Lessons · Lesson 5 of 6
The split that walked off two cliffs at once
A correct decision to place a continuity programme in two origins, taken correctly, that cost five times what the same hedge cost in a different shape.
Lesson 5 of 6 · 18 min
A rule with a good reason behind it
Ingrid Halvarn has one standing instruction on continuity lines: never place the whole of one in a single country. It is not a slogan. Keeping a second origin working is a hedge: you pay a little more all the time so that one country's problem cannot stop the line. In the six years Osbern has sold BP-140, one drop arrived late enough to need air freight to protect the shelf, and that cost 22,000. Spread over the record, keeping a second origin working is worth about 3,666.67 a year.
So in February the programme was split. Egypt takes 60 of every hundred pieces, Bangladesh takes 40: 144,000 and 96,000 units a year, across the same four drops. Amriya was told, Rupsha was told, both agreed, and the costing sheet was updated with the two landed costs from lesson 1.
Every decision in that paragraph is defensible. The outcome was 33,946.62 a year, against a hedge worth 3,666.67.
Three thresholds, none of them in the costing sheet
A costing sheet holds prices. A file holds thresholds, and a threshold is a cliff rather than a slope: the number does not drift as volume moves, it jumps at a line and then sits still. There were three in this programme.
Rupsha's price break. The 4.10 Marek was quoting is Rupsha's price at 50,000 units or more per order. Below that it is 4.29. Splitting by percentage gave Rupsha 24,000 an order, so the quote that made Bangladesh attractive stopped applying the day the split was agreed. At 4.29, Bangladesh lands at 5.0518 rather than 4.8376 — 0.3213 a unit above Egypt instead of 0.1071.
Nihal's minimum shade run. The qualifying piqué that makes Amriya's goods originate is knitted to order, and Nihal will not put a shade on the machine below 3,500 kg. BP-140 runs in five colours: navy 34%, white 27%, black 19%, olive 12%, rust 8%. At 0.245 kg a garment:
| Volume placed in Egypt | Rust units | Rust cloth, kg | Nihal will run it |
|---|---|---|---|
| 240,000 | 19,200 | 4,704 | yes |
| 144,000 | 11,520 | 2,822.4 | no |
So Amriya bought rust piqué from its cheap non-qualifying source, which is the sensible thing for a factory to do when a mill declines a run. Those 11,520 units stopped originating in Egypt and were declared at the general rate, at 4.8817 landed instead of 4.7305 — the 0.1512 from lesson 2, running backwards.
The entry. One style now arrives in the same container as both a qualifying and a non-qualifying good, so every customs entry has to be split. Osbern's broker charges 340 for it, four times a year: 1,360.00.
Two complaints, one cause
In March the costing desk raised a ticket: Amriya's landed cost has gone up and nobody changed the FOB. In April the compliance desk raised a different ticket: one colour of BP-140 is being declared at the general rate although Amriya holds duty-free access.
They went to two people and were investigated separately, and neither found anything wrong at Amriya, because there was nothing wrong at Amriya. Both tickets are the February split, arriving six and eight weeks later wearing different clothes.
What the same hedge costs in a different shape
The instruction Ingrid gave was keep a second origin working. It said nothing about how. Compare two shapes of the same instruction, against a baseline of the whole programme in Egypt at 1,135,320.00 a year.
| Split every drop 60/40 | Give Bangladesh one whole drop | |
|---|---|---|
| Egypt a year, units | 144,000 | 180,000 |
| Bangladesh a year, units | 96,000 | 60,000 |
| Bangladesh units per order | 24,000 | 60,000 |
| Rupsha's price applying | 4.29 | 4.10 |
| Rust shade at Nihal | below the minimum | runs |
| Extra cost of Bangladesh volume | 30,844.80 | 6,426.00 |
| Extra cost of the lost preference | 1,741.82 | 0 |
| Broker entry splitting | 1,360.00 | 0 |
| Extra a year over the baseline | 33,946.62 | 6,426.00 |
5.28 times, for the same instruction, the same two factories and a very similar quantity of insurance. The difference is entirely that one shape crosses two thresholds and the other clears both. It clears the tighter of them by 28 kg, because at 180,000 units the rust shade is 14,400 pieces and 3,528 kg against a minimum of 3,500.
Write that 28 down. It is the most fragile number in the file, and next season's colour mix will move it.
The ending that is not tidy
Even the good shape is not free. 6,426.00 a year against a hedge measured at 3,666.67 means Osbern is paying 1.75 times what its own six-year record says it is worth.
That is a judgement and it should be recorded as one. You pay it if you believe the next disruption is larger than the last, which is a reasonable belief and not an arithmetic one — six years is a thin record, and the thing you are insuring against is rare by definition. What is not acceptable is the version where nobody computes the price at all, calls it prudence, and pays 9.26 times instead.
Check yourselfWhy did the percentage split break a shade minimum while the whole-drop split did not, when both moved a similar share of the volume away from Egypt?Show the answer
Because a percentage split shrinks every colour at once, and the smallest colour is closest to the floor. Rust is 8% of the programme, so it is the shade with the least headroom, and taking 40% out of the Egyptian volume took it under. The whole-drop split leaves Egypt with 75% of the volume across all five colours, which keeps even rust above the line — narrowly. The size of a split is what a spreadsheet shows you. The shape of it is what decides which thresholds you cross.
Course 27.1 covers the vendor-base side of this — why you keep an alternative supplier at all, and what retaining one is worth. What this lesson adds is that the value of a second origin depends on the shape you buy it in, and that the shape is decided by thresholds nobody prints in a quotation.