A quoted price is not a cost.

What the model is doing

Landed cost, term by term

The last line is the one people leave out. You pay for every unit that ships, including the ones quality rejects on receipt, so the cost of a unit you can actually sell is the paid cost divided by the yield. A 2.3 percent scrap rate is not a rounding error on a two-percent margin.

The carrying term prices the capital sitting in the pipeline. Eleven weeks of inventory on the water is eleven weeks of money you cannot use, and at 18 percent a year that is about four cents on an eleven-dollar unit. Small, but it moves in the same direction as everything else that makes a long supply line expensive, and the whole point of the exercise is that those small terms add up to the entire apparent price advantage.

The five suppliers, and the trap in the table

At the rates the planning horizon opens with, the quotes and the landed costs rank almost in opposite orders:

Pearl River looks a third cheaper than Bajío and lands about seventeen cents dearer. Vistula and Bajío quote within five cents of each other and land more than a dollar apart. And the supplier with the lowest landed cost in the game is one you cannot buy from at all until you spend $120,000 and wait two quarters — which is a decision, not a footnote. Whether it pays depends on the scenario, and in one of the three it clearly does not.

Why lead time is quantised

Lead times are quoted in weeks and converted to whole quarters: two weeks or less lands in the quarter you order, three to eight lands next quarter, more than eight lands in two. That is a real simplification and it is the most consequential one in the model. It is what turns the cheap supplier from a purchase into a bet: an order placed with Pearl River in quarter 4 serves demand in quarter 6, against a forecast made in quarter 4 that is wrong by a few percent in a direction you will not learn for six months.

It also means the last two quarters are structurally different. An order that cannot land before the planning horizon closes is not allowed, so the long-lead suppliers simply disappear from the table in quarters 7 and 8, and the expensive domestic supplier is all you have left. That end-of-horizon effect is an artefact of the game having an end — but it is also exactly what a buyer faces before a plant shutdown, a model changeover, or the last quarter of a contract.

Hedging, honestly

A forward contract for the coming quarter is quoted at last quarter's spot rate plus 0.6 percent. If you cover 100 percent of a quarter and the currency then moves in your favour, you paid the premium and gave up the gain. If it moves against you, the premium was the cheapest money you spent all year. Over the eight quarters of A strong dollar, then a weak one, a policy of always hedging and a policy of never hedging end up within a fraction of a percent of each other, and the difference between them is almost entirely the accumulated premium.

That is the correct lesson and it is an unsatisfying one. Hedging does not make sourcing cheaper. It makes the cost of sourcing knowable, which is worth something to a business that has to commit to a price list, and worth nothing to a game that is scored on total spend. If you find yourself trying to time the hedge, notice that you are being asked to forecast an exchange rate, and that nobody can.

The three scenarios

• The tariff year. Currencies drift, everybody ships what they promised, and one trade measure lands: duty on Chinese goods rises from 22 to 35.5 percent, announced in quarter 5 and in force from quarter 6. The sting is that a two-quarter lead time means orders placed before the announcement are assessed at the new rate when they arrive. This is the scenario where qualifying the duty-free alternative pays, if you start early enough.
• A strong dollar, then a weak one. No tariff changes, no supply problems, and every foreign currency falls about fifteen percent and comes back. Nothing about the suppliers changes; the entire movement in your cost base comes from outside the supply chain.
• The supplier that stumbles. Pearl River delivers 62 and 71 percent of its commitments in quarters 5 and 6. The order that fails was committed in quarter 4, before there was anything to see. This is the scenario that prices concentration: the concentration index sits in the analytics all year doing nothing, and then it explains everything.

The two benchmarks

Follow the cheapest quote buys from whoever quotes lowest at the opening exchange rate, and never reconsiders. It is the policy the whole exercise argues against, and it is worth noticing how close it comes — between half a percent and a bit over two percent dearer, depending on the scenario. On a four-million-dollar spend that is $28,000 to $94,000. Sourcing decisions are not usually rescued or ruined by a single insight; they are ground out in fractions of a percent, and a game that showed a naive buyer losing twenty percent would be lying to you.

Lowest landed cost, with hindsight ranks by the cost that turns out to apply on arrival, knowing every exchange rate, duty rate and delivery failure in advance, carries a four percent buffer against the losses it can see coming, and qualifies the locked supplier when that turns out to pay. Nobody has that information. It is the bar a perfectly informed buyer reaches with a rule this simple, and the gap between it and the naive policy is the value of looking at landed cost at all.

Assumptions, stated plainly

• No fixed cost of supplier count. Managing five relationships costs the same here as managing one, which is false and which pushes the model toward fragmentation. The concentration index in the analytics exists to keep that question visible even though the cost function ignores it.
• No volume discounts, no contract tiers, no renegotiation. Prices are posted and take it or leave it.
• Payment on arrival. No terms, no letters of credit, no working-capital timing beyond the pipeline carrying charge.
• Quality is a scrap rate. No warranty tail, no recall, no customer discovering the defect after you shipped it.
• Everything is deterministic. Exchange rates, duty changes and delivery failures follow a fixed path per scenario. Two players who make the same calls get the same result, which is what makes them comparable — and which also means you should not read the ability to do well here as evidence that anyone can forecast a currency.
• One part number. No mix, no shared tooling, no minimum order quantity beyond quarterly capacity.

Where this sits in the practice

Total cost of ownership, landed cost modelling, and supply base concentration are the standard frames here, and the arithmetic in this game is the arithmetic in a procurement TCO template — quote, freight, duty, quality, and carrying, with the last two most often missing. What a spreadsheet cannot show is the sequencing: that the commitment is made two quarters before the cost is known, that the duty rate is set by someone else after you have signed, and that the supplier you would most like to use is the one you have not qualified. Those are the reasons sourcing is a planning problem and not a pricing exercise.