Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- Pull every week 38 and week 39 quote for Shanghai to Rotterdam into one sheet by 9 September 2026, with identical scope columns for origin charges, destination charges, free time and allocation type, and calculate the spread to the dollar before any negotiation starts.
- Split September volume so no single carrier holds more than 40% and at least 30% sits on the lowest all-in quote by 12 September 2026; the target is a blended rate within 10% of the cheapest quote in the set, which on this week's numbers means under 3,982 dollars per FEU.
- Re-quote every Monday and Thursday from 9 September 2026 and log the week-on-week change per carrier; trigger a re-tender the moment any single carrier moves more than 5% in one week, which would have fired on ONE's 12.5% and CMA CGM's 10 to 11% moves.
- Fix 60% of October volume before 18 September 2026 on a named-sailing allocation with a written surcharge list and stated free time at both ends, and keep 40% floating so the position can still be adjusted after the holiday push clears.
- Reconcile every online quote against the actual invoice on the first three bookings after 20 September 2026; target variance under 2%, and put any line item you cannot map into writing with the carrier before the second sailing.
Weekly online quotes for Week 38/39 show a fragmented Asia-Europe market. Maersk cut its Shanghai-Rotterdam rate to $3,620/FEU (-$100 week-on-week), while CMA CGM lists $4,425-$4,475 for early September and $4,925 for the second half, and ONE quotes $4,008 rising to $4,508. HMM sits at $3,930. With pre-holiday 'double-festival' booking pressure, shippers see sharper week-to-week repricing and should lock allocations early.
The Analysis
Maersk took Shanghai to Rotterdam down to 3,620 dollars per FEU last week, a cut of 100. That is the sentence the market read, and it is 2.7%. On the same port pair, in the same week, the top quote in the set is 4,925. The spread is 1,305 dollars, or 36% of the lowest quote. One of those numbers is worth a negotiation. The other is worth a headline.
The coverage called the week mixed. Mixed is generous. Two carriers moved up by double digits inside the same fortnight, one moved down by less than three percent, and the fourth is quoted without a week-on-week comparison. A market where the dispersion is thirteen times the direction is not a market with a direction. It is a market with a distribution, and you buy from the distribution, not from the average.
So let me lay the quotes out on one line and read them the way I read a rate sheet: what is included, what is the spread, and what a month of volume actually costs.
Here is the set as it printed for weeks 38 and 39. Maersk at 3,620 per FEU, down 100, which puts it at 3,720 the week before, so the cut is 2.7%. HMM at 3,930, which is 310 above Maersk, or 8.6%. ONE moved from 4,008 to 4,508, which is 500, or 12.5%, and it moved inside a fortnight. CMA CGM listed 4,425 to 4,475 for the first half of September and 4,925 for the second half, a rise of 450 to 500, somewhere between 10.1% and 11.3%. Two of the four carry explicit upward moves and both are double-digit. The one that went down went down by an amount four and a half times smaller than the smallest upward step in the same set.
The unit question has to come next, because this is where the 3,620 gets misused. The same week's SCFI reading puts Far East to North Europe at 2,643 per TEU. Halve the Maersk quote to put it on a twenty-foot basis and you get 1,810. That is 833 below the index, or 31.5% below it. Two numbers for roughly the same port pair in the same week, 31.5% apart. I am not going to pretend one of them is wrong, because both can be right. An online quote and an index measure different things.
The quote is a bookable price for a defined sailing with a defined allocation and a scope of charges you have to read line by line. The index is a weighted basket of claimed rates for a week. The halving is itself crude, since a forty-foot box is not two twenty-foot boxes and has never been priced as if it were. What survives all of that is the warning: you cannot put 3,620 into a cost model until you know which charges sit inside it, and you cannot use 2,643 to check it.
The dispersion is not noise. It is inventory. Every quote here is attached to specific sailings, and what each carrier is selling this week is whatever space it has left on those sailings before the holiday push. A carrier with a light book on a late-September sailing cuts to fill it. A carrier that is nearly full on the same sailing has no reason to cut and every reason to test 4,925. Neither is telling you what the market is doing. They are telling you what their own remaining allocation is doing. That is why the spread widens exactly when the calendar tightens: the ones who are short of cargo discount harder, the ones who are long hold. The average of the set is a number no carrier quoted and no shipper paid.
Who gets hurt depends on which number you are standing on. A small shipper without a contract who books late is buying the top of the range, and the gap between booking in the first half of September and booking in the second half at ONE alone is 500 per FEU. A shipper holding a service contract signed when the lane looked firmer, say 4,400 all in, is 780 above the lowest quote in the same week, which is 17.7% of its own contract rate. That inversion is what contracts were never designed for. A contract protects you when the market rises and quietly costs you when the market splits. A large buyer with an MQC and three carriers on a panel is the only comfortable party in this picture, because dispersion is precisely what a panel is for.
The clock matters more than the level. These are week 38 and week 39 quotes and ONE moved 500 inside that window. If your booking cycle is monthly, you are sampling a weekly variable once a month and calling the result a budget. The holiday push compresses the cycle further: allocation tightens, quotes shorten their validity, and the number shown on Monday is not always the number the desk honours on Thursday. Freight at this frequency is not a price. It is a quote with an expiry, and the expiry is the part that costs money.
Here is what the coverage missed by writing about direction. The spread on this lane is 1,305 per FEU, which is 36% of the lowest quote. The week-on-week move that made the headline is 100, which is 2.7%. The dispersion is thirteen times the direction. Spend the week negotiating direction and you have negotiated a variable one-thirteenth the size of the one left on the table. The asymmetry is worse than that ratio suggests. Wait a week, Maersk cuts another 100, you save 100 a box. Wait a week, the market follows ONE to 4,508, you pay 888 more than the 3,620 that was on the screen. Eight hundred and eighty-eight against one hundred. That is the actual bet being placed by everyone who says they will wait for the market to come to them.
Put it on volume. Assume thirty FEU a month on this lane, which is my assumption and not a figure from the source. All thirty at 3,620 is 108,600 dollars. All thirty at 4,925 is 147,750. The difference is 39,150 dollars a month, and you can capture or lose it without the market moving a point, purely on where in the range you booked. Split the month half at the low quote and half at the high and the blended rate is 4,272.50, which is 652.50 above the low, or 19,575 dollars a month. Set that next to the 3,000 dollars a month that Maersk's 100-dollar cut is worth on the same thirty boxes. The negotiation that gets you the spread is worth thirteen times the negotiation that gets you the direction.
The counterfactual is unpleasant and worth stating. Anyone who fixed October volume at ONE's 4,008 before it went to 4,508 saved 500 a box, which is 15,000 dollars a month on thirty boxes, or 45,000 over the quarter, and did it by acting on a quote rather than on a forecast. Anyone who signed at 4,400 in the belief that a contract is always cheaper is now 780 above the spot market and cannot reprice until review. Both decisions were made without a view on where the market was going. They were made on where the quotes were.
The trap this week is comparing quotes that are not the same quote. Check whether origin and destination charges are inside, whether it is a named-sailing allocation or a general promise, what free time it carries at each end, and whether the surcharge list is closed or open-ended. A 300-dollar gap disappears the moment one quote excludes a terminal handling charge and the other does not, and a 1,305-dollar gap can disappear inside four days of demurrage at destination if the box lands before your warehouse is ready. My trigger is simple. If the spread between your best and worst all-in quote for the same sailing week is above 15%, split the booking and stop shopping for a direction. Below 5%, take the cheapest and go do something useful. This week the spread is 36%, and the useful thing was not the 100 dollars.
↑ Back to the key points