What it means
Every business is really a bundle of activities it has chosen to perform itself. A make-or-buy decision asks, for one of those activities, whether the company is better off doing it or paying someone else to.
The same framework covers a machined part, a call centre and a finance department. The financial test rests on relevant costs, meaning costs that differ between the two options.
Variable production costs, avoidable supervisory salaries and equipment leases that could be ended all count, while general overheads such as head office rent usually do not, because they continue regardless. Loading unavoidable overhead into the make cost is the classic way of talking yourself into outsourcing something that was actually cheaper in-house.
Opportunity cost belongs in the calculation too. If buying the part frees a production line that can then make something else profitable, that extra contribution is a genuine benefit of buying.
If the freed capacity would simply sit idle, the saving is smaller than it first appears. The non-financial factors often decide the case.
Control over quality, protection of intellectual property, supplier reliability, the flexibility to change volumes quickly and the effect on staff morale all matter, and none of them appear neatly in a spreadsheet. A decision that saves a modest amount per unit but gives a competitor sight of your design is rarely a good trade.
Make-or-buy is not a permanent answer. Supplier pricing, exchange rates, wage costs and your own volumes all move, so a sensible business revisits the comparison every year or two.
Long outsourcing contracts with heavy exit penalties are risky precisely because they freeze a decision that should stay open.
In practice
Real-world examples
Example 1
A food producer compares running its own delivery fleet against a third-party logistics contract. The fleet costs more per drop but guarantees temperature control and next-morning slots, so the company keeps chilled routes in-house and outsources ambient pallet deliveries.
Example 2
A software company decides whether to build its own billing engine or license one. Building would cost roughly $340,000 of engineering time in year one, so it buys, then revisits the decision three years later when subscription fees have risen and its needs have become unusual.
Example 3
A furniture maker outsources powder coating after calculating that its own line runs at 30% capacity. It keeps a small booth for urgent jobs, accepting a slightly higher blended cost in return for the ability to turn samples around in a day.
Think of it
Deciding whether to cook dinner at home or order takeaway. Cooking yourself uses your own groceries and time, while ordering out costs more money but saves you effort.
Formula
Calculation
Cost to make = (variable cost per unit x volume) + avoidable fixed costs
Cost to buy = (purchase price per unit + acquisition costs per unit) x volume
A pump manufacturer needs 60,000 valve assemblies a year.
Making them costs $9.00 in materials, $4.50 in direct labour and $2.00 in variable overhead, giving a variable cost of $9.00 + $4.50 + $2.00 = $15.50 a unit. The company also carries $180,000 of avoidable fixed costs for a dedicated supervisor and a machine lease that would end if production stopped.
Cost to make = (60,000 x $15.50) + $180,000 = $930,000 + $180,000 = $1,110,000, which is $1,110,000 / 60,000 = $18.50 a unit.
A supplier quotes $17.20 a unit, and inbound freight and goods-in inspection add $0.40 a unit, giving a delivered cost of $17.20 + $0.40 = $17.60.
Cost to buy = 60,000 x $17.60 = $1,056,000.
Buying is cheaper by $1,110,000 - $1,056,000 = $54,000 a year, or $0.90 a unit. That $54,000 is the annual price of keeping the work in-house, and management has to decide whether direct control of quality and supply is worth it. If the freed floor space could instead produce another product generating $70,000 of contribution, the case for buying strengthens to $54,000 + $70,000 = $124,000 a year.Case study
Seen in the real world
This is a fictional, illustrative example. Vantage Cycle Works made its own aluminium handlebars and had done so since the founder started in a shed. A new operations director produced an analysis showing an imported equivalent would cost $6.80 delivered against an in-house cost of $9.40, saving roughly $130,000 a year across 50,000 units.
Closer inspection showed that $2.10 of the $9.40 was an allocation of factory rates that would not disappear if the line closed. On a relevant-cost basis the real in-house figure was $7.30, cutting the saving to $25,000 a year before any redundancy costs.
In this illustrative case Vantage kept production in-house, and the operations director redirected the effort towards a genuine make-or-buy candidate: the anodising process, which was running two shifts short of viable volume.
Watch out
Common mistakes
- Including allocated head office or factory overhead in the make cost when those costs would continue whether the item is made or bought.
- Comparing a supplier's headline price with a fully loaded internal cost, and forgetting freight, inspection, tooling and supplier management time.
- Ignoring what happens to the freed capacity, which can either add contribution from other work or simply become an idle cost.
Questions
People also ask
Which costs are relevant to a make-or-buy decision?
Only those that change between the options, meaning variable costs, avoidable fixed costs and any opportunity cost of capacity, never sunk costs already spent.
Should the cheapest option always win?
No, because supply risk, quality control, intellectual property and the ability to flex volumes can easily outweigh a small per-unit saving.
Does make-or-buy apply outside manufacturing?
Yes, the same analysis is used for payroll, IT support, cleaning, marketing and almost any internal service that a supplier also offers.