A buyer offers you $215 per unit for 1,000 extra units. Your books say each unit costs $200 to make. Easy $15,000 profit — so you accept, run the extra batch, and watch the month's profit come in lower than if you had said no. What happened? The $200 was your average cost across everything you already produce. The extra 1,000 units — made on overtime, with rush-ordered materials and a rented machine — cost $238 apiece. You lost $23 on every unit of a deal that looked profitable.
That gap between what one more unit costs and what your average says is marginal cost, and pricing or production decisions made without it are guesses dressed up as math. This guide shows you how to compute it, why it falls and then rises as you grow, and the exact point where it tells you to stop producing.
The One-Question Definition
Marginal cost answers a single question: what will it cost to make just one more?
Not the average across your whole run. Not your total spend divided by total output. The incremental cost of the next unit — or the next batch — given everything you are already producing. If making 100 widgets costs $1,000 in total and making 101 costs $1,009, the marginal cost of that 101st widget is $9.
Formally:
Marginal Cost = (Change in Total Cost) / (Change in Quantity Produced)
Or, in the shorthand you will see in every economics text: MC = change in TC / change in Q.
One simplification makes this practical: fixed costs — rent, insurance, salaried overhead — do not change when you produce one more unit, at least in the short run. So the change in total cost is really the change in variable cost: direct materials, direct labor, production supplies, energy, extra maintenance, and anything else that moves with volume. You can write the formula that way instead:
MC = (Change in Variable Cost) / (Change in Quantity Produced)
That is the whole apparatus. Everything below is about using it without fooling yourself.
How to Calculate Marginal Cost in Four Steps
Take a concrete example: you run a small candle studio. Last month you poured 2,000 candles at a total production cost of $6,400. This month you poured 2,500 candles and total production cost came to $7,700. What did each of those extra 500 candles cost?
Step 1: Find the change in total cost. Subtract the old total from the new total: $7,700 − $6,400 = $1,300.
Step 2: Find the change in quantity. Subtract the old output from the new output: 2,500 − 2,000 = 500 candles.
Step 3: Divide the cost change by the quantity change. $1,300 / 500 = $2.60 per candle. That is your marginal cost over that range.
Step 4: Sanity-check it against your average. Last month's average cost was $6,400 / 2,000 = $3.20 per candle. The extra candles cost less than the average — $2.60 versus $3.20 — because the fixed costs (studio rent, insurance) were already covered. This is the normal, healthy pattern when you have spare capacity, and it is exactly why average cost misleads: it blends the cheap extra units with the expensive early ones.
Two practical notes before you trust any number you compute. First, calculate over small changes. The formula assumes costs behave smoothly across the range you measure, and a jump from 2,000 to 10,000 units will drag in step changes — new equipment, a new hire — that make one blended figure meaningless. Second, include all variable costs in the change: rush shipping on materials, overtime premiums, the extra kiln firing, card processing fees on the additional sales. The costs people forget are always variable ones.
Marginal Cost vs. Average Cost: The Distinction That Decides Real Money
These two numbers answer different questions, and using the wrong one is the most expensive mistake in this guide:
- Average cost = total cost / total quantity. It tells you whether the business as a whole is viable at a given price. It must be covered over time.
- Marginal cost = change in cost / change in quantity. It tells you whether the next units are worth producing. It must be covered on each decision.
Return to the opening example. Average cost $200, offer price $215, marginal cost $238. The average said yes; the margin said no; the margin was right. Any deal priced between your marginal cost and your average cost loses money on those units while looking profitable on a blended report. That is how businesses grow revenue and shrink profit in the same quarter.
The reverse trap exists too. Suppose your average cost is $200, your marginal cost on spare capacity is $140, and a buyer offers $160 for a one-off order you would otherwise not produce. Refusing because "$160 is below our $200 cost" leaves $20 per unit of contribution on the table — money that would have helped cover fixed costs that you pay whether or not you take the order. When capacity is idle, marginal cost is the floor, not average cost.
The rule of thumb: use average cost to judge the business, marginal cost to judge the next decision. Confusing the two costs you money in both directions.
Why Marginal Cost Falls, Then Rises: The U-Shaped Curve
Plot marginal cost against output and you typically get a U shape — falling at first, bottoming out, then climbing. Each phase has a physical cause you can feel in your own operation.
The falling phase: economies of scale
Early on, each additional unit gets cheaper to make. Workers develop routines and get faster. Machines run closer to their efficient operating point instead of idling. Fixed costs spread across more units. A bakery whose ovens are already hot and whose staff is already on shift can bake one more tray of loaves for little more than flour and yeast. During this phase, marginal cost sits below average cost and pulls the average down with it — the $2.60-versus-$3.20 pattern from the candle example.
The rising phase: diminishing returns and capacity walls
Nothing falls forever. As output approaches capacity, efficiency gains taper off and then reverse: machines need more maintenance, workers earn overtime, storage overflows, error rates climb, and rush orders carry premiums. The bakery that has to add a night shift or lease a second oven faces a much higher marginal cost per loaf than it did at midday with idle oven space.
Sometimes the rise is not gradual but a step: output is cheap right up until you need one more delivery van, one more employee, one more software tier — and then the marginal cost of the units that forced the step includes the whole step. Lumpy costs like these are why the smooth U-curve of textbooks kinks in real businesses, and why measuring over small quantity changes matters.
What the bottom of the U tells you
The inflection point — where marginal cost stops falling and starts rising — marks the limit of your current setup's efficiency. Everything past it costs more per unit than the units before it. You do not need to chart a curve to use this: when you notice overtime, expediting fees, and error rework creeping into a batch, you have crossed it. That is the moment to price the next order off the rising marginal cost, not last month's comfortable average.
The Produce-or-Stop Decision: Where Marginal Cost Meets Marginal Revenue
Marginal cost becomes a decision rule the moment you pair it with its twin: marginal revenue, the additional revenue from selling one more unit. The rule is the most useful sentence in production economics:
Keep producing while marginal revenue exceeds marginal cost. Stop where they meet.
If the next unit costs $9 to make and sells for $20, making it adds $11 to profit — produce it. If it costs $22 to make and sells for $20, making it subtracts $2 — stop. The profit-maximizing output is exactly where marginal cost equals marginal revenue: producing one unit less would leave money on the table, and one unit more would cost more than it earns.
Walk through a small table to feel the logic. A print shop sells posters at a flat $20 each (so marginal revenue is $20 all the way down) and tracks the marginal cost of each additional hundred posters:
| Batch (posters) | Marginal cost per poster | Marginal revenue | Add to profit? |
|---|---|---|---|
| 1–100 | $24.00 | $20.00 | No — still climbing toward efficiency |
| 101–200 | $18.00 | $20.00 | Yes — $2.00 each |
| 201–300 | $14.00 | $20.00 | Yes — $6.00 each |
| 301–400 | $16.00 | $20.00 | Yes — $4.00 each |
| 401–500 | $21.00 | $20.00 | No — overtime kicked in |
The shop should produce 400 posters and stop. The first hundred loses money per unit (setup dominates), the middle batches are the sweet spot, and the fifth batch costs more than it earns. Note the uncomfortable implication: the profitable decision is to turn down work past 400 posters at $20 — or reprice it above $21. Operating past the point where the curves cross loses money on every additional unit even while the overall operation stays profitable, which is why the loss hides so well in monthly totals.
Three caveats keep this rule honest. First, it assumes you can sell what you produce at the assumed price; flooding your own market can push marginal revenue down as volume rises. Second, it is a short-run rule — in the long run, prices must cover average cost including fixed investment, or the business liquidates no matter how healthy each margin looks. Third, real marginal cost moves with input prices and wages, so last quarter's crossing point is a starting estimate, not a permanent setting.
Tracking It in Your Books: Make Change-in-Cost Readable
You cannot compute change-in-cost if your books bury every batch in one undifferentiated expense pile. The fix is to record production costs so that each batch — or each month, shift, or run — stands alone. In a plain-text ledger, per-batch subaccounts do this with no extra tooling:
2026-10-01 * "October batch 1 — 500 candles, regular hours"
Expenses:Production:Batch-01:Materials 410.00 USD
Expenses:Production:Batch-01:Labor 320.00 USD
Expenses:Production:Batch-01:Energy 45.00 USD
Assets:Checking -775.00 USD
2026-10-15 * "October batch 2 — 500 candles, overtime + rush wax order"
Expenses:Production:Batch-02:Materials 468.00 USD
Expenses:Production:Batch-02:Labor 512.00 USD
Expenses:Production:Batch-02:Energy 58.00 USD
Assets:Checking -1038.00 USDNow the marginal cost of expanding output from 500 to 1,000 candles reads straight off the ledger: batch 2's own $1,038 for 500 candles, or $2.08 per candle. Batch 1 cost $1.55 per candle; batch 2 cost $2.08. The $0.53 jump is the rising arm of the U, itemized: $58 of rush materials, $192 of overtime labor, $13 of extra energy. When a buyer next asks for 500 more candles on a rush timeline, you quote off $2.08 — not the $1.81 average — and the overtime decision prices itself.
The general pattern: post variable production costs to a subaccount per batch, period, or job, keep fixed overhead in its own untouched accounts, and the change-in-cost for any marginal calculation is one balance query away. The Beancount documentation covers balance queries and account hierarchies if you are setting this structure up for the first time.
Five Mistakes That Corrupt Every Marginal Cost Calculation
1. Pricing off average cost. Covered above, worth repeating because it is the most common: averages blend cheap marginal units with expensive early ones. Every accept-or-reject decision on additional volume should clear the marginal cost of that volume.
2. Forgetting step costs. If the next 100 units require a new hire, a new machine, or a new software tier, that cost belongs in the marginal calculation for the units that trigger it. Spread the step over the units it unlocks — or better, treat the step as its own go/no-go decision with its own revenue case.
3. Counting sunk costs as marginal. Money already spent — the machine you bought last year, the batch already poured — does not change with the next unit and must not enter the formula. Only forward-looking, incremental costs count. This is the same discipline as ignoring sunk joint costs in any incremental decision.
4. Using a stale marginal cost. Material prices, wage rates, and energy tariffs move. A marginal cost computed in January can be fiction by October, especially for energy-intensive or imported inputs. Recompute from recent batches — which is free if your books carry the per-batch structure above — before any large quote.
5. Applying a short-run number to a long-run decision. Marginal cost properly excludes fixed costs only while capacity is fixed. Deciding whether to sign a five-year lease or launch a product line on marginal cost alone silently assumes the fixed investment is free. For capacity and entry decisions, the hurdle is average cost at the planned volume, with a margin above it.
Your Next Step: Compute One Real Number This Week
Pick your most recent production run, batch, or busy week. Pull its variable costs — materials, hourly labor, energy, fees — out of your books, preferably from their own subaccounts. Divide the cost change by the quantity change against the run before it. Compare the result to the price of your last accepted order and to your average cost. If the order cleared the margin, your pricing is sound; if it only cleared the average, you now know exactly how much the next quote needs to rise.
