Supply Chain Cases 2026: Costing by Driver, Worked

The one published supply chain case whose arithmetic is complete

Firms publish cases, and almost all of them hide the numbers in an exhibit image. Oliver Wyman's Urban Grocers is the exception we found, and it is worth working through in full because it is the only firm-published chain case whose cost arithmetic is entirely in text.

The client "runs a successful grocery and fresh produce business in the US" and is "considering changing their fleet from 53- to 58-foot-long trucks, as this will increase the volume of goods that can be delivered for potential sale", checked 30 September 2026 at oliverwyman.com. Volume rises: "The larger trailers would carry 10% more volume with the same number of trips".

Then comes the sentence that is the entire lesson of this case type, and the firm states it as a note rather than burying it: "They do not expect costs to rise in a 1:1 proportion, because different types of costs are expected to change in different ways".

That is the instruction. Four cost lines are given, each with its own driver named, and the work is to apply the right percentage to the right line.

Four lines, four drivers, one total

Here is the published cost data, with the driver each line follows.

Recomputing each line in the order the case poses it: $700,000 times 8 percent is $56,000. $500,000 times 10 percent is $50,000, and the driver line adds nothing because the trip count is unchanged. The lease and maintenance step costs are $80,000 plus $15,000, so $95,000. $60,000 times 5 percent is $3,000.

The published total is "$56,000 + $50,000 + $95,000 + $3,000 = $204,000". It reconciles.

Read the driver column rather than the cost column and the structure of the answer appears. Driver labor is the largest line on the page and it does not move, because the case is about trailer length rather than trip count. Loader labor is smaller and it moves in full proportion, because loaders handle volume. Getting that pair the right way round is most of the credit available.

What the recommendation turns on

The revenue side is short. The case says to "Assume Urban Grocers trucks deliver $8,000,000 of revenue generating product per year" and that "the marginal profit on these goods is 30%".

So a 10 percent volume rise is $8,000,000 times 10 percent, which is $800,000 of additional revenue. At 30 percent marginal profit that is $240,000. Against $204,000 of additional cost, the deal earns $36,000 a year.

Now compute the ratio the case does not print. The four cost lines it names total $700,000 plus $500,000 plus $900,000 plus $60,000, so $2,160,000. The $204,000 increase is 9.44 percent of that base, against a 10 percent volume increase.

State the comparison carefully, because it is the point of the page. Costs rose 9.4 percent for 10 percent more volume. Nothing in this cost base was idle, so almost none of the extra volume fell through to profit. That is the absence of operating leverage, measured on a firm's own published case, and it is why a $8,000,000 revenue line produces $36,000.

Note the caveat honestly: the case does not state current lease cost or overhead, so $2,160,000 is a floor and the true base is larger. A larger base makes the percentage smaller and the conclusion stronger.

Solving a supply chain case backwards for the margin of safety

The published recommendation says the deal is "marginally profitable on paper" and advises against it. Turn that judgment into a number, which is what an interviewer is waiting for.

Three steps, in order.

1. The bar. Additional gross profit has to equal $204,000 for the deal to reach zero.
2. The revenue that produces it. At a 30 percent marginal profit, $204,000 divided by 0.30 is $680,000 of additional revenue.
3. The volume that produces that. On an $8,000,000 base, $680,000 is an 8.5 percent uplift.

So the assumed uplift is 10 percent and the break-even uplift is 8.5 percent. One and a half points of the ten is the entire cushion.

Two more ratios, and keep their labels apart. The $36,000 kept out of $240,000 of incremental gross profit is 15.0 percent, which is the share of the new gross profit that survives the new cost. The return on the incremental cost is a different figure: $36,000 over $204,000, which is 17.6 percent. Both are correct and they answer different questions, so name the denominator whenever you quote either.

The case then names the risk that eats exactly that cushion, and it is a physical constraint rather than a financial one: "Some products are highly perishable and may go stale before being sold, especially since the shelf space in stores is not changing, nor is warehouse space". The trucks carry 10 percent more. The shelves hold the same. Whatever share of the extra produce spoils comes straight out of a 1.5 point cushion.

That is the branch to raise unprompted. A volume gain in a chain is only worth anything at the point where the chain ends, and here the chain ends at a shelf that did not get bigger.

The supply chain input price nobody in the case controls

One more sensitivity belongs in the recommendation, and the case leaves it out. The fuel line rises 8 percent because the trucks are heavier. Meanwhile the price of the fuel itself moves for reasons that have nothing to do with trailer length.

The Energy Information Administration puts the US average retail price of No 2 diesel at $3.748 a gallon in September 2025 and $6.291 in September 2026, checked 30 September 2026 at eia.gov. That is a factor of 1.6785, a rise of 67.8 percent.

Apply it to the same $700,000 fuel line and the increase is $474,600. Against $36,000 of annual profit from the trailer decision, the fuel price move is 13.2 times as large.

Word the comparison precisely, because it is easy to overstate. The decision adds 8 percent to the fuel line. The market added 67.8 percent to the same line in twelve months, which is roughly eight times further. The dollar swing from the market is more than thirteen times the profit the decision produces.

What that licenses in a recommendation is one sentence. A project whose entire return is smaller than the annual noise in one of its own cost lines needs a hedge or a pass, and the arithmetic says which.

Firms disagree on whether a supply chain case has numbers

Prepare for both versions, because two major firms publish chain cases with opposite conventions.

Deloitte publishes one where a general contractor is behind schedule because "A project concrete subcontractor has not been able to pour concrete due to poor weather conditions" and "a project plastics supplier has not been able to deliver piping due to supply chain impacts", checked 30 September 2026 at usrecruiting.deloitte.com. It contains three questions and not one arithmetic operation. The suggested answers ask which stakeholders to involve, which documents to request, and which schedule remedies exist, listing options such as "Increasing manpower Increasing # of subs Increasing labor hours / second shifts / night shifts Resequencing schedule activities".

So on the same family of problem, one firm publishes a full cost model and the other publishes a stakeholder and evidence exercise. Both are real published cases from large firms.

The practical move is to open by asking what the client needs decided. If it is whether a change pays, you are in a cost model. If it is how to recover a slipping commitment, you are in a sequencing and contract problem where the numbers arrive later. Guessing wrong costs the first three minutes.

Context on slack is worth one line here. The Federal Reserve's release of 18 September 2026 reports capacity utilization at 76.3 percent for total industry and 75.7 percent for manufacturing, against long-run averages of 79.4 and 78.2 percent, published at federalreserve.gov. Aggregate slack exists in the economy. It did not exist in Urban Grocers' cost base, which is why its volume gain bought so little.

One label that slips, and why it matters

The case's arithmetic is correct throughout. One label is not, and pointing at it teaches something no correct page can.

The revenue step reads: "With a 10% increase in volume, we see that each truck would generate an additional $8,000,000 x 10% = $800,000 revenue per year". The assumption it draws on is a fleet figure, since the case says "Urban Grocers trucks deliver $8,000,000 of revenue generating product per year". The final recommendation returns to the fleet reading, describing "an additional $36,000 total for Urban Grocers annually".

So the multiplication is right and the words "each truck" attach a fleet base to a single vehicle. We quote the sentence as published rather than tidying it.

The lesson generalizes past this case. The failure mode on chain arithmetic is rarely the multiplication. It is the label: per truck against per fleet, per trip against per year, per pallet against per case. Check the label after every product, because a wrong label produces a plausible total and a false conclusion.

Where the points go on a chain case

The scheme that follows belongs to us. The dimensions and weights are read from the grading code in src/lib/grading, and since no consulting firm publishes a marking scheme, nothing here is any firm's official standard.

What the published record does not contain

We looked for a first-hand candidate account of a supply chain case and found none we can cite. Reddit refuses connections from this network, and no video transcript tooling was available in the session. Oliver Wyman publishes two consultant profiles and neither mentions its practice cases. Deloitte publishes a tips video and no account of a specific case. Competitor preparation blogs cover chain cases heavily and are not citable sources.

Four gaps matter. Urban Grocers states four cost lines and says overhead does not change without stating it, so the $2,160,000 base is a floor and every percentage computed on it is conservative. The case names spoilage as decisive and never prices it, so this page does not price it either. The Deloitte delay case carries no figures at all, which is the nature of that case rather than a limit of our reading. And no firm publishes how often a chain case is quantitative, so we describe the two published conventions and stop there.