Case study: make or buy
Who asked: the CFO of Harrow Bicycles (made up for practice). Frames come from an outside supplier. Operations wants to build its own frame plant in part of the warehouse that's now leased out. A feasibility study, already paid for, says the plant would cut the cost per frame.
The data: volumes, the supplier's price, the plant's build cost, variable and fixed costs, the warehouse rent that would be given up, the study's cost, the plant's life, the discount rate and the CFO's payback limit, on the Plant sheet.
What done looks like: the annual saving with the right costs in and out, NPV, payback and IRR, NPV across volumes, the break-even volume, and a recommendation that says where NPV and payback disagree.
A make-or-buy decision compares only the costs that change with the choice: the relevant costs.
- Sunk costs stay out. Money already spent, like the feasibility study, is gone whichever way Harrow decides.
- Opportunity costs go in. If the plant uses space that now earns rent, that lost rent is a cost of making, even though no one writes a check for it.
With the annual saving right, the project is a capital budgeting question: NPV, payback and IRR, then how sensitive the answer is to the volume forecast. A plant has high fixed costs, so its value swings with volume: the break-even volume is where the saving per frame just covers the plant's costs and repays the build over its life.
NPV and payback can disagree. When they do, NPV is the better guide to value, but payback is a rough flag for risk. Say which one you're following and why.
Buy 40,000 x 92 = 3,680,000
Make 40,000 x 61 + 420,000 + 180,000 rent = 3,040,000
Saving 640,000 a year
The lost rent is in; the 150,000 study is not.