“We’ve already put too much into this to stop.”
That sentence sounds like commitment. It can also be a request to spend more money protecting money you can no longer recover. The past gets a vote in the future even when it has no useful information to contribute.
Imagine a business owner who has spent months building a course. The topic interested people, but nobody has purchased the specific offer. Finishing will require another substantial investment in editing, design, and promotion. The owner’s strongest argument for continuing is the amount of work already completed.
That work matters as history. It may have produced useful assets and knowledge. But it doesn’t automatically make the next investment worthwhile. The question is what the owner can reasonably expect from the resources still available.
A sunk cost is a past cost you can’t recover by choosing differently now. Money spent on an unusable advertisement is gone whether you run another advertisement or stop. Hours spent drafting an abandoned manuscript won’t return because you finish it.
In their 1985 paper on the psychology of sunk cost, Hal Arkes and Catherine Blumer examined how prior investments can increase willingness to continue an endeavor. Their research included a theater subscription field study and questionnaire studies. The paper offers evidence of the effect; it doesn’t establish that persistence is always irrational or that every continuing project is a mistake.
The distinction is crucial. A project may be worth finishing because the remaining cost is low and the likely future benefit is high. The fact that you’ve spent a lot doesn’t prove you should continue. It also doesn’t prove you should quit.
“Ignore sunk costs” is a decision principle, not an instruction to forget everything you learned while incurring them.
Draw a line across a page. Above it, list the history: money spent, time invested, promises made, assets created. Below it, list the options currently available and what each would require from here.
For the hypothetical course creator, the options might include finishing the full program, testing a smaller paid workshop, licensing selected materials, or stopping development. Each has different future costs and possible benefits. “Finish” and “waste everything” are rarely the only choices.
Ask what you would do if you inherited the project today. You’d receive the existing materials and knowledge, but you wouldn’t have to defend the original decision. Would you invest the next month of work at the current price? Would you choose this opportunity over the alternatives you now have?
This thought experiment won’t remove your emotions. It can expose whether your argument concerns future value or past embarrassment. Both deserve attention, but they require different responses.
Leaving can create new costs. Those belong in the decision. If changing a software system requires migration, training, and temporary disruption, you can’t dismiss those expenses by announcing that sunk costs are irrelevant.
Similarly, a contract may create obligations that continue even if you stop using the service. A project may involve promises to customers or responsibilities to employees. The costs of fulfilling or responsibly ending those commitments are future consequences, not merely memories.
Suppose a company dislikes its current scheduling platform. The subscription fees paid last year are sunk. The cost of exporting records and training staff on a replacement is prospective. The time employees will keep losing if the system remains unsuitable is prospective, too. A useful comparison includes both sides.
List consequences according to whether today’s choice can change them. That question is more reliable than sorting expenses into emotional categories such as painful, annoying, or embarrassing. A past mistake can feel enormous while having no bearing on the best remaining option.
Continuing a project consumes more than its direct expenses. It occupies time and attention that could serve something else. That alternative doesn’t have to be a glamorous new venture. It might be delivering well for existing clients, improving a functioning product, or recovering enough capacity to make decent decisions.
Imagine a consultant maintaining an unprofitable offer because it took a year to develop. The offer requires regular customization and attracts buyers who need substantial support. Keeping it may prevent the consultant from serving a simpler, more valuable problem for a better-fitting audience.
The comparison should use realistic alternatives. Don’t weigh your frustrating real project against an imaginary business with effortless demand and no operational problems. Estimate the work, uncertainty, and delay involved in the alternative as honestly as you estimate the current project.
If there is no better immediate use of a resource, that matters. It still doesn’t make every continuation sensible. A project can consume attention, create liabilities, or delay necessary maintenance even when your next big idea hasn’t arrived.
Some worthwhile projects require persistence before the evidence becomes informative. You won’t learn much about a monthly service by abandoning it after one awkward conversation. You also won’t learn much by changing the offer after every refusal.
Before continuing, name the question the next investment will answer. Then specify the amount of time or money required to answer it and the results that would support each next step.
For example, a founder might decide to present a defined offer to a relevant set of budget owners over an appropriate buying period. The purpose is to learn whether the problem is urgent, who can authorize a purchase, and what blocks a decision. The test needs enough structure to produce interpretable evidence.
Compare that with “We’ll keep promoting until people understand it.” That sentence has no budget, endpoint, or result capable of challenging the original belief. It is a rescue mission without a return date.
This is where confirmation bias can reinforce the sunk cost trap. Every encouraging comment becomes permission to continue. Every refusal becomes proof that more explanation is needed. A fair test allows the idea to earn further resources or lose them.
Quitting badly can destroy value that could have been preserved. A decision to stop should include an exit plan rather than an emotional disappearance.
Identify what can be reused, sold, transferred, documented, or responsibly retired. Existing research may support a different offer. A customer list may require specific privacy protections. A collaborator may need notice and a clear account of outstanding responsibilities. Some materials may be worth keeping simply because they prevent you from repeating the same experiment later.
Don’t turn salvage into another sprawling project. Spending six weeks organizing something you’ve decided to stop can become a continuation under a different name. Choose a proportionate closeout process with a defined finish.
Communicate plainly with affected people. Explain what is ending, why, what commitments remain, and what happens next. You don’t need a heroic narrative about failure being success in disguise. You need an accurate account and dependable follow-through.
When trust is involved, the manner of stopping can matter as much as the decision itself. Clear expectations are part of repairing trust if earlier commitments can no longer be met.
One reason people keep going is that stopping seems to turn their past work into a verdict on their intelligence. If the project ends, they imagine the investment becomes proof they were foolish.
Separate the original decision from the current one. You may have made a reasonable bet with limited information. You may also have ignored warnings. Either way, the question now is whether the next investment is justified by what you know today.
Write a short account of what changed. Which assumptions were supported? Which failed? Which signals did you miss? What would you test earlier next time? That record gives the experience somewhere useful to go.
Be specific. “Never trust anyone” is an expensive lesson because it damages future opportunities without explaining the error. “Verify that the buyer controls a budget before designing a custom proposal” is a practical improvement. The quality of the lesson matters more than the drama of the disappointment.
If several people own the decision, discuss the stopping conditions while everyone is still willing to consider them. A hypothetical partnership might agree that another development phase requires a paying pilot customer, a working prototype, and a delivery budget within an agreed range. Those conditions should reflect the project’s actual economics, rather than arbitrary milestones chosen because they’re easy to celebrate.
Also identify who can authorize an extension and what evidence they must provide. Otherwise, each missed milestone can quietly become another exception. An exception may be justified, but it deserves a new decision with an explicit cost. Calling it a temporary delay doesn’t make the resources free.
Leave room for genuinely new information. A valuable unexpected result may support a different path. Document that path as a fresh proposal rather than treating it as proof that the original plan was correct all along. You can preserve curiosity and flexibility while still requiring each additional commitment to make sense.
Before your next continuation decision, compare three options: proceed as planned, change the commitment, and stop responsibly. For each, write the remaining cost, the plausible benefit, the main uncertainty, and the condition that would trigger another review.
Don’t pretend those estimates are precise when they aren’t. Ranges and explicit assumptions are more useful than confident numbers built to justify a preferred conclusion. If a small test can resolve the main uncertainty, consider whether it deserves a limited budget.
You may decide to keep going. If the future case is strong, persistence can be entirely sensible. You may decide to stop. If the future case has collapsed, stopping can preserve resources for something that can still work.
The important change is where your justification begins. It begins with the situation you have, the choices you can make, and the consequences still open to influence.
Your past investment bought you today’s position, including whatever you learned. It doesn’t have to purchase the rest of your life. Give the next hour, the next dollar, and the next promise a reason that belongs to the future.