← All posts

Freight

Sea, rail or air, and why the answer is often all three

The cheapest freight decision is rarely one mode for the whole order. It is the cheapest mode that still arrives in time.

Marcus VolstedCo-founder, Restocio Published 7 August 2026Updated 10 October 202613 min read

Facts about Stocky and Shopify's built-in tools are taken from Shopify's own Help Center. Restocio is an independent product and is not affiliated with or endorsed by Shopify.

The rule everything else hangs on

The cheapest way to move an order is the slowest freight that still lands before you run out. When no single mode manages that, split the order: the bulk goes by sea, and a smaller bridge quantity travels by rail or air, sized to cover only the days you would otherwise be empty. That is the whole method. The rest of this page is what each mode actually buys you, how to size the bridge, and what the arithmetic does when you decide a week late.

Roughly, and it varies by lane, season and carrier:

Treat these as shapes, not quotes. Freight rates move with the lane, the season, fuel prices and whatever happened in the news this quarter, which is why this page quotes none. The only rates that matter for your planning are the ones on your own forwarder's invoices. What does not move is the ratio: air is not slightly more expensive than sea, it is a different order of magnitude, and that single fact drives every decision below.

What door to door really contains

Most stores plan against the transit time on a freight quote, and it is the wrong number. The clock starts the day you decide to order, and the goods can be sold only after all of this has happened:

• Order handling. The purchase order agreed, the deposit paid, the production slot or sailing booked. Days, sometimes weeks.
• Production. Made-to-order goods do not exist yet. For clothing this is routinely 30 to 45 days on its own, and it appears on no freight quote. Air freight does not skip the factory: production delays every route equally, which is why a made-to-order product with a fast flight booked is still a slow product.
• Transit. The part the quote actually covers. Port to port, or terminal to terminal.
• Customs and inland. Clearance, duties, and the truck from the port to your warehouse.
• Putaway. Goods on the receiving dock are not goods you can sell.

Add those up per mode and per supplier, from real past orders rather than from quotes, and you have the lead times the rest of this page runs on. Planning with port-to-port times is the quiet reason so many orders where "the boat was on time" still arrived after the shelf went empty.

The decision ladder

With honest lead times per mode, choosing the mode is mechanical.

First check whether your stock outlasts the sea journey. Days of stock is units on hand divided by units sold per day. If that number is comfortably bigger than the sea lead time plus its safety margin, everything travels by sea, because sea is the cheapest and paying for speed you do not need is a donation to an airline.

If stock runs out before the boat lands, compute the gap: sea lead time minus days of stock. That is how many days the shelf would sit empty. Then buy only enough faster freight to cover those days. Rail first, because it is the cheaper of the fast options. Air only for whatever part of the gap sits in front of rail's own arrival. Each faster mode covers the stretch between the day it lands and the day the next slower mode lands.

Two details make or break this in practice:

• The bridge is added on top of the sea order, never carved out of it. The sea order's job is unchanged: cover the coverage period at the cheap rate. The bridge's only job is to keep the shelf alive until the boat docks. Treat them as one budget and you will under-order both.
• Whatever is already moving reduces the bridge, mode by mode. Units already on a boat reduce the sea need, units already on a train or plane reduce that bridge. Skip the netting and you double-buy exactly the products you are most nervous about, then watch the original shipment land two weeks later.

And never upgrade the whole order. Putting a full container's worth of goods on a plane because 300 units are needed in week two is the single most expensive mistake available in this entire process. The 300 fly. The rest sails. The replenishment guide walks the same logic from the quantity side.

One boundary case, stated plainly: if stock will not last even the air lead time, no freight decision rescues you. You are choosing which customers to disappoint, and the real lesson is upstream, in ordering earlier next time.

A split you can check by hand

Take the product used in examples across this site: it sells 10 a day and has 100 units left. Sea takes 70 days door to door, rail 20, air 5. Safety margins are 30 days on sea, 7 on rail, 3 on air, and the order should cover 30 days of demand.

Now put money on it, using ratios rather than rates. Assume every unit weighs the same, and pick example ratios of rail at twice sea per kilo and air at five times, chosen here for clean arithmetic, not as market data. Call the freight bill for sending all 1,700 units by sea the baseline. The split above costs about 1.6 times that baseline, because the 700 bridge units pay the fast rates. Flying the whole order costs 5 times the baseline. And the baseline itself was never really available, because all-sea means an empty shelf from day 10 to day 70.

The interesting part is what the same product costs at different decision dates. Same demand, same lead times, same targets. The only thing that changes is how many days of stock remain when you finally act:

Nothing about the product changed between those rows. Only the calendar did. That is why the order deadline, the last day everything can still go by boat, is worth more attention than any negotiation you will ever have with a forwarder.

Safety days per mode, and why sea needs the most

Each mode gets its own safety margin, in days, on top of its lead time. The example used 30 for sea, 7 for rail, 3 for air, and the pattern is not arbitrary. The longer and more complicated the journey, the more days it has available to lose.

A sea shipment can miss its sailing, sit at anchor outside a congested port, wait for transshipment, or lose a week to a customs inspection, and each of those lands on a journey that is already months long. A flight has fewer moving pieces and fewer days to lose. In practice the absolute spread grows with the journey: a boat being somewhat late is measured in weeks, a plane in days. Plan to the average arrival and you will be late roughly half the time, which for an importer means an unplanned air bridge roughly half the time.

Worth admitting: a flat margin per mode is a compromise. The more honest version sizes the margin per product, from how variable that product's demand and its supplier's lead time actually are, and it is worth asking any tool vendor whether their margin is flat or computed. A flat margin is still far better than none. It is simply a compromise you should know you are making.

Weight data comes first

Freight is priced by weight, with volume taking over for bulky goods, and weight is the working currency of every comparison on this page. The consequence is blunt: without a weight per unit, modes cannot be compared at all. "Air is several times sea" means nothing for a specific product until you know how many kilos of it you are moving.

Unit weights usually already exist in your Shopify catalogue, entered once for shipping rates and never looked at again. Pull them out, fill the gaps, and treat a product without a weight as a product without a freight plan. A tool should refuse to estimate here rather than invent: Restocio marks a product with a missing weight as not calculable instead of printing a made-up margin, and the honest spreadsheet equivalent is a blank cell, not a guess.

One supplier, one shipment, many deadlines

Freight is bought per shipment, not per product, and that changes how orders should go out. The rhythm that works for most importers: deadlines decide what is due, and a periodic review per supplier decides when orders actually leave, so that everything from one factory travels together. Consolidating by supplier is often worth more than optimising any single product, because it saves an entire shipment.

The interaction with each product's own deadline has one governing rule: the shipment date must respect the earliest deadline in the group. Products whose deadlines fall later simply get ordered a little early, which parks a little extra cash and is usually cheap. The expensive direction is the other one. Hold the group past the tightest product's deadline and that product now needs its own air bridge, which tends to cost more than everything the consolidation saved.

Consolidation also solves the smallest-shipment problem. Three units of one product by air will never pay for the freight they ride on. But ten products needing ten fast units each from the same supplier are one hundred-unit air shipment, which might well pay for itself. Judge minimum shipment sizes at the shipment level, across every product sharing the supplier and the mode, not per product line.

The calendar moves every deadline

Two seasonal effects are worth writing into a calendar rather than rediscovering annually.

Chinese New Year. Factories close for weeks, and capacity is booked out for weeks before and after. Orders that normally take 60 days can take roughly twice that. Every product whose deadline falls anywhere near the closure effectively has its deadline pulled forward, because the lead time behind that deadline just grew.

Peak season. Freight in the run-up to Q4 is slower and more expensive at the same time. Space gets scarce, sailings roll, and the safety days that were adequate in May are optimistic in September.

Neither of these appears in your sales history, and no forecast computed from that history will warn you. This is exactly the kind of knowledge the person has and the data does not, which is why a planning tool should surface the deadline and let you act early on what you know is coming.

What a late decision costs, per day

This mechanism deserves spelling out, because the bill never says "hesitation" on it. Every day you delay past the sea deadline, the boat lands one day further past the point your stock runs out, so the bridge grows by one day of sales. At 10 units a day, deciding after the weekend adds about 20 units to the premium-freight quantity. A week of "let me think about it" adds 70. The table above is the same effect at full scale: between acting at 75 days of stock and acting at 10, the example freight bill more than doubles.

The freight bill for lateness grows quietly, and it is almost never attributed to the decision that caused it. The air invoice arrives weeks after the deadline passed, gets filed under freight costs, and the next quarter it happens again.

Landed cost keeps the score

The reason to care about modes is not the freight invoice itself. It is what freight does to the margin of every unit it carries. Landed cost is the supplier price plus that unit's share of freight plus duty, and it differs per mode: the same product flown in carries more freight cost per unit than the same product shipped by sea, so it keeps a smaller share of its selling price. Seen side by side per mode, that is the real purchasing trade-off, and a product with a healthy sea margin can be a marginal product by air.

Margin computed on the supplier price alone hides all of this, which is how stores end up proudly scaling products that lose money every time they fly. The full method, including duty and what to do when your costs already include freight, is in the landed cost guide.

What to track

• Realistic door-to-door lead time per mode, including order handling, production, customs and putaway, taken from real orders rather than quotes.
• A safety margin per mode, largest for sea, because delays are normal and planning to the average is planning to be late half the time.
• Weight per unit. Freight is priced per kilo, so without weights you cannot compare modes at all.
• Your actual freight rate per kilo per mode, taken from a real invoice rather than a list price.
• What is already in transit, per mode. Units on a boat reduce the sea need; units on a plane reduce the bridge.
• Landed cost per unit, per mode. Goods plus freight plus duties, the only number that tells you the true margin of the units you are about to move.

Where this method stops

Restocio does not book freight, does not quote rates, and has no forwarder integration. It decides the quantities: how many units of each product should travel by each mode, by which date, net of what is already moving. You take that split to your forwarder, and your invoices supply the rates. For most importers that is the right division of labour, but if what you want is a booking platform, this is not one.

The flat per-mode safety margin discussed above applies here too, and one more limit deserves stating: a first order on a new lane is a guess. There is no history of how your goods clear your port, and no arithmetic on this page substitutes for asking the forwarder hard questions about the first shipment. The method gets sharp from the second order onwards, when the lead times come from your own experience.

Common questions

Is it cheaper to ship everything by sea?

Per kilo, yes. But if the shelf empties before the ship arrives, the lost sales and the emergency air freight that follows usually cost more than the saving. The cheapest total outcome is normally sea for the bulk plus a small air or rail bridge.

Should I split one order across sea, rail and air?

When stock will not outlast the sea journey, yes. The bulk goes by sea and a smaller quantity travels faster, sized to cover only the days you would be empty before the boat lands. The bridge is added on top of the sea order, not taken out of it, and upgrading the entire order to a faster mode is almost always the most expensive option.

How many safety days should each freight mode get?

More for slower modes. A sea journey has months of moving pieces and can slip by weeks, while a flight can slip by days, so sea deserves the largest margin and air the smallest. A flat number per mode is a workable compromise; a margin computed per product from real variability is better.

Does production time change the freight decision?

It moves the deadline, not the ladder. Made-to-order goods add production days ahead of every mode equally, because a plane cannot leave before the goods exist. So the order deadline arrives earlier, but the sea, rail and air comparison itself is unchanged.

What is landed cost?

What one unit actually costs you on your shelf: the price paid to the supplier plus its share of freight, plus duties and any other import costs. Margin calculated on the supplier price alone overstates profit, sometimes badly.

Does Restocio book the freight?

No. It calculates how many units should travel by each mode and by when, netting off what is already in transit, and you book the shipment with your own forwarder. Rates come from your invoices, not from the app.

Marcus Volsted

Co-founder, Restocio

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

• Landed cost, the number that keeps the score
• Shopify replenishment, end to end
• Stock replenishment, the complete guide