I stop scoring when a candidate opens with "procurement, labor, overhead" and starts listing cuts. That is a brainstorm. A cost-reduction case is won on how you classify the cost base—fixed versus variable, direct versus indirect, addressable versus contractual, one-off versus recurring—and on whether you price the second-order effects: revenue at risk, service levels, morale, and the cash cost of getting the savings out.
After this piece you should be able to open a cost case with a structure map, rank levers by time-to-impact and reversibility, close with a quantified recommendation that includes payback and risk, and avoid the three arithmetic traps that quietly kill otherwise clean math.
What the interviewer is listening for
Cost cases sit inside the same profit equation as any profitability case: Profit = Revenue − Cost. The interviewer already knows you can name levers. The sheet rewards whether you isolate which dollars can move, on what clock, and at what collateral cost.
Firms' own practice material already treats cost as structured math. McKinsey's Electro-Light case forces a break-even from a $2 wholesale price, a $1.90 unit cost, and $40 million of fixed launch cost—contribution and fixed cost, not a shopping list of cuts. McKinsey Operations work on integrated cost scans stresses cross-functional levers over siloed line-item hacking. Bain's Accelerated Performance Transformation page publishes that zero-based budgeting can cut costs by 20 to 40 percent when every dollar must justify itself against strategy.
Ned's rule. Name the structure of the dollar before you name the lever. If you cannot say whether a cost is fixed or variable, contractual or discretionary, and run-rate or one-off, you are not ready to cut it.
Map the cost base before you cut
Ask for a breakdown early. Then force four splits on whatever the interviewer gives you. You only need one clean partition at a time; stacking all four in the first minute sounds like a framework dump.
| Split | Question it answers | Why interviewers care |
|---|---|---|
| Fixed vs variable | Does this dollar fall when volume falls? | Variable cuts scale with volume; fixed cuts need a structural move |
| Direct vs indirect | Does this dollar sit in COGS or in overhead? | Direct often moves with product redesign; indirect with org design |
| Addressable vs contractual | Can we change it this year without a breach? | Contractual savings have a clock and a legal risk |
| Recurring vs one-off | Is this a run-rate or a cash event? | Mixing the two is the fastest way to fake a P&L win |
A $50 million "procurement opportunity" that is 60 percent locked in multi-year contracts is not a $50 million year-one lever. Say that out loud. Interviewers mark the candidate who shrinks addressable spend before inventing a percentage cut.
For manufacturing and ops prompts, pair this map with the bottleneck logic in operations cases: utilisation and yield often dominate headcount theatre.
Levers by time-to-impact and reversibility
Once the base is mapped, rank initiatives. Impact matters, but so do how fast cash shows up and how hard the cut is to undo if demand recovers.
| Lever | Typical time to impact | Reversibility | Second-order risk to name |
|---|---|---|---|
| Discretionary spend freeze (travel, events, contractors) | 0–3 months | High | Morale; deferred project slip |
| Vendor renegotiation / volume consolidation | 3–9 months | Medium | Supply risk; quality; switching cost |
| Cloud / IT rate and usage rightsizing | 1–6 months | High | Latency, reliability, product velocity |
| Shift patterns / overtime / temps | 1–3 months | High | Quality, safety, attrition |
| Headcount reduction (voluntary then involuntary) | 3–12 months | Low | Service levels, revenue, morale, severance cash |
| Plant footprint / line consolidation | 12–36 months | Very low | Capacity, lead time, customer fill rate |
| Product redesign / specification change | 6–24 months | Low | Margin mix, warranty, customer acceptance |
Bain's manufacturing practice cites average manufacturing cost reductions in a 5–20 percent range when plants are diagnosed end to end—useful as a public benchmark for how large structural programs can be, not as a number you invent mid-case without data.
Mini-case 1: Northfork Metalworks (manufacturing)
Prompt. Northfork stamps and machines aluminum housings for industrial equipment. Revenue $280 million. Operating margin 5 percent versus a peer set at 9 percent. Close the four-point gap (~$11.2 million of EBIT) without stranding customers.
Cost base (invented, internally consistent). Revenue $280M, EBIT 5% = $14M, so total costs = $266M.
| Bucket | Annual cost | Notes |
|---|---|---|
| Direct materials | $126M | 45% of revenue; 40% contractual, 60% addressable |
| Direct labor | $48M | Two plants; Plant A at 62% utilisation, Plant B at 88% |
| Variable plant overhead | $22M | Energy, consumables |
| Fixed plant overhead | $36M | Depreciation, salaried maintenance, lease |
| Freight | $16M | Variable with shipments |
| SG&A | $18M | Mostly fixed |
| Total | $266M | EBIT $14M |
Structure first. Materials are the largest addressable pool ($126M × 60% = $75.6M addressable). Labour is not "cut 10%"—Plant A is under-utilised, so unit labour is inflated by fixed shift cost. Fixed plant overhead only moves with footprint or shift design.
Levers with numbers.
- Procurement consolidation. Consolidate three of five commodity alloys onto two suppliers; 8% off $50M of addressable spend → $4.0M run-rate. Switching cost: $0.6M one-time. Conservative year-one cash (half-year): ~$2.0M.
- Utilisation. Move 15% of Plant A volume to Plant B; idle one weekend shift at A. Labour and variable overhead save $3.2M run-rate. Redeployment reserve: $0.8M. Fill-rate risk: 1% of revenue ($2.8M) at 25% contribution → $0.7M EBIT at risk; haircut to ~$0.2M expected drag if execution is careful.
- Freight densification. Better load planning: 6% of $16M → $1.0M.
Recommendation. Procurement + utilisation + freight → ~$8.2M run-rate before risk, ~$8.0M after the utilisation haircut—most of the $11.2M gap in 12–18 months without a plant closure. Restructuring cash ~$1.4M; payback under three months once run-rate lands. Hold plant closure: more fixed overhead unlocked, but low reversibility and lead-time risk.
Mini-case 2: LatticeOps (B2B software / services)
Prompt. LatticeOps sells workflow software with an attached implementation services bench. ARR $90 million, services revenue $30 million. Gross margin on software 78%, on services 22%. Company wants $12 million of annual cost takeout ahead of a fundraising.
Cost base.
| Bucket | Annual cost | Nature |
|---|---|---|
| Product and engineering headcount | $38M | Mostly fixed; ~12% contractors |
| Cloud infrastructure | $14M | Mixed; 30% reserved instances (contractual), 70% on-demand |
| Customer success + support | $16M | Fixed headcount keyed to logo count |
| Sales (fully loaded) | $22M | Mix of base (fixed) and commission (variable) |
| G&A and vendors (legal, security tools, agencies) | $11M | ~$4M multi-tool overlap addressable |
| Services delivery bench | $19M | Mostly fixed utilisation problem |
| Total opex + delivery | $120M | Cost base to cut against $120M revenue |
Find $12M without torching NRR.
Structure first. Cloud is partly contractual—do not claim 20% of $14M if only $9.8M is on-demand. CS headcount cuts look large but threaten net revenue retention. A low-utilisation services bench is a capacity problem, not a blunt labour cut.
Levers with numbers.
- Cloud rightsizing. 35% of on-demand spend idle or oversized: $9.8M × 35% = $3.4M. Engineering sprint $0.4M; hold $0.3M performance contingency → ~$3.1M durable.
- Vendor consolidation. Drop two overlapping security tools and one agency: $2.2M. Switching $0.3M. Low revenue risk.
- Services bench. Utilisation 61%. Flex 20 contractors and redeploy 8 FTEs onto billable work. Cash leave $4.5M; if $1.5M returns as services contribution, net takeout ~$3.0M—state which definition you use.
- Hiring freeze + attrition in G&A and non-quota sales support: $2.0M, high reversibility.
- Reject: 15% CS cut (~$2.4M). Case assumption: every 10% CS cut loses 2 NRR points on $90M ARR → $1.8M ARR at 78% GM ≈ $1.4M gross profit at risk. Not a net save.
Recommendation. Cloud + vendors + bench flex + freeze → about $10.3–12.3M, with ~$0.7M one-time cost and payback under two months. Flag CS cuts as false economy. These moves stay mostly reversible—useful before a fundraise.
Arithmetic traps that look like savings
Three mistakes show up constantly. Drill them on /drills/math and in case math practice until they feel automatic.
Percent of a percent. "Cut 10% of the 30% of costs that are overhead" is 3% of the total base, not 10%. Say the absolute dollars.
Run-rate versus one-off. A $5M restructuring charge that unlocks $4M of annual savings is not "$9M of value." Report run-rate savings, cash cost to achieve, and months to payback separately—the same discipline as in ROI and payback cases.
Savings on a shrinking base. If volume is falling 10% and variable costs fall with it, that is not a cost program; it is the P&L breathing. Attribute only the incremental reduction beyond the volume effect. McKinsey's Electro-Light break-even logic is the same habit: hold unit cost and fixed cost distinct so volume stories do not contaminate cost stories.
How to open and close in the room
Open in under a minute: restate the cost target in dollars, ask for the cost split, propose your partition (fixed/variable or value chain), and say you will rank levers by impact, speed, and reversibility before recommending.
Close in under a minute: one number for run-rate savings, one for one-time cost, payback in months, the main risk (revenue, service, or capacity), and what you would not cut. That is what a partner can take to a client steering committee.
CoachNed is not affiliated with McKinsey, BCG, Bain, or any firm named here.
Practice this today
Take any profitability prompt and force a four-split cost map before you name a single lever. Then run one timed math set on /drills/math—percent-of-percent and run-rate versus one-off problems only. When you want a full case with a scored debrief, start at /start (three typed turns, no account) or sit the live voice case at /interview. Everything is open for seven days, no card; then $120 for a recruiting season or $49 a month.
Frequently asked questions
How do you structure a cost reduction case interview?
Restate the savings target in dollars, map the cost base with one clear partition (fixed versus variable or value chain), separate addressable from contractual spend, then rank levers by impact, time-to-impact, and reversibility while calling out revenue and service risk. Close with run-rate savings, one-time cost, and payback.
What is the difference between fixed and variable costs in a case?
Variable costs move with volume (materials, shipping, cloud usage tied to traffic). Fixed costs stay in the short run (salaried headcount, plant leases, reserved capacity). Cutting variable costs scales with volume; cutting fixed costs usually needs a structural decision and a cash cost to exit.
Should I recommend layoffs in a cost case?
Only after you have quantified addressable non-people spend and utilisation problems, and only with severance, service-level risk, and payback stated explicitly. Many cases award more points for procurement, footprint, or vendor moves that protect revenue.
How do I calculate payback on a cost initiative?
Divide one-time implementation and restructuring cash by annual run-rate savings (after a realistic risk haircut). State the answer in months. Do not add one-time and run-rate into a single "impact" figure.
Are cost reduction cases the same as operations cases?
They overlap. Operations cases often centre on a bottleneck or flow constraint; cost cases centre on the P&L. Use operations logic when utilisation or yield is the driver—see operations cost frameworks—and keep the close on dollars and payback.
Sources
- Electro-Light | McKinsey Careers — Official practice case with unit cost, fixed cost, and break-even math. Checked 2026-09-24.
- How to untap the full potential: An integrated—not isolated—view on cost | McKinsey Operations — Cross-functional cost levers and Operations Full Potential Scan framing. Checked 2026-09-24.
- Accelerated Performance Transformation | Bain & Company — Published APT outcomes and zero-based budgeting savings range (20–40%). Checked 2026-09-24.
- Manufacturing Consulting | Bain & Company — Public 5–20% average manufacturing cost reduction range. Checked 2026-09-24.
- Cleansheet | McKinsey Operations — Bottom-up product cost structure for identifying procurement and design savings. Checked 2026-09-24.