Sunk cost
A sunk cost is money, time or effort already spent that cannot be recovered. In business, it should not determine what you do next: compare the future costs and benefits of your options instead.
Key takeaways
- A cost is sunk only to the extent that you cannot recover it through a refund, resale or another route.
- Spending $120,000 on a product does not justify spending another $40,000 if the next step will lose money.
- Sunk costs and fixed costs are different. A fixed subscription fee might be avoidable next month; last month's nonrefundable fee is sunk.
- Ignoring sunk costs does not mean ignoring lessons learned. Past results can change your forecast of future returns.
How sunk costs affect decisions
The test is simple: can your decision today change or recover this cost? If not, leave it out of the comparison between your options.
Examples include a nonrefundable advertising campaign that has already run, wages paid for completed work, and a research project with no resale value. A refundable deposit is not entirely sunk. Neither is equipment you can sell, although its resale value might be much lower than its purchase price.
The sunk cost fallacy happens when past spending becomes a reason to keep spending. For a founder, that often sounds like needing to finish a product because the team has already spent a year building it.
That year matters as evidence. It may reveal customer demand, technical limits or the team's ability to deliver. But the time itself cannot be recovered by launching.
Switching projects improves the forward result by $100,000
A startup has spent $120,000 developing a prototype. That money is nonrefundable, and the prototype has no resale value.
The founder is comparing three options over the next 12 months. The figures below are forecasts, not guaranteed results, and ignore differences in cash-flow timing within the year.
| Option | Future cash receipts | Future cash costs | Net future cash flow |
|---|---|---|---|
| Finish the current product | $180,000 | $220,000 | -$40,000 |
| Switch to a smaller product | $160,000 | $100,000 | $60,000 |
| Stop both projects | $0 | $0 | $0 |
The arithmetic is:
- Finish: $180,000 − $220,000 = −$40,000.
- Switch: $160,000 − $100,000 = $60,000.
- Difference: $60,000 − (−$40,000) = $100,000.
The $120,000 already spent is the same under every option. Including it would reduce every total by $120,000 without changing the ranking.
Switching has the highest expected cash return under these forecasts. It does not erase the earlier loss: the combined result would still be −$120,000 + $60,000 = −$60,000. The founder's job is to choose the best available future, not make the historical total look profitable.
Before committing, test the forecasts and whether the company has enough cash to carry out the plan. Calling an old cost sunk does not make the next investment sensible.
How an appraisal document treats sunk costs
The UK Treasury's Green Book, a public-investment appraisal guide, excludes sunk costs from the appraisal of options. It distinguishes them from opportunity costs, including the value of resources that could be used elsewhere.
For a startup, that means an old equipment purchase price should not drive the decision. But cash you could receive by selling that equipment belongs in the comparison: keeping it means giving up that alternative.
Sunk cost vs. fixed cost
These terms answer different questions.
| Term | What it describes | Example |
|---|---|---|
| Sunk cost | Whether a past cost can be recovered | Last month's nonrefundable software fee |
| Fixed cost | Whether a cost changes with output over a given period | A flat monthly software fee, regardless of usage |
A cost can be both fixed and sunk. But a future fixed cost might still be avoidable, such as next month's subscription if you cancel before renewal.
Questions people ask
Why is sunk cost a fallacy?
The cost is not a fallacy; the reasoning is. The mistake is treating an unrecoverable past investment as a benefit of continuing. Continue when the expected future benefits justify the future costs, not because stopping would make the earlier spending feel wasted.
What is an example of a sunk cost?
A business pays $5,000 for an advertisement that has already run, with no refund available. That $5,000 is sunk. Whether to buy another advertisement should depend on its expected return, informed by the first campaign's results.
What is sunk cost fallacy in relationships?
People use the phrase when someone stays mainly because of the years or effort already invested. That history cannot be recovered by staying; the relevant question is what the relationship is likely to offer from here.
