I stop listening when a candidate opens a supply-chain case with a profitability tree and never mentions days. The sheet rewards flow arithmetic: inventory days, service level, cost-to-serve, landed cost, and Little's law. Name a framework if you must. Then price the pipeline.
Most guides teach you to "think end to end." Interviewers already assume that. What separates a second-round pass from a polite no is whether you can hold service constant, rebuild unit cost into landed cost, and show how many dollars of inventory buy the next point of fill rate. After this piece you should be able to run that tradeoff on clean numbers in under five minutes.
What the interviewer is actually grading
In a typical first-round slate of eight, I expect two or three people to get the logistics frame right. The rest compare sticker price, say "lean," and move on. The scoring sheet is blunter:
| What they say | What the sheet hears |
|---|---|
| "$4.20 China beats $5.10 Mexico" | Incomplete cost object |
| "We should cut inventory" | No service constraint |
| "Dual source for resilience" | Cost of buffer never priced |
| "Landed cost is X; days of stock for Y% fill is Z" | Passable structure |
Clarify three definitions before you touch the exhibit. Fill rate is the share of ordered units (or lines, or orders—pick one and hold it) shipped from available stock. Cycle service level is the probability of no stockout in a replenishment cycle. They are related and not identical; mixing them mid-case is how people lose the plot. Inventory days is average inventory divided by average daily usage—the same idea the ASCM Supply Chain Dictionary treats as inventory turnover inverted.
If the prompt says "OTIF" or "in-stock," ask which numerator they mean. Do not invent a third metric.
Landed cost is the cost object, not the PO price
Retail filings make the point without consulting folklore. Target's fiscal 2025 10-K states that inventory cost includes what they pay suppliers, freight into DCs and stores, and import costs (tariffs, duties). Cost of sales also pulls in outbound shipping and supply-chain facility costs. The PO price is one line. The network is the rest.
Walmart's fiscal 2025 Form 10-K reports $511.8 billion of cost of sales and $56.4 billion of year-end inventories. Against the prior-year inventory balance of $54.9 billion, that is roughly nine turns, or about 40 days of inventory. Target's same-year pair—$75.5 billion cost of sales and about $12.5 billion average inventory—is closer to six turns, or about 60 days, which matches Target's own note that inventory typically turns in less than three months.
Those are retailer-scale disclosures, not interview answers. Use them as anchors: when a case invents "15 days of stock" for an ocean-sourced durable, ask whether that is DC stock only or the whole pipeline.
Freight is not a rounding error. As of 17 September 2026, Drewry's World Container Index stood at $4,500 per 40-ft container, with Shanghai–Los Angeles at $7,712 and Shanghai–New York at $10,394. Put 800 units in that FEU and the Transpacific West Coast leg alone is nearly $10 per unit before inland haul, duty, and the carrying cost of 40–50 days at sea. If your China vs nearshore slide ignores that, you are grading procurement theater.
A usable landed-cost build in a case:
Landed unit = product + freight + duty + packaging/yield loss + pipeline carrying
Pipeline carrying ≈ landed value × (pipeline days / 365) × carrying rate. Cases usually hand you 15–25% for the carrying rate; do not debate WACC unless asked.
Little's law is the bridge between days and dollars
Little's law says average inventory in a stable system equals throughput times average time in the system: L = λW, or in operations language WIP = throughput × cycle time. You do not need the 1961 proof. You need the unit check.
If HarborGlow ships 120,000 units a year and the ocean-plus-inland pipeline is 45 days, average units in transit are:
120,000 × (45 / 365) ≈ 14,800 units
At a $48 landed unit cost into the DC, that is about $710k of pipeline inventory before the first pallet hits the rack. At a 20% carrying rate, that pipeline alone costs about $142k per year. Candidates who "optimize the DC" and never price the water lose that line.
U.S. business logistics costs are large enough that interviewers expect you to treat freight, inventory, and service as one system. Per CSCMP's 2026 State of Logistics Report as covered by FreightWaves, those costs totaled $2.4 trillion, or about 7.8% of GDP, for the latest reported year (down from $2.6 trillion / 8.7% the year before).
Worked example: buying three points of fill rate
Prompt (invented). Nestline Home sells a mid-market blender through U.S. retail. Steady demand is 120,000 units/year (~329/day). Landed cost into the DC is $48. Contribution margin after landed cost and variable selling is $22. Replenishment lead time is 30 days truck from a regional plant (no ocean in this SKU). Current policy targets a 95% cycle service level. Average on-hand inventory is 22 days of demand. Carrying cost is 20%/year. Lost sales are assumed equal to unmet demand (no backorders). Management wants 98% fill on this A-SKU. How much inventory does that buy, and is it worth it?
Step 1 — Current inventory economics
On-hand units = 22 × 329 ≈ 7,240
Inventory value = 7,240 × $48 ≈ $347,500
Annual carrying = 347,500 × 0.20 ≈ $69,500
At 95% fill, unmet demand ≈ 5% × 120,000 = 6,000 units
Lost contribution ≈ 6,000 × $22 = $132,000
Step 2 — Move to 98% fill
Unmet demand falls to 2% × 120,000 = 2,400 units
Lost contribution falls to 2,400 × $22 = $52,800
Recovered contribution = 132,000 − 52,800 = $79,200
Safety stock does not rise linearly with fill rate. For interview math, use the z-score step the exhibit implies, or a stated days increase. Here the planner's model says on-hand rises from 22 to 31 days to hit 98% given lead-time demand variability (case input).
New on-hand = 31 × 329 ≈ 10,200 units
New value ≈ 10,200 × $48 ≈ $489,600
New carrying ≈ 489,600 × 0.20 ≈ $97,900
Extra carrying = 97,900 − 69,500 = $28,400
Step 3 — Decision
Net annual benefit ≈ 79,200 − 28,400 = $50,800
Recommend the 98% policy for this A-SKU. Hold C-SKUs at 92–95% unless the retail contract forces a single rate. State the assumption: if stockouts become backorders rather than lost sales, the "recovered contribution" shrinks and you need the expedite and goodwill lines instead.
| Policy | Days on hand | Carrying $/yr | Lost margin $/yr | Total of those two |
|---|---|---|---|---|
| 95% fill | 22 | 69,500 | 132,000 | 201,500 |
| 98% fill | 31 | 97,900 | 52,800 | 150,700 |
The interviewer is listening for the trade, not the slogan "higher service is better." You just bought three points of fill for about $28k of carrying and pocketed roughly $51k net on the lost-margin math.
Ned's rule. Never cut inventory without naming the service metric you are holding constant, and never raise service without pricing the days. If either number is missing, you are advocating, not analyzing.
Cost-to-serve and when the cheap lane loses
Cost-to-serve is landed cost plus the customer-specific extras: rush freight, returns, small-drop deliveries, and the inventory you park for that customer's fill contract. A national account that demands 99% fill on 400 SKUs can destroy the margin on a lane that looked fine on average unit cost.
In the room, build a simple strip:
- Base landed cost (as above).
- Inventory days required by the service clause (Little's law + safety stock).
- Expedite rate × expected expedite volume.
- Returns / reverse logistics if the exhibit shows them.
If the case is make-vs-buy or China-vs-Mexico, keep service and volume pattern fixed while you swap the landed build. Spiky demand plus a 50-day ocean pipeline is a service problem first; the carrying-cost line is the second-order effect.
How to open in sixty seconds
Say this, then ask for the exhibit:
"I would treat this as a flow problem: define the service metric, rebuild unit cost into landed cost including freight and pipeline carrying, then size inventory days for that service using Little's law on the pipeline and a safety-stock step for variability. I would separate A and C SKUs and price dual-source as an insurance cost, not a slogan."
That is enough structure. Then do the arithmetic.
Practice this today
Take any dual-source or inventory exhibit you already have and force three numbers onto one page: landed $/unit, days of stock, fill %. Change one; recompute the other two. Then run a scored math drill on CoachNed—everything is open for seven days, no card; then $120 for a recruiting season or $49 a month—at /drills/math or start a live case at /interview.
Frequently asked questions
What is a supply chain case interview testing?
Whether you can optimize landed cost, working capital, and service as one system. Interviewers score the flow math—days, fill, freight, pipeline—more than the label of your framework.
How do you calculate inventory days in a case?
Divide average inventory by average daily demand (or invert turns: 365 ÷ cost-of-sales/average inventory). State whether the figure is DC only or includes pipeline.
What is the difference between fill rate and service level?
Fill rate is the share of demand shipped from stock. Cycle service level is the chance of no stockout in a replenishment cycle. Ask which one the case uses and keep it fixed when you compare options.
How does Little's law show up in consulting cases?
Average units in the system equal throughput times time in the system. Use it to convert lead time into pipeline inventory dollars before you debate DC safety stock.
Should I always choose the lower unit-cost supplier?
No. Rebuild to landed cost with freight, duty, yield, and pipeline carrying, then check whether the lead time can hit the contracted fill rate without uneconomic expedites.
Sources
- ASCM Supply Chain Dictionary — professional definitions for inventory turnover and related metrics. Checked 2026-09-24.
- Target Corporation Item 8 financial statements (fiscal 2025) — inventory and cost-of-sales composition including freight and import costs; inventory ~$12.3B and cost of sales $75.5B. Checked 2026-09-24.
- Target Corporation Item 7 MD&A (fiscal 2025) — note that inventory typically turns in less than three months. Checked 2026-09-24.
- Walmart Inc. Form 10-K for fiscal year ended January 31, 2025 — cost of sales $511.8B and inventories $56.4B; logistics included in cost of sales description. Checked 2026-09-24.
- Drewry World Container Index assessment (17 Sep 2026) — composite $4,500/FEU; Shanghai–LA $7,712; Shanghai–NY $10,394. Checked 2026-09-24.
- FreightWaves coverage of CSCMP 2026 State of Logistics Report — U.S. business logistics costs $2.4T (7.8% of GDP). Checked 2026-09-24.
CoachNed is independent and unaffiliated with McKinsey, BCG, Bain, Deloitte, or any firm named here.