A sunk cost is a cost you've already paid and can't get back by any means. Because it's gone whichever way you decide, Corporate Finance Institute (CFI) says it shouldn't play any part in your next decision; only relevant costs, the ones that change depending on what you choose, should.
OpenStax, the free economics textbook, says the same thing: sunk costs happened in the past, can't be recovered and should be disregarded in current decisions. CFI lists a handful of other names you might come across, including past cost, embedded cost, prior year cost, stranded cost, sunk capital and retrospective cost.
Sunk costs vs fixed costs: the difference that trips people up
CFI makes a distinction most people don't expect. All sunk costs are fixed costs, but not all fixed costs are sunk costs. The test is whether you can get the cost back. Equipment is a fixed cost, but you can resell it or return it at a set price, so it isn't sunk.
| Cost | Can you get it back? | Sunk? |
|---|---|---|
| Equipment you can resell or return (CFI) | Yes, at a set price | No |
| Staff training that's already been delivered (CFI's ERP example) | No | Yes |
| A completed marketing study (CFI) | No | Yes |
That matters when you're deciding whether to stop. The resale value of your equipment is still on the table, so it belongs in the decision. The training is gone.
Three business examples from CFI
All three use CFI's US dollar figures, as of October 2026.
| Example (CFI) | Already spent (sunk) | What matters now |
|---|---|---|
| Airplane | $5 million | Finish it for $1 million more, or build the plane customers want for $4 million |
| ERP training | $10,000 | Whether the ERP system is worth keeping |
| Marketing study | $10 million | Whether the product will be profitable (the study says no) |
In the first, a company has spent $5 million building an airplane when its managers realise nobody wants it. It could finish it for another $1 million, or build the plane there's demand for at $4 million. The $5 million is sunk. The only relevant cost, in CFI's words, is the $4 million. The cheaper option looks tempting, but it buys a plane nobody will purchase.
In the second, a company spends $10,000 training staff on a new ERP system that turns out to be confusing and unreliable. Management wants to drop it, and CFI says the $10,000 shouldn't come into that decision.
In the third, a company spends $10 million on a marketing study that concludes its planned product would be heavily unprofitable. CFI says it shouldn't go ahead with the launch, and the cost of the study shouldn't sway it either way.
The sunk cost fallacy
CFI defines the fallacy as a mistake in reasoning: counting what you've already spent when you decide whether to carry on, because it feels like the resources would otherwise be wasted. It's the urge to keep pouring resources into something simply because so much has gone in already.
CFI's everyday examples make it easy to spot:
- Tom buys a $12.50 movie ticket online, finds the film boring, and sits through it anyway because he's paid.
- Jennifer pays a $100 entry fee for a tutoring club, finds it unhelpful after 4 of 7 sessions, and goes to the last 3 because of the fee.
- Someone books a $200 snowboarding trip, then a better $100 one, and the dates clash with no refunds. CFI says most people would pick the more expensive trip because losing it feels worse, even though it may be less fun.
Why a spreadsheet model helps
CFI says the fallacy is easier to avoid with a financial model, because a discounted cash flow (DCF) model only looks at future cash flows and gives no weight to the past. Building the comparison in Excel takes the emotion out and shows which option gives the best net present value from here. You don't need a full DCF to get that benefit. A simple two-column sheet of future costs and future returns for each option does most of the job.
Sunk costs are the mirror image of opportunity cost, which is the value of what you're about to give up. One is behind you; the other is in front of you, and only that one belongs in the decision.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
A five-step check before you keep going
- Write down everything you've already spent on the project.
- Mark which of those costs you could still recover through resale, refund or reuse.
- Cross out the ones you can't recover. They're sunk.
- Compare only the future costs and benefits of carrying on versus stopping.
- Pick the option with the better future outcome.
If you want a broader process for decisions like this, my guide to making better financial decisions in your business goes further, and these budgeting mistakes that hold back profit include a few that start with a sunk cost.
Where I come in
Marketing is full of sunk costs: the course you bought, the agency retainer you've paid, the channel you've spent a year building. That's why teams keep running channels that stopped working. I've spent 23 years in marketing, and I help small businesses and SaaS marketing teams look at each channel on what it'll bring in from here, then decide which to keep and which to stop. You can see how I work on my work with me page.
My verdict
What you've already spent is gone whichever way you go, so the only useful question is what the next pound or dollar will earn. If you'd start the project today knowing what you know now, carry on. If you wouldn't, stop. Stopping a bad project isn't a failure. It's a decision, and usually a good one.
The cost that still matters is what you give up by carrying on, which is the subject of opportunity cost.
Frequently asked questions
What is a sunk cost in simple terms?
It's a cost you've already paid and can't get back, so it shouldn't affect what you decide to do next.
Are sunk costs the same as fixed costs?
No. CFI says all sunk costs are fixed costs, but not all fixed costs are sunk. Equipment you can resell isn't sunk.
What is the sunk cost fallacy?
CFI defines it as counting what you've already spent when deciding whether to continue, because it feels like the resources would otherwise be wasted.
What's an example of a sunk cost in business?
In CFI's example, a company spends $10,000 training staff on an ERP system that doesn't work. That $10,000 shouldn't affect whether it drops the system.
How do you avoid the sunk cost fallacy?
CFI suggests a financial model that only looks at future cash flows, so the past gets no weight in the decision.
