Relevant costs: special orders, make or buy and scarce resources
Most business decisions turn on what changes. A cost that will be incurred whatever is decided cannot help to choose, and a cost that looks small on paper can be the one that matters. This lesson sorts costs into relevant and irrelevant, then applies the test to three decisions with a fictional bicycle maker: whether to accept a special order, whether to make a part or buy it, and what to produce when something is scarce. Apple's and Microsoft's capital spending then show how much capacity each has already committed.
- Level
- Beginner
- Reading time
- About 24 minutes
- Know this first
- Product costing methods
By the end you will be able to:
- say which costs and revenues are relevant to a decision and which are not, including sunk and opportunity costs
- decide whether to accept a special order, with and without spare capacity
- compare making and buying a component using avoidable costs and the value of released capacity
- rank products by contribution per unit of a scarce resource
Deciding with the numbers that change
The earlier lessons in this module sorted costs by behaviour, traced them to products and measured how profit responds to volume. This one uses them to decide. A manager rarely chooses between a good business and a bad one. The choice is usually between two actions, such as taking an order or turning it down, and the question is how profit differs between them. The answer depends on the costs and revenues that differ, and on nothing else.
The running example is the fictional bicycle maker once more. Its plant can make 10,000 bicycles a year, it regularly sells 8,000 at $200, and the full cost of a regular bicycle at that volume is $155, which is the $30 of fixed cost per bicycle plus $125 of variable cost from the first lesson of the module. Apple and Microsoft appear near the end.
What makes a cost relevant
A cost or revenue is relevant if it will influence the decision. In practice that means it is a future amount that differs between the alternatives. This is differential analysis: look only at the differences in costs and revenues between the choices. Three kinds of cost are left out or handled with care.
A sunk cost has already been incurred and cannot be avoided, so it is the same whichever choice is made and is irrelevant. The price paid last year for a machine is sunk, whatever the machine is now worth. A future cost that does not differ between the alternatives is also irrelevant, such as the rent on a factory that will be paid anyway. An opportunity cost is the benefit lost by not choosing the next-best alternative. It is not a payment, so it never appears in the books, but it is relevant, and ignoring it is a common way to get a decision wrong.
Numbers are not the whole decision. Qualitative factors, those that cannot be measured in numbers, can outweigh a small financial gain: how customers will react, employee morale, the dependability of a supplier, brand image. The lesson works through the quantitative side, and each decision below ends with the qualitative factors that deserve a hearing.
A special order
A distributor offers to buy bicycles directly, at a price below the usual one. Should the maker accept? It depends on capacity. The plant has room for 2,000 more bicycles a year than it now sells. Within that room, the order uses idle capacity, and the fixed costs, which are paid anyway, are irrelevant. Only the extra costs matter: $90 of materials and $30 of labour, so $120 a bicycle. The $5 sales commission is not paid on a direct order, so it is not a cost of this decision. Beyond the spare 2,000, every bicycle made for the order displaces a regular sale that would have contributed $75, and that lost contribution is an opportunity cost. If the order needs anything extra, such as special packaging or tooling, that cost is relevant too.
Price a special order
Set the offered price, the number of bicycles and any extra costs the order brings. The left page lists the order's revenue and the costs that exist only because of it. The right page puts the answer per bicycle, next to the lowest price that would break even and the full cost of a regular bicycle.
Accept on the numbers. The order brings in $225,000 and causes variable costs of $180,000, so it adds $45,000. The lowest price that breaks even is $120.00 a bicycle, against an offer of $150.00. A rule that compared the offer with the full cost of $155.00 would turn it down and lose that profit.
Four experiments. First, the starting offer of $150 for 1,500 bicycles fits within the spare capacity. It brings in $225,000, causes $180,000 of variable costs and adds $45,000 to profit, even though $150 is below the full cost of $155. A rule that compared price with full cost would show a loss of $7,500 on the order and turn it down, because the $30 of fixed cost in the full figure would be paid whether or not the order is taken. Second, raise the order to 3,000. The first 2,000 use idle capacity, but the other 1,000 displace regular sales and give up $75,000 of contribution. Doubling the order does not double the gain: it is $15,000, and the lowest break-even price rises to $145.00. Third, cut the price to $115, below the $120 variable cost, and the order loses $7,500; no extra volume can fix a price below the variable cost. Fourth, go back to the starting order and add $40,000 of extra costs. The gain drops to $5,000 and the break-even price rises to $146.67.
Even when the numbers favour an order, the qualitative factors count. Existing customers may hear of the lower price and ask for it. A cheaper version may take sales from the regular product. A large order may crowd out regular customers. In some countries, selling the same goods at different prices to different buyers is regulated. Against these, an order may also open the door to a relationship with a new customer. The rule is to accept an order when its price covers every cost that the order causes, and then weigh these factors before saying yes.
Make or buy
The same logic applies to a part. The maker produces wheel sets in house, 10,000 a year, at a variable cost of $30 each ($18 of materials, $9 of labour and $3 of variable overhead). The factory charges each set $12 of fixed overhead. A supplier offers to deliver them at $38. The question is how much of the $12 would disappear if the sets were bought. Suppose $4 of it is the lease on the wheel-building equipment, which would end, and the other $8 is a share of the building and supervision, which would stay. Only the $4 is avoidable and relevant.
There is one more consideration. If buying frees up factory capacity that could earn money on other work, that profit is the opportunity cost of making. Suppose the freed capacity could earn $70,000 a year, or $7 for each set.
| Making 10,000 wheel sets a year | Full-cost view | Relevant costs only | Relevant costs and capacity |
|---|---|---|---|
| Variable cost to make (materials, labour, variable overhead) | $30.00 | $30.00 | $30.00 |
| Fixed overhead charged to each set | $12.00 | $4.00 (avoidable only) | $4.00 (avoidable only) |
| Profit forgone on the capacity released | not considered | not considered | $7.00 |
| Cost of making | $42.00 | $34.00 | $41.00 |
| Price to buy from a supplier | $38.00 | $38.00 | $38.00 |
| Decision | Buy, saving $40,000 | Make, saving $40,000 | Buy, saving $30,000 |
The full-cost view spreads the whole $12 of overhead over each set and recommends buying. Using only avoidable costs reverses it, because making then costs $34 against a price of $38, and saves $40,000. If the released capacity could earn $70,000 elsewhere, making costs $41 per set and buying is better, by $30,000. The same facts give three answers, and only the last two ask the right question: what does each choice change? Before outsourcing, qualitative factors need checking as well: whether the supplier's quality and delivery can be relied on, whether the quoted price will last, whether it can cope with changes in volume, and what the effect would be on staff.
When something is scarce
Sometimes the limit is not demand but a scarce resource, such as machine time, skilled labour or space. A constraint is a scarce resource that limits output, and it is called a bottleneck when it slows production. With a constraint, ranking products by contribution per unit is misleading. The right measure is the contribution per unit of the constrained resource.
Suppose the maker also builds a road model, and both models pass through the same assembly stage, which has 10,000 hours a year. The road model sells for $260 with $140 of variable cost and takes 4 assembly hours; the mountain model, the $200 bicycle of earlier lessons, takes 2 hours and has a variable cost of $125 (commission included). Customers would buy up to 2,000 road and 4,000 mountain bicycles, which together would need 16,000 hours.
| Two models, 10,000 assembly hours | Road model | Mountain model |
|---|---|---|
| Selling price | $260 | $200 |
| Variable cost | $140 | $125 |
| Contribution per bicycle | $120 | $75 |
| Assembly hours per bicycle | 4 | 2 |
| Contribution per assembly hour | $30.00 | $37.50 |
| Most that can be sold | 2,000 | 4,000 |
The road model contributes more per bicycle, $120 against $75, but it uses twice the hours, so each hour of assembly earns $30 on the road model and $37.50 on the mountain model. Hours are the limit, so the mountain model comes first.
| Plan for 10,000 assembly hours | Road model made | Mountain model made | Total contribution |
|---|---|---|---|
| Mountain first (highest contribution per hour) | 500 | 4,000 | $360,000 |
| Road first (highest contribution per bicycle) | 2,000 | 1,000 | $315,000 |
Making the mountain model first uses 8,000 hours for 4,000 bicycles and leaves 2,000 hours, enough for 500 road bicycles, for a contribution of $360,000. Ranking by contribution per bicycle puts the road model first, uses 8,000 hours for 2,000 road bicycles and leaves 2,000 hours for 1,000 mountain bicycles, for $315,000. The correct ranking earns $45,000 more from the same hours. It also tells the maker what an extra assembly hour is worth: the road model is the use that is not fully met, so one more hour is worth $30. If more hours can be bought for less than that, they are worth buying.
Committed capacity at Apple and Microsoft
The published statements cannot show a company's special orders or its make-or-buy decisions, which are made inside the business. What they do show is how much capacity each company has built, and capacity is where fixed, sunk and relevant costs meet. Once a plant or a data centre is built, the money spent is sunk and, for short decisions, only the extra costs of using it matter, as with the special order above. Whether to build more is a different question, in which the whole cost is relevant.
| Fiscal 2025, $ million | Apple | Microsoft |
|---|---|---|
| Revenue | 416,161 | 281,724 |
| Capital spending | 12,715 | 64,551 |
| Capital spending as a share of revenue | 3.1% | 22.9% |
| Depreciation and amortisation | 11,698 | 34,153 |
| Depreciation and amortisation as a share of revenue | 2.8% | 12.1% |
In fiscal 2025 Apple spent $12.7 billion on property, plant and equipment, 3.1% of its revenue, and Microsoft spent $64.6 billion, 22.9%. Depreciation and amortisation, which spreads past spending over the years, was 2.8% of revenue at Apple and 12.1% at Microsoft. Microsoft is therefore carrying a much larger committed cost base per dollar of sales, and is adding to it faster. Such costs do not change when sales do, which is the cost structure that the lesson on operating leverage described. These ratios say nothing about whether either company's choices are wise. They only show how much of each has already been committed to capacity.
What these rules cannot tell you
- They assume the estimates are right. A relevant cost is only as good as the forecast of what will change, and the avoidable part of a fixed cost is often the hardest to judge.
- They look at one decision at a time. A policy of accepting every order that covers variable cost would leave the business with no way of covering its fixed costs. Over time, the average price must cover everything.
- The short run is not the long run. Fixed costs are irrelevant while capacity sits idle. When capacity has to be added or replaced, they are relevant again.
- Constraints move. Fixing the bottleneck often shifts it to another stage, and the ranking can change.
- Qualitative factors can decide. A small gain may not justify annoying a loyal customer or depending on one supplier.
Common mistakes
Comparing the price with the full cost
The full cost includes fixed costs that may be paid whether or not the order is taken. Compare the price with the costs the order causes.
Treating a sunk cost as a reason
What was paid for a machine last year cannot be recovered by using it. Only its future costs and resale value count.
Forgetting the opportunity cost
If an order or a make-or-buy choice uses capacity that could earn something elsewhere, that lost profit belongs in the comparison.
Treating all fixed costs alike
A fixed cost that would end with the decision, such as an equipment lease, is relevant. One that would continue is not.
Ranking by contribution per unit when a resource is scarce
The product with the largest contribution per unit may use the most of the scarce resource. Rank by contribution per unit of the resource.
Try it yourself
1. A plant has idle capacity. A buyer offers $160 a unit for 500 units. The variable cost is $120, and normal full cost is $170. Should the order be accepted, and what does it add to profit?
Show the working
The order uses idle capacity, so fixed costs are irrelevant. Gain per unit = 160 − 120 = $40, so the order adds 40 × 500 = $20,000. A full-cost rule would reject it because $160 is below $170, but the full cost includes fixed costs that are paid anyway.
2. The same order would need $15,000 of extra costs, such as a temporary supervisor. What is the lowest acceptable price, and what is the gain at $160?
Show the working
Lowest price = 120 + 15,000 ÷ 500 = $150. At $160 the gain is (160 − 150) × 500 = $5,000, which is 20,000 less the 15,000 of extra costs.
3. A part costs $12 in variable costs to make, plus $8 of allocated fixed overhead, of which $3 would be avoided if the part were bought. A supplier quotes $16. Which is cheaper per unit, and by how much?
Show the working
The relevant cost of making is 12 + 3 = $15. That is $1 below the supplier's $16, so making is cheaper by $1 per unit, before opportunity cost and qualitative factors. The full cost of $20 would wrongly suggest buying.
4. Product A contributes $40 a unit and uses 2 hours of a scarce machine. Product B contributes $45 a unit and uses 3 hours. The machine has 600 hours and demand for each is ample. Which should be made, and what is the contribution?
Show the working
Contribution per hour is 40 ÷ 2 = $20 for A and 45 ÷ 3 = $15 for B. A is better. 600 hours make 300 units of A for 300 × 40 = $12,000, against 200 units of B for 200 × 45 = $9,000.
5. A machine bought last year for $80,000 can be sold for $20,000 today. Is the $80,000 relevant to the decision to keep it or sell it?
Show the working
No, it is a sunk cost and the same either way. The relevant figure is the $20,000 that selling would bring in, which is the opportunity cost of keeping the machine, together with what the machine will earn if kept.
Where this is used
- A manufacturer with spare capacity decides on the lowest price at which it will accept a bulk order from a new customer.
- A business reviewing its supply chain compares the avoidable cost of making a part with a supplier's quote, and the value of the space and people released.
- A factory with a bottleneck machine decides which products get the machine time, and what an extra hour of it is worth.
- A manager deciding whether to keep or close a product line looks at the costs that would actually go away.
Key terms
- Relevant cost
- A future cost that differs between the alternatives and so affects the decision.
- Differential analysis
- Comparing alternatives by looking only at the differences in their costs and revenues.
- Sunk cost
- A cost already incurred that cannot be avoided, and so is irrelevant.
- Opportunity cost
- The benefit given up by not choosing the next-best alternative.
- Avoidable cost
- A cost that would stop if a particular decision were taken.
- Special order
- A one-off order, usually at a price below the normal one.
- Constraint
- A scarce resource that limits output. A bottleneck is a constraint that slows production.
- Qualitative factor
- A consideration that cannot be measured in numbers, such as customer reaction or supplier reliability.
Questions people ask
Does this mean a business should price everything at variable cost?
No. The rule is about one-off decisions with idle capacity. Regular prices must cover all costs and leave a profit, otherwise the business loses money over time.
Why do accountants allocate fixed costs if they are not relevant?
For external reporting and for long-run pricing, every cost must be recovered somewhere. Allocation is for that purpose, and not for choosing between alternatives.
What if the offer is below variable cost but would win a major new customer?
Then the loss is an investment in a relationship, and management should treat it as one: estimate what the customer is likely to be worth over time and compare. The numbers on the single order do not show that.
Is opportunity cost an actual cost?
It is not a payment and does not appear in the accounts, but it is a real cost of the decision, since the benefit it measures is lost.
Sources and review
- Apple Inc., Form 8-K, Exhibit 99.1: fourth quarter and full-year fiscal 2025 results, filed with the US Securities and Exchange Commission. Source of Apple's statements.
- Microsoft Corporation, Form 8-K, Exhibit 99.1: fourth quarter and fiscal year 2025 results, filed with the US Securities and Exchange Commission. Source of Microsoft's statements.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 10.1: Identify relevant information for decision-making, for relevant, sunk and opportunity costs and qualitative factors.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 10.2: Evaluate and determine whether to accept or reject a special order, for the special order rule with idle and full capacity.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 10.3: Evaluate and determine whether to make or buy a component, for avoidable fixed costs in make-or-buy decisions.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 10.6: Evaluate and determine how to make decisions when resources are constrained, for constraints and contribution per unit of the scarce resource.
How the numbers were checked: every special-order figure, the make-or-buy comparison and both production plans were recomputed in plain code, the contribution-per-hour ranking was checked against the alternative plan, and the calculator, which runs separately, was compared with that code across a grid of inputs. The capital spending, depreciation and revenue for Apple and Microsoft come from the filed results.