Cost structures: fixed, variable, marginal and unit cost

When a deal lands on your desk at an unexpected price, the real question isn't "is this price good?", it's "does it clear my costs?" The same £15 sale is a gift to one business and a slow death to another, and the difference is the cost structure.

The idea in depth

Start with the cut that organises everything else: does a cost change when you make or sell one more thing? Fixed costs stay put across a range of activity, the office lease, your core engineering salaries, the annual software licence. They are owed whether you sell nothing or sell out. Variable costs move with volume, raw materials, the card-processing fee on each order, the cloud bill that ticks up with usage. The line isn't moral or permanent (a "fixed" salary becomes variable the day you hire per-order contractors); it's about behaviour within the range you're operating in.

From that distinction fall the two numbers leaders most often muddle. Marginal cost is the change in total cost from producing one more unit, in most short-run cases just its variable cost, because the fixed costs are already paid. Unit cost (average cost) is total cost divided by units made, so it carries a slice of the fixed costs and shrinks as you spread them over more units. Marginal cost asks "what does the next one cost me?"; unit cost asks "what does the average one cost me?", and only the second tells you whether the business as a whole is viable.

flowchart TD
  T(["Total cost of running the business"]) --> F(["Fixed costs, don't move with volume"])
  T --> V(["Variable costs, rise per unit"])
  V --> M(["Marginal cost = cost of one MORE unit"])
  F --> U(["Unit cost = (fixed + variable) ÷ units"])
  V --> U
  M --> Q1(["Sets your floor price"])
  U --> Q2(["Sets your sustainable price"])
					

In practice: for any product line, write down the cost that genuinely rises with each sale, that, and only that, is your marginal cost and your true price floor. Everything else is fixed cost you'll recover through volume. Knowing the gap between the two tells you how much room you have to discount a one-off deal without lying to yourself about the business underneath.

Why the fixed/variable mix is really a bet on risk

The fixed-to-variable mix has a name when you zoom out: operating leverage. AccountingTools defines it plainly, operating leverage "measures a company's fixed costs as a percentage of its total costs", and the consequence is the part leaders should feel in their stomach: a high-fixed-cost business turns small revenue swings into large profit swings, in both directions. As the same source notes, "higher fixed costs can increase risk, so companies must ensure projected sales growth justifies the investment," and businesses heavy with fixed costs are more exposed in a downturn because "they must cover fixed costs regardless of sales levels."

This is why a software company and a consultancy with identical revenue are not the same business. The software firm pours money into building the product once, then ships copies at almost no marginal cost; the consultancy's costs are mostly the people doing the work, so they scale up and down with the workload. Carl Shapiro and Hal Varian made this the defining feature of information goods in Information Rules (1999): producing the first copy involves "high fixed costs but low marginal costs," and every copy after is "negligible." Their conclusion follows directly, "cost-based pricing just doesn't work: a 10 or 20 percent markup on unit cost makes no sense when unit cost is zero. You must price your information goods according to consumer value, not according to your production cost."

"Production of an information good involves high fixed costs but low marginal costs.", Shapiro & Varian, Information Rules (1999)

So: read your own cost structure as a risk profile before you read it as a spreadsheet. If you're fixed-cost-heavy, your job is volume and utilisation, every empty seat, idle server or unsold ticket is pure loss, but every extra sale above break-even is almost pure profit. If you're variable-cost-heavy, you're more resilient when sales dip but you'll never get the same explosive margin from scale. Neither is "better"; they demand different leadership. One rewards filling capacity, the other rewards protecting margin per unit.

Where this model breaks down

The neat fixed/variable split is a short-run convenience, and treating it as a law is the classic trap. Three honest limitations. First, costs are only fixed within a relevant range: cross a threshold and they jump, a fourth shift, a bigger warehouse, another team lead. These "step costs" mean unit cost doesn't glide down forever. Second, almost nothing is purely one or the other; most real costs are mixed (a phone plan with a base fee plus per-minute charges), and forcing them into a clean bucket smuggles in error. Third, the dangerous one, marginal-cost pricing is a floor for the occasional spare-capacity deal, not a strategy: price every sale at marginal cost and you never recover the fixed costs, so a business that looks profitable on each transaction quietly drowns. Shapiro and Varian's warning cuts both ways, when marginal cost is near zero, competition can drag price toward zero too, so you need value-based pricing or genuine lock-in to make any profit at all.

Two habits guard against this. Label step costs explicitly on your model and mark the volumes where they trigger, so a growth plan doesn't sail past a cost cliff unnoticed. And ring-fence marginal-cost pricing as an exception you grant deliberately, for one-off capacity fill, never as the default that quietly sets fire to your fixed-cost recovery.

A worked example

Illustrative figures, a stand-in business, not a benchmark. Picture a small print studio, "Marlow Press," making custom posters. Rent, the leased printer and one salary come to £8,000 a month in fixed costs. Each poster eats £3 of paper, ink and packaging, its variable, and so its marginal, cost. They sell at £11.

Every poster contributes £11 − £3 = £8 toward the fixed costs (its contribution margin). To break even, Marlow must cover £8,000 ÷ £8 = 1,000 posters a month. Below that, the fixed costs aren't yet paid and the studio loses money no matter how busy it looks; above it, each extra poster drops £8 straight to profit, because the £8,000 is already covered.

Now watch unit cost move. At 1,000 posters, total cost is £8,000 fixed + £3,000 variable = £11,000, so unit cost is £11, exactly the price, which is why that's break-even. Sell 2,000 and total cost is £14,000: unit cost falls to £7, because the same £8,000 now spreads over twice the posters. Marginal cost never budged from £3, but average cost halved its fixed-cost burden. That gap is the whole story of scale.

Here's where the two numbers earn their keep. A customer offers £6 each for 200 posters to fill an idle Friday, below unit cost, so reject it? Not necessarily. Each still clears the £3 marginal cost and contributes £3 toward fixed costs you owe anyway, so the deal adds £600 you wouldn't otherwise have. Accept it as an exception. But price all output at £6 and you'd contribute £3 a poster against the £8,000 monthly nut, needing 2,667 posters just to break even, a much harder business. Same costs, same product; the verdict flips on whether £6 is the exception or the rule.

flowchart TD
  S(["Marlow Press: £8,000 fixed/mo · £3 variable · £11 price"]) --> CM(["Contribution = £11 − £3 = £8 per poster"])
  CM --> BE(["Break-even = £8,000 ÷ £8 = 1,000 posters"])
  S --> U1(["At 1,000 units → unit cost £11"])
  S --> U2(["At 2,000 units → unit cost £7"])
  BE --> D{"£6 spare-capacity offer?"}
  D --> Y(["As an exception: clears £3 marginal, adds £600, accept"])
  D --> N(["As the standard price: break-even jumps to 2,667, refuse"])
					

Frequently asked questions

What's the difference between marginal cost and unit cost?

Marginal cost is what one more unit costs, typically just its variable cost, since the fixed costs are already committed. Unit cost (average cost) is total cost, fixed and variable, divided by all units produced, so it carries a share of the fixed overhead and falls as volume rises. Marginal cost is roughly flat per unit; unit cost shrinks with scale. You price the floor off marginal cost and judge viability off unit cost.

Is a high-fixed-cost business bad?

No, it's a different bet. High fixed costs (high operating leverage) mean more risk when sales fall, because you owe those costs regardless, but more reward when sales rise, because revenue above break-even is nearly all profit. The question isn't whether fixed costs are high, it's whether you can reliably fill the capacity they buy. Airlines, software and gyms live on this trade.

How does the cost mix change my break-even?

Break-even is fixed costs divided by the contribution margin per unit (price minus variable cost). More fixed cost raises the break-even volume; a fatter contribution margin lowers it. That's why a price cut is dangerous in disguise, shaving the price shrinks the contribution margin, which can push break-even up by far more units than the discount seems to imply.

Should I ever sell below my unit cost?

Sometimes, deliberately, as an exception. If a one-off sale clears your marginal cost it still contributes something to fixed costs you already owe, so it can beat leaving capacity idle. But this is a spare-capacity tactic, not a pricing strategy: if your standard price doesn't clear unit cost, you never recover your fixed costs and lose money at every volume.

Where do step costs trip people up?

People assume unit cost keeps falling smoothly forever, so growth looks ever more profitable. In reality fixed costs jump in steps, a new shift, a bigger lease, another manager, each time you outgrow the current setup. Map those thresholds onto your volume forecast so a plan doesn't quietly march past a cost cliff and turn a "profitable" scale-up into a loss.

Related in the Toolkit

• Financial statements (P&L, balance sheet, cash flow), your cost structure is what the P&L's cost lines actually represent; this is how to read them.
• Reading annual reports, fixed-vs-variable mix and operating leverage are exactly what you reverse-engineer from a company's cost disclosures.
• Management vs financial accounting, marginal and contribution costing live in management accounting, the internal view built for decisions like these.
• Accounting standards & revenue recognition (IFRS 15 / GAAP, subscription revenue), how costs are matched to revenue shapes the unit-cost picture, especially for subscriptions.
• Budgeting (OPEX, CAPEX, annual planning vs actuals), the OPEX/CAPEX split is the fixed/variable distinction by another name, on the planning side.
• Forecasting, FP&A & variance analysis, knowing which costs are fixed lets you forecast profit at different volumes and explain the variances.
• Sales & operations planning (S&OP) & demand planning, matching demand to capacity is how a fixed-cost business avoids paying for idle capacity.
• Engineering productivity & delivery metrics (DORA), in software, engineering is the big fixed cost, so its productivity is the lever on your whole cost structure.

Where to go next

• Operating leverage, AccountingTools, a clear, practitioner-grade definition of how the fixed/variable mix turns into profit swings and risk.
• Information Rules, Carl Shapiro & Hal Varian (1999), the book that explains why high-fixed/low-marginal cost structures force value-based pricing; the classic for digital businesses.
• Cost-Volume-Profit Analysis, Managerial Accounting (open textbook), a free, worked chapter on contribution margin and break-even, if you want the full method behind the example here.
• Cost Volume Profit Analysis (CVP): calculating the Break-Even Point (YouTube), a short, numbers-on-screen walkthrough of break-even from fixed costs and contribution margin.