Which Costs Move?

Two people can run the exact same business and one of them goes broke. The difference is rarely the price they charge or how hard they work. It is which of their costs move when volume moves, and which ones sit there regardless.

Start with a smoothie stand, because the whole idea fits inside one. You have two ways to run it. Option A keeps commitments small: a pop-up permit, a rented blender, about $200 a week you pay no matter what, and then about $5 of labor, ingredients and card fees riding on every cup. Option B signs a lease, leases equipment, puts staff on scheduled shifts, and carries $960 a week before the first cup is poured. In exchange the ingredients come cheaper, so each cup costs about $2.10. At 60 cups a week, A wins by a mile, $500 against roughly $1,345, because B is paying for a lease and a batch of ingredients it cannot use. At 280 cups, B is cheaper. Around 262 cups the two tie, and that crossover is the one number in the story that tells you what to do. Expect volume above it and commit. Expect volume below it, or have no idea, and stay flexible.

So a cost is a decision about how you want to grow. Rent, a lease, a salaried employee, a piece of equipment: these are fixed costs, and they are a hole you dig at the start of every month. Ingredients, hourly labor, shipping, card fees are variable, and they climb with every unit. Neither shape is better. High fixed costs mean you need volume to climb out of the hole, and once you are out, every extra unit drops almost entirely to profit. High variable costs mean you are never deep in a hole, but every dollar of revenue drags cost along with it, and margin is hard to keep.

The second idea is that one profit number has to answer two different questions, and each question needs the costs cut differently. “How did we do last quarter?” is answered by the income statement, which subtracts costs by what they were for: the cost of the product first, then selling and administration, then depreciation, interest and tax, each subtotal answering something a specific person cares about. “What happens if we grow?” is answered by reclassifying the same costs by how they behave. Out of that second cut comes the contribution margin, price minus variable cost, which is what one more sale actually puts in your hands. And break-even, which is just fixed cost divided by contribution: how many sales dig you out of the hole.

Near break-even, profit is extraordinarily sensitive to volume. A 100-seat theater that needs 60 tickets to cover the night’s fixed cost makes nothing at seat 60, and the 61st ticket takes it from zero to a full margin. The 62nd doubles profit. By seat 95 another ticket is welcome and barely moves the percentage. That is operating leverage, and it runs in both directions. On the way up it feels like magic. On the way down you carry the full cost of capacity you cannot easily undo, and profit falls faster than sales. The judgment tool is the margin of safety: how far the volume you expect sits above the volume you need.

The third idea turns the definitions into decisions. There is no universally correct way to classify a cost. The decision in front of you picks the classification. Ask what changes if you say yes and what changes if you say no; anything that is the same either way is noise, including money already spent, which is sunk and does not get a vote. And “fixed” is always fixed with respect to a time horizon. Payroll is fixed next month and negotiable over a decade.

So here is the caterer’s phone call. She charges $50 a head, pays about $30 a plate in food, disposables and staff time, and carries $4,000 a month for a kitchen that is already paid for. A corporate client wants to add 50 guests to next month’s event at $45 a head instead of $50. Holding the price feels like the right instinct. Which costs actually change if she says yes? That question is the whole week.

In class this week

The cost module runs across three sessions and one case, and the smoothie stand above is the reading that opened it. Tuesday takes up two perspectives on profit and loss. The anchor is two friends with lemonade stands on the same Saturday: Ava rents a $150 cart and can charge $1.50 a cup, Ben sets up a folding table for nothing and can only charge $1.00, and both pay $0.50 a cup. Ava makes twice as much on every cup but starts $150 in the hole. Ben breaks even on his first cup. Which would you rather be? It depends entirely on how many cups Saturday brings, and the two tie at 300. The sticking point is that income statement lines do not sort themselves into fixed and variable. Cost of goods sold has a salaried lemonade-mixer inside it; selling expense has commission inside it. You have to take each line and ask how it behaves.

Thursday is profit-informed operating decisions, where the tools stop being definitions. We take the caterer’s call and a handful of other decisions, a discounted order, a venue choice, a subscription model that loses money per customer, and run each one: which costs are relevant, what the contribution is, what has to be true for it to work, and how much risk the move carries. Bring a calculator.

The stories

1. Will a Turnaround Plan Revive Volkswagen or Merely Delay Its Demise? The New York Times, September 5, 2026

Volkswagen won union approval this week for the largest restructuring in its 89-year history: about 50,000 more job cuts, bringing the total to roughly 100,000 by 2030, half the number of models, and production capacity cut by more than 500,000 vehicles a year, to 9 million against a pre-pandemic target of 12 million. Chief executive Oliver Blume estimates the company’s costs run 30 percent higher than rivals’. Four German factories have no clear future beyond 2030, and one idea on the table is converting them to defense production; a Citi analyst puts the chance of saving all that capacity that way at zero. Workers hold half the seats on the supervisory board, and the deal came only after management threatened to go around it.

Read this one three times, once per lesson. It is a cost-structure story: VW built for 12 million cars and is selling far fewer, so the fixed cost of plants, people and model lineups is spread over too little volume, which is what makes its unit costs 30 percent above competitors who scaled to their actual demand. It is an operating leverage story on the way down: the fixed base that made a growing VW efficient now sits there as volume shrinks, and profit falls faster than sales. And it is a time-horizon story. Whether a cost is fixed depends on how far out you are looking, and here you can watch it happen. Payroll is untouchable this month and negotiable by 2030, with a co-determination system deciding how slowly “the long run” arrives. The factories are the sunk-cost test. What they cost to build does not matter anymore; what they cost to keep running from here is the only relevant number.

Read it

2. David Ellison Is Promising at Least 30 Movies a Year. Hollywood Is Skeptical. The Wall Street Journal, September 2, 2026

Paramount’s chief executive has promised theater owners at least 30 theatrical releases a year once his company absorbs Warner Bros. Discovery, in written agreements that run three years. No studio has released 30 films in a year since 2007; the major-studio average this century is about 15, and Disney, which put out 30-plus a year in the 1990s, released 16 last year. The article’s explanation is the part to read. A quarter-century ago DVD sales made most films profitable, so more releases meant more money. Now DVDs are gone, streaming pays less, and budgets of $200 million to $300 million with $100 million marketing campaigns are common, so each film is a much bigger bet with a thinner cushion. The combined company will also carry nearly $80 billion of debt.

Every film is its own break-even problem: a fixed production and marketing cost that ticket sales have to dig out of before the studio makes anything. When DVDs added a second revenue stream, the contribution from each film was large enough that a mediocre box office still cleared the hole, and volume was the strategy. Take that stream away and raise the fixed cost per film, and the margin of safety on each release collapses, which is why slates shrank. Theaters are the other side of the same math. The New York Times reported on August 18 that the three big chains, about 1,230 theaters and 16,700 screens, endorsed the merger in exchange for the 30-film pledge, because a screen is a fixed cost that earns nothing on an empty night. A theater does not care about a studio’s profit per film. It cares about volume, and 30 releases is a promise about volume. The Times piece is from August 18, a little older than I usually send.

Read it

3. Ryanair Cuts Traffic Target to Reduce Exposure to Unhedged Winter Oil The Wall Street Journal, September 3, 2026

Ryanair trimmed its passenger target for the fiscal year ending March 2027 from 216 million to 214 million by cutting winter capacity, the November-to-March schedule it calls “unprofitable.” The airline says the cut will reduce winter losses by 70 million to 100 million euros. It has hedged 80 percent of the year’s jet fuel at about $67 a barrel, and with oil prices high it expects “some less well-hedged competitors” to struggle to keep flying this winter.

This is Thursday’s relevant-cost test applied to an airplane. The aircraft, the crew contracts and the airport slots do not change whether or not a January flight operates. What changes is the fuel Ryanair has not already locked in at $67, plus the fares those flights would bring in. When the unhedged fuel on a flight costs more than the tickets contribute, the flight has negative contribution, and cancelling it saves money even though the plane sits idle. Notice also what a hedge is in this language: it converts a variable cost that swings with the oil market into something close to a fixed one, and doing that better than your rivals is a cost-structure advantage that shows up precisely when volume and prices are under stress.

Read it

4. The Airline CEO Filling a Spirit-Sized Hole in the Market The Wall Street Journal, September 4, 2026

An interview with Jimmy Dempsey, who runs Frontier, the largest U.S. budget airline still standing after Spirit shut down this year. Frontier’s costs run about 40 percent below the rest of the industry, and it is now adding first-class seats and satellite Wi-Fi. The detail to sit with: two and a half years ago Frontier began blocking the middle seat in its first two rows and selling the extra room as “Upfront Plus.” The revenue from that space improved enough that it is now replacing those rows with two-by-two first-class seats. Frontier and Spirit overlapped on about 100 routes.

Blocking a middle seat is a decision about one unit. Its cost is the fare you give up on that seat; its return is what passengers will pay for the room next to it. Frontier ran that test on two rows, read the contribution per row, and only then committed to the fixed cost of new seats. That is the sequence the chapter recommends: know you are making money on the unit before you scale it. Spirit is the other half of the lesson. An ultra-low-cost structure is only an advantage if the volume shows up, and when fuel rose and customers moved upmarket, the carrier that had bet everything on cost per seat had the least room to absorb the miss.

Read it

5. Jersey Mike’s Is Winning the Sandwich War. It Still Needs to Win Over Gen Z. The Wall Street Journal, September 2, 2026. Heard on the Street

Jersey Mike’s, newly public and valued around $7.5 billion, has posted 20 straight years of same-store sales growth and $4.3 billion in systemwide sales by charging $15 to $20 for a sub, drink and chips to a customer base that is about 70 percent Gen X and Boomer and only 2 percent Gen Z. It plans to grow from about 3,300 stores to 15,000. It spent about 1 percent of its marketing budget on social media in 2025, against 10 to 25 percent at peers. Its 12.5 million loyalty members visit about three times as often as other customers. It has tested a $10.99 “Boardwalk Bundle,” and the columnist warns that leaning on discounts is how Subway eroded its own position.

Thursday’s chapter runs a unit-economics example where a customer worth $15 a month costs $40 to acquire and only pays off if they stay three months, then asks which lever to pull: acquisition cost, retention, price, or variable cost. Jersey Mike’s is choosing. The loyalty program is the retention lever, and a three-times visit frequency is the number that justifies it. The shift of ad dollars from television to TikTok is a bet that a different channel acquires a younger customer more cheaply. And the bundle is the caterer’s dilemma at chain scale. A $10.99 sandwich still carries positive contribution, but if the discount leaks into what the $15-to-$20 customer expects to pay, the deal costs more than the plates it was meant to fill.

Read it

One question to carry

Take any story above and answer two things. Which costs in it move when volume moves, and which sit there regardless? Then find the one decision in the story that only makes sense once you know the difference.

For my Oberlin students

Several of these sit behind subscriptions the college already pays for, so there is no reason to pay twice. Activate your access once and the links above just open.

• New York Times: register at nytimes.com/grouppass. You have to be on campus the first time, and it renews yearly.
• Wall Street Journal: register at wsj.com/oberlin. Works off campus and covers the last four years.
• Harvard Business Review: open it through EBSCO Business Source Complete and search the title - HBR in Business Source Complete.
• MIT Sloan Management Review and California Management Review: same database, search the title and authors.
• Forbes: search the title and author in Factiva through the library databases.
• The Economist: the library’s Economist archive, through the databases page.
• McKinsey: free on the open web; a free McKinsey account unlocks the PDF.

If anything here hits a paywall anyway, email me and I will get you a copy.