulky product the box cubes out long before the weight figure says it is full, and the app will let you plan 600 cartons' worth of units into what it thinks is half a box. The cartons-per-container arithmetic above has to be done before the units reach us, and if your products fill boxes by volume, the honest reading of our container line is a lower bound on the number of boxes, never the count.
Two more, plainly. The number of days one order is meant to last is fixed at 30 inside the engine today and is not a setting you can change, so Restocio never stretches an order to fill a box. It sizes on demand and lead time, and a fill-up is a decision you make in the order dialog, not one the app makes for you. And the economic cycle above, an EOQ-style order size that weighs freight against holding cost per product, is a proposal in our own planning notes rather than a built feature. It would need the fixed cost per order that we do not collect, and until it exists the arithmetic in this post is yours to run. We do not quote freight, and we do not know the road weight limit at your destination either.
Common questions
How many cartons fit in a 40ft container?
Fewer than the volume division says. A 40 foot high cube is about 76 cubic metres inside on Maersk's and Hapag-Lloyd's own sheets, so a 60 by 40 by 40 cm carton divides into it 791 times by volume, but only 600 cartons load one way round, because 350 mm of width and 297 mm of height are left over. Turning some cartons reaches 660, and standing one layer on end reaches about 700. Work it from the inside dimensions and your own carton, then check the weight against the payload.
How many units fit in a 20ft container?
Take the inside dimensions, about 5,896 by 2,350 by 2,393 mm on Maersk's sheet, divide each by the matching dimension of your carton, round each down and multiply the three, then multiply by units per carton. For a 60 by 40 by 40 cm carton that is 9 by 5 by 5, 225 cartons, or 5,400 units at 24 a carton, against 343 cartons by volume. Then check that 225 cartons weigh less than the payload, around 28 tonnes, and less than what the roads at your end allow.
Should I order enough to fill a container?
Only when the units that fill it were going to be bought within weeks anyway. Once a box is booked the extra cartons ride free, but they still cost their purchase price today and their holding cost until they sell. In the worked example a product selling 25 a day was cheaper shipped LCL than in a box, and cheaper in a box with 51 cartons of space left than in one filled, while at 100 a day the full box won clearly. Treat a fill-up as a minimum order quantity you set for yourself and run the surplus-cover test on it.
Is a 40ft high cube worth it over a standard 40ft?
Only if your carton stacks into the extra 30 centimetres. A high cube is 2,697 mm high inside against 2,393, which for 400 mm cartons is six layers instead of five, a fifth more cartons for the same floor. If your product weighs out before it cubes out, the extra height is air you are not allowed to fill and the high cube gains you nothing. Check the payload first, then the stack.
Does EOQ work for goods shipped by container?
As a starting point, not as an answer. EOQ assumes a fixed cost per order that does not change with the quantity. Paid by the cubic metre that roughly holds, so the formula works for LCL. Paid by the box it does not: the fixed cost jumps at each container boundary and the marginal freight drops to zero in between, so the formula gives a different answer depending on which cost you feed it. Use it to find the neighbourhood, then decide which side of the box boundary to land on by hand.
Does Restocio tell me whether my order fills a container?
By weight, not by volume. Type the price per container and the weight it carries under Settings, Freight and Margin, and Restocio prices container units by weight and says how many boxes the order's kilos fill. It has no cubic metre field, so a light, bulky product can fill a box by volume before the weight count says it is full. The carton arithmetic is done outside the app, and the app never stretches an order to fill a box on its own.
Sources
Every source below was opened and read on the date shown against it. Rules, rates and schedules change, so check the current page before you act on a number that matters. If anything here is out of date or wrong, tell us and we will correct it.
- Maersk, Our fleet, equipment specifications, dry containers (PDF) (read 29 September 2026)
- Maersk, What is the weight limit for the cargo (support FAQ) (read 29 September 2026)
- Hapag-Lloyd, Container Specification booklet (PDF, March 2016 imprint) (read 29 September 2026)
- Shapiro, LCL (Less Than Container Load) or FCL (Full Container Load), that is the question (read 29 September 2026)
- iContainers, LCL vs FCL: Differences, Costs and How to Choose (updated 18 September 2026) (read 29 September 2026)
- Harris, F. W. (1913), How many parts to make at once, Factory 10(2), as reprinted in Operations Research 38(6), 1990, pp. 947 to 950 (copy hosted by Brooklyn College, CUNY) (read 29 September 2026)
Marcus co-founded Restocio and works on it daily with a Swedish importer who plans their purchasing in it every working day. Restocio is built in Sweden by two founders, one Swedish and one Danish, and Marcus is the Danish one. Why we are building it.
Related reading
- FCL vs LCL, and the CBM where it flips
- Minimum order quantities, and what to do when the MOQ is more than you need
- Landed cost for Shopify stores, and why your margin is wrong
- How many days of stock should you hold
- How much cash to hold for the next container
- Importing for Shopify, and where to start
- Shopify replenishment, end to end
Restocio plans purchasing for Shopify stores that import. It works out what to order, how much, and whether it should travel by sea, rail or air, so you pay air freight only for the units that genuinely cannot wait.