Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- For any promotion with a fixed date, work backwards from the delivery date: subtract 7 days for air, 25 for rail and 60 for ocean, then confirm the production start falls after the order is placed, target: a written date plan per SKU within 3 working days.
- Settle the refrigerant handling arrangement with the airline and the ground handler before booking space, target: handling approval in hand at least 72 hours before the planned departure.
- Ask the factory for its committed daily allocation to your order rather than its nameplate capacity, target: a written daily output number that covers the whole promotion quantity.
- Make rail the default for any order with 30 days of lead time, target: at least half of non-urgent volume on rail, with air reserved for the units that must hit the promotion date.
- Confirm dock appointments with the retailer's distribution centres before the first shipment leaves, target: an appointment for every inbound container and no unit arriving without a slot.
A sustained European heatwave turned air conditioners and fans into best-sellers, with a French supermarket's July 2 promotion of 200,000 AC units causing chaos. A Chinese factory is running at up to 30,000 units per day, using air freight for urgent French orders despite refrigerant handling rules, while ocean and rail take 2 months and 25 days respectively.
The Analysis
On July 2 a French supermarket put 200,000 air conditioners on promotion and the queues formed before the doors opened. Behind that number sits a Chinese factory running up to 30,000 units a day, air freight for the urgent French orders, two months for ocean and 25 days for rail. Everything in this story is a comparison of clocks, and only one of them fits inside a heatwave.
Two hundred thousand units is not a demand signal. It is a point load, dropped on one retailer on one date. Capacity planning is built for ramps, not for a single spike at a single address, and that mismatch is where the mess actually happens.
Here is how the hours stack up on each mode, who ends up holding the stock, and what I would book before the next heat dome forms.
Run the production figure against the promotion figure. At 30,000 units a day, 200,000 units is 6.7 days of the factory's entire output, and that assumes every unit made in those days goes to this order and nothing else does. No factory runs that way. If this promotion took even 60 percent of output, the order occupies the line for about eleven days. That 60 percent is my assumption; the source does not break the split. Eleven days of production is not a crisis. The problem is that those eleven days start when the purchase order lands, and the promotion date does not move.
Now the modes, in days. Ocean is two months, so from July 2 the cargo lands in early September, after the heat. Rail is 25 days, which is the end of July, still inside the window. Air is days, but not as few as the sales team tends to assume, because these machines carry a refrigerant charge and cannot be tendered as ordinary general cargo. That does not make air impossible; it makes air a two-step process, where the handling arrangement is settled before the aircraft is booked, and that step has its own clock. Everything in this story that looks like a capacity problem is really a sequencing problem.
The straight comparison belongs on a whiteboard. Ocean: 60 days, arrives after the season, cheapest per unit. Rail: 25 days, arrives with three to five weeks of heat left, middle cost. Air: five to seven days door to door if the refrigerant handling was arranged in advance, ten to fourteen if it was not, and those ranges are my working assumption rather than published figures. Only two of the three land while the product is still worth full price, and one of those two costs a fraction of the other.
Who gets squeezed depends on where they sit in the calendar, not on how big they are. A brand with a retail commitment to a specific promotion date has no choice: it flies, it pays, and the refrigerant handling becomes a fixed cost of doing that promotion. A mid-sized seller with no fixed date should be on rail, and most of them are not, because rail is neither the fastest nor the cheapest and therefore loses both arguments in a meeting. A small seller who booked ocean in June is competing with 200,000 units arriving in September, sells at a discount, or holds until next June and pays eleven months of storage.
The importer and the exporter are having different emergencies. The French retailer's emergency is at the dock: 200,000 units arriving into a distribution network within weeks means each regional warehouse takes tens of thousands, each store takes hundreds, and a store backroom holds dozens. The chaos in the story is that queue, and it shows up at the store rather than at the port. The Chinese exporter's emergency is earlier and quieter: line time, refrigerant handling, and a booking that has to be confirmed before the goods exist.
The point the coverage leaves out is that a point load does not stress the lane, it stresses the receiving end. Two hundred thousand units on one promotion date is roughly a national retail distribution network's throughput for a category, compressed into a window. You can fly every unit perfectly and still fail, because the constraint stops being aircraft and becomes dock doors, appointment slots and backroom space. Which means the useful question is not whether you can get space. It is whether your customer can take delivery, and that answer was decided weeks before the first booking.
Second thing nobody says out loud: the 30,000-a-day figure is a nameplate number. It is what the line can do on a good day with every shift running and no component shortage. A promotion landing on top of existing orders does not get 30,000 a day; it gets whatever is left over. If you are the buyer, the question is not capacity but priority: where does my order sit in the queue when two other customers are chasing the same heat.
Arithmetic worth doing, assumptions stated. Take 200,000 units and assume a split system ships as two cartons, indoor and outdoor, at 0.35 cubic metres together; the packaging data is my assumption, the story gives none. Total volume is 70,000 cubic metres. At 68 cubic metres per forty-foot container that is over a thousand containers if the whole promotion went by sea. Nobody moves a single promotion in a thousand boxes, which is exactly why this one got split across modes, and why part of it arrived or will arrive late.
The trap is measuring a promotion by units sold rather than units delivered on the day. A retailer that sells 200,000 and hands over 140,000 has a customer service problem that outlasts the heatwave. The second trap is treating rail as the slow option. At 25 days it is the only mode that lands inside a six-week heat window at a cost the product can actually carry, and it is the one everybody skips because it wins neither the speed argument nor the price argument on its own.
So the decision order is: fix the date, pick the mode that hits it, settle the refrigerant handling, then book. Most teams run it backwards and discover the handling requirement at the counter. Get the units into the country while the country is still hot; a September air conditioner is a discount, not a product.
↑ Back to the key points