Most candidates treat unit economics as a memorized tree of revenue, cost, and "LTV over CAC." That is not what the scoring sheet rewards. In the room, the first mark is whether you named the right unit before you touched a calculator. The second is whether you proved contribution on that unit before talking about growth.
This piece stays on those two moves: unit selection and contribution math. Lifetime-value math and the broader scale-versus-profitability trade-off belong elsewhere. After you finish, you should be able to pick a unit in thirty seconds, run contribution cleanly, and catch yourself when switching the unit would reverse the answer.
What filings actually disclose
Public companies do not publish a universal "unit economics framework." They publish a unit they chose, then a contribution figure defined against that unit. The definitions differ, which is the point: the firm picks the unit that matches how the business scales.
| Company | Unit implied by the metric | What they disclosed | Source period |
|---|---|---|---|
| Uber | Core platform (rides + delivery trips as a platform) | 9% Core Platform Contribution Margin for 2018; top-five countries ranged roughly 54% to (10)% in early 2019 | S-1 / 424B4, 2019 |
| Lyft | Platform revenue after direct costs | Contribution of $920.8 million on $2.2 billion revenue (42.7% Contribution Margin) in 2018 | S-1/A, 2019 |
| DoorDash | Marketplace order | Illustrative 2019 order retains $4.90 after merchant and Dasher shares; company Contribution Profit (Loss) was $(200) million in 2019 on 263 million Total Orders | S-1, 2020 |
| Starbucks | Store (company-operated vs licensed) | Company-operated stores drove most revenue; licensed stores carry lower gross margin but higher operating margin for Starbucks because licensees fund operating cost and capex | FY2025 10-K |
Uber's own prospectus is candid that contribution margin is not profitability, and that the same platform can look healthy in one country and underwater in another. DoorDash's order diagram shows take-rate economics at the transaction; its cohort charts then shift the unit toward the consumer over time. Starbucks reports comparable store sales as transactions times ticket, which is a reminder that retail cases often need both the cup and the store.
None of those filings hand you a case structure. They show that serious operators argue in units, then in contribution, and they state which definition they are using.
Choose the unit from the decision, not the industry
I stop listening when a candidate opens with "I'll look at LTV and CAC" before clarifying what the client is deciding. Ask one question first: what would we be buying more of if we said yes?
| Client decision | Natural unit | Contribution test | When that unit is wrong |
|---|---|---|---|
| Raise prices or cut discounts | Order, ticket, ride, seat | Price minus truly variable cost | If fixed capacity is the bottleneck, move to seat or flight |
| Open another site | Store, kitchen, hub, studio | Contribution after site fixed cost at realistic utilization | If the network is the product, city or region may dominate |
| Grow marketplace volume | Transaction or order | Take rate minus payment, support, and fulfillment share | If promo spend is targeted by user cohort, the consumer unit appears |
| Fill an asset | Flight, class, charger session, hotel night | Revenue per used capacity unit vs variable cost | Averaging across empty and full days hides the real story |
| Cut COGS | SKU or shipment | Materials, labor, freight that move with volume | Do not load rent into "variable cost" unless the lease truly steps |
The unit is not a personality preference. It is the smallest repeatable chunk that still answers the decision. A coffee chain asked whether to open cafe number twelve needs the store unit even if every latte looks fine. A delivery marketplace asked whether take rates cover Dasher pay needs the order unit first. Mixing them mid-case is how candidates talk themselves into the wrong recommendation.
Managerial accounting textbooks define contribution margin the same way interviewers use it: selling price minus variable cost per unit, then volume against fixed cost for breakeven (OpenStax). Keep the definition boring. Put the creativity into naming the unit and sorting costs.
Contribution math the sheet rewards
Once the unit is named, the calculation is short. Interviewers are listening for three separations, not for fancy formulas.
- Revenue per unit. List price, mix, discounts, tips that pass through, and fees you keep. Pass-through tips are not your revenue if they leave the platform.
- Variable cost per unit. Only costs that move when that unit happens: ingredients, payment processing, per-trip insurance, packaging, delivery payouts, energy per session. Salaried managers and rent are usually fixed or step-fixed.
- Contribution, then fixed-cost coverage. Contribution = revenue − variable cost. Units needed to cover fixed cost = fixed cost ÷ contribution per unit. Utilization is how many units the asset actually produces, not the brochure capacity.
Say the numbers aloud as you build them. A clean board looks like:
| Line | Per order | Notes |
|---|---|---|
| Revenue kept | $11.00 | After promo discount |
| Variable cost | $7.60 | Food, packaging, card fees |
| Contribution | $3.40 | Do not call this profit yet |
| Site fixed cost / month | $18,000 | Lease, minimum labor, utilities |
| Breakeven orders | 5,294 | $18,000 ÷ $3.40 |
If the site currently runs 4,200 orders, you are not "almost profitable." You are short about 1,100 orders, or you need roughly $0.90 more contribution per order. That is the recommendation language the interviewer wants.
Worked example: same business, two units, opposite answers
The client is AtlasFit, a three-studio boutique fitness chain. Management wants to open a fourth studio in the same metro. Average class ticket is $32. Instructor pay tied to the class is $8 per attended seat. Supplies and payment fees are $2 per seat. Studios run 22 classes a week with 14 attended seats on average. Monthly studio fixed cost (rent, front desk, utilities, equipment lease) is $28,000.
Unit A: attended seat (the class ticket).
- Revenue: $32
- Variable cost: $8 + $2 = $10
- Contribution: $22 per seat
A weak candidate stops here. "$22 contribution is strong, so open studio four and fill seats with intro discounts."
Unit B: studio month (the decision unit).
- Seats per month: 22 classes × 14 seats × 4.3 weeks ≈ 1,324 seats
- Contribution per studio: 1,324 × $22 ≈ $29,130
- Fixed cost: $28,000
- Studio profit before overhead: about $1,130
Barely positive. Now apply the growth plan: a 30% intro discount on half the seats to "fill the new studio." Discounted seats contribute $32 × 0.7 − $10 = $12.40, not $22.
| Mix at studio 4 | Seats | Contribution each | Monthly contribution |
|---|---|---|---|
| Full price | 662 | $22.00 | $14,564 |
| Intro (30% off) | 662 | $12.40 | $8,209 |
| Total | 1,324 | — | $22,773 |
| Fixed cost | — | — | $28,000 |
| Gap | — | — | ($5,227) |
Same company. Same capacity. Change the unit from seat to studio, and "open with discounts" flips from obvious to a loss of roughly five thousand dollars a month. The right recommendation is usually: stabilize utilization and price at the existing three studios, test whether full-price demand can clear about 1,275 seats a month ($28,000 ÷ $22) at a new site, and do not buy growth that destroys contribution.
That flip is the skill. If your answer would reverse under a better unit, you have not finished the case.
Ned's rule. Name the unit in your first structure sentence, then refuse to discuss growth until contribution on that unit is positive at a utilization the client already achieves somewhere. If contribution is negative, scale multiplies the loss.
Mistakes that are really unit mistakes
Calling average cost "variable cost." Rent divided by last month's orders is not variable. It falls when volume rises and rises when volume falls. Interviewers score the distinction.
Skipping pass-throughs. DoorDash's illustrative order charges the consumer $32.90 but retains $4.90. If you treat gross merchandise as platform revenue, every take-rate case is wrong.
Forcing one unit for the whole case. Start with the decision unit. You may still check a secondary unit for a sensitivity (seat economics inside a studio case). Say when you switch.
Treating contribution margin as GAAP profit. Uber's S-1 says Core Platform Contribution Margin is useful and also that it is not a measure of profitability. Echo that honesty in the interview: contribution funds fixed cost and overhead; it is not the bottom line.
Jumping to lifetime value. Useful when the decision is about buying customers over time. Distracting when the prompt is about opening a kitchen, cutting packaging cost, or pricing a single ride. Keep CLV for the cases that earn it.
Practice the unit switch today
Take any growth prompt and force two units before you recommend. Write both contribution calculations on paper. If they disagree, the decision unit wins, and the other becomes a diagnostic.
CoachNed is built for that habit. Start with a five-minute typed case at /start (three turns, instant scores, no account required), then run the live voice case at /interview when you want an eighteen-minute debrief. For raw contribution and breakeven reps, use /drills/math. Everything is open for seven days with no card; then $120 for a recruiting season or $49 a month.
Frequently asked questions
What is unit economics in a case interview?
It is the practice of defining a repeatable unit tied to the client decision, calculating revenue and variable cost for that unit, and checking whether contribution covers the fixed cost that growth would add. It is not a branded framework you recite by name.
How do I choose the right unit quickly?
Ask what you would buy more of if you said yes: orders, customers, stores, seats, rides, shipments. Pick the smallest unit that still includes the costs the decision creates. If opening a site is on the table, the site usually beats the order.
Is contribution margin the same as profit?
No. Contribution is revenue minus variable cost. Profit also absorbs fixed cost, shared overhead, and often one-time items. Filings that highlight contribution margin usually warn that it is not GAAP profitability.
Should I always calculate LTV and CAC?
Only when the unit is a customer or account and acquisition spend is part of the decision. For factories, flights, kitchens, and most cost cases, start with contribution per operational unit and leave lifetime value alone unless the interviewer steers you there.
What does a strong recommendation sound like?
State the unit, the contribution number, the utilization or volume required to cover fixed cost, and the action: expand, pause, reprice, or cut variable cost. Mention the sensitivity that would flip the call. That is enough. CoachNed is independent and is not affiliated with McKinsey, BCG, Bain, Deloitte, Uber, Lyft, DoorDash, Starbucks, or any firm named above.
Sources
- Uber Technologies, Inc. Form S-1 — Core Platform Contribution Margin definition and 2018 figure of 9%. Checked 2026-09-24.
- Lyft, Inc. Form S-1/A — 2018 Contribution of $920.8 million and 42.7% Contribution Margin. Checked 2026-09-24.
- DoorDash, Inc. Form S-1 — Illustrative 2019 Marketplace order economics ($4.90 retained) and 2019 Contribution Profit (Loss) of $(200) million on 263 million Total Orders. Checked 2026-09-24.
- Starbucks Corporation Form 10-K (fiscal 2025) — Company-operated versus licensed store economics and store as the operating unit. Checked 2026-09-24.
- OpenStax, Principles of Managerial Accounting, §3.2 — Contribution margin and breakeven definitions used in case math. Checked 2026-09-24.