Sea-Air vs Rail vs Ocean vs Air: A Four-Column Read

Nobody Sells You a Comparable Set of Four Numbers

Ask for a quote on a China-to-Europe shipment and you will receive four prices. You will not receive four comparable transit figures, and the reason is not dishonesty — the numbers are not constructed the same way. An ocean schedule counts from a vessel departure. A rail schedule counts from a terminal acceptance that depends on a wagon being available. A sea–air lane is two schedules joined at a hub whose dwell belongs to neither. Express counts from a scan. Put the four brochures side by side and you are comparing four different start lines, which is why the cheapest column so often arrives as the most expensive decision.

The useful version of this comparison is not a table somebody publishes. It is a table you build once from your own freight, because the variables that decide which column wins are specific to your density, your destination, and how long your cash is tied up. What follows is the structure of that table, and what each column can be held to.

Fix the Denominator Before Comparing Anything

Freight cost per shipment is the wrong denominator because it hides the two things that actually differ: how much capital the journey immobilises, and how much shelf time it consumes. State the comparison as cost per unit delivered plus days of stock in transit, and the columns separate properly.

A worked arithmetic on a hypothetical consignment of 1,000 units, each carried at a landed cost of EUR 40, with inventory that sells at 25 units a week:

• A 35-day arrival holds roughly 1,400 unit-weeks of stock in the pipeline. A 15-day arrival holds 600.
• The 20-day difference frees 800 unit-weeks of capital — in this example, the equivalent of about three weeks of selling, which is the quantity a buyer should be weighing against the freight gap rather than against the freight total.
• A stock-out, meanwhile, is not a carrying cost at all. If the listing loses rank while out, the cheap column has to pay for the advertising that buys the position back.

The arithmetic is illustrative, not a quote; substitute your own unit cost and sell-through. What does not change is that a mode comparison without a carrying-cost term is only a rate comparison wearing a coat.

The Ocean Column: Cheapest Per Unit, Longest Tail

Full containers and groupage dominate the low end of cost per unit for goods that are heavy relative to their volume, and 2026 made the timing term less predictable rather than removing it. Partial returns to the short route through the Suez Canal coexisted with operators still routing the long way, which means two carriers on the same trade can honestly quote a fortnight apart. Take from the ocean column what it is good at — cost, capacity, no weight penalty on dense cargo — and treat its arrival date as a distribution rather than a date. Book it for stock you can plan a month ahead, not for a launch.

The Rail Column: The One With Published 2026 Data

China–Europe rail freight is the only column in this comparison with official volume statistics you can check for yourself. In the first quarter of 2026 the service handled 5,460 train trips carrying 546,000 twenty-foot equivalent units, up 29% and 22% respectively against the same quarter a year earlier, reaching 235 cities in 26 European countries. The operator also reported that on-time scheduled products cut transit by more than 30% against comparable services on the same corridor.

Those figures support two conclusions and not a third. They support the idea that the corridor has capacity and frequency, because a service growing at that rate is not a weekly experiment. They support the idea that a scheduled product differs materially from a regular one on the same path. They do not support any specific number of days from your factory to your customer’s door, because border dwell, gauge change and terminal acceptance are not in the statistic — and those are the three things a rail transit claim actually depends on.

The Hybrid Column: Where the Two Schedules Meet

Sea–air buys back part of the ocean column’s time without paying for the whole air move, and its real characteristic is that it is the only column whose transit depends on a handover rather than on a vehicle. Every number in a hybrid quotation is the sum of a vessel arrival, a dwell at the transfer point, a flight build-up cut-off, and an uplift. The dwell belongs to no published schedule. Ask who is accountable for it, in days, and ask what happens if the hub’s gateway changes while the container is at sea.

The Air Column: Fast, Priced by Capacity, Capped by Convention

Airfreight cost tracks capacity released to the trade, and belly-hold space is the elastic part — which is why a lane can be re-priced between two quotations in the same week without anything about your cargo changing. Two further properties belong in the comparison. Weight matters differently: light, bulky cartons are billed on volume rather than on the scale, so the air column punishes packing that the ocean column barely notices. And recovery differs: under the Montreal framework the carrier’s liability for cargo is a per-kilogram figure — a limit of 26 SDR per kilogram of gross weight — which is not a policy and will not pay for a pallet of electronics.

Build the Table From Six of Your Own Shipments

1. Pick six consignments of the same SKU family, similar density, similar destination, spread across a season rather than clustered in one week.
2. Record five timestamps for each: goods ready, carrier acceptance, departure, arrival at the port or hub of entry, delivery to door.
3. Compute two spans, never one: transport days and total days from ready to door. The gap between them is origin dwell, and it is usually the largest single thing you can fix.
4. Add your own carrying term: days in transit multiplied by units per day multiplied by unit cost. Then add the freight invoice.
5. Repeat for whichever column looked cheapest. The comparison is only valid when the same denominator is applied to all four.

Six shipments is enough to see whether a mode’s spread is tight or bimodal, which is the information a published range cannot give you. Where the spread is bimodal, the honest conclusion is that you are not choosing a mode — you are choosing a schedule, and the mode is only the label on it.

Three numbers make a forwarder’s answer worth having: your density, the volume of cartons per month, and the number of unit-weeks of stock you are currently carrying. With those, a comparison of four columns is arithmetic. Without them, it is a brochure. We will run all four against the same denominator and mark the assumption in every figure, because a transit we have not measured is an estimate, and estimates should arrive labelled as such. Send the six timestamps if you have them; if you do not, send the SKU, the destination and your sell-through rate, and we will tell you which column we would book and why we might be wrong.