Named as a behavioral pattern by Richard Thaler (1980) and shown in experiments by Hal Arkes and Catherine Blumer (The Psychology of Sunk Cost, 1985).
Sets aside everything already spent and asks one question: from today, does continuing return more than the best alternative?
Money already spent never moves the forward comparison
A claims system rebuild, a year late. Slide the amount already spent from $0 to $4M: the whole-project figure swings, but the point on the chart only slides sideways, because the money is gone either way. Only the cost to finish, the value of finishing, and the best alternative move the point across the line and change the answer. Illustrative.
From todayFinish it: from today it beats the best alternative.Stop and switch: from today the alternative is worth more.
An example
A regional insurer was two years and $2.5 million into rebuilding its claims system in house. The project was a year late and needed $1.2 million more to finish. The steering committee's first instinct was to press on, since walking away meant writing off everything spent so far.
The finance lead asked them to make the decision as if the project were a new proposal arriving that week. From today, finishing would cost $1.2 million and bring about $1.5 million in savings over the system's life, a net of $0.3 million. A packaged claims system would cost $0.6 million to license and fit and bring about $1.3 million, a net of $0.7 million. The $2.5 million appeared on neither line, because no choice in the room could get it back.
The committee chose the package. It kept the two finished modules that fit the package cleanly and moved four of the in-house developers onto the integration work.
Why it works
Money already spent is gone whichever way the decision goes, so it cannot make one option better than another. Only what each choice will cost and return from now on differs between them. The rule needs stating because people feel the past spending: stopping makes the loss real and visible, while continuing keeps alive the hope of recovering it, and part of the pull is a wish not to look wasteful.
Restating the decision from today removes the number that carries the feeling and leaves the numbers that carry the choice. The test cuts both ways. A project whose forward case is strong survives it, and is protected from a nervous cut by people who only see how much it has cost so far.
The thinking behind it
Richard H. Thaler Richard H. Thaler, a University of Chicago economist, named the pattern in Toward a Positive Theory of Consumer Choice (1980), in the Journal of Economic Behavior and Organization. The paper cataloged ways real consumers depart from the standard theory of choice, among them the endowment effect and the tendency to honor sunk costs, letting a past payment decide how much a purchase gets used. He received the 2017 Nobel Memorial Prize in Economic Sciences for his contributions to behavioral economics.
Hal R. Arkes Hal R. Arkes, a psychologist, documented the effect with Catherine Blumer in The Psychology of Sunk Cost (1985), in Organizational Behavior and Human Decision Processes. They defined it as a greater tendency to continue an endeavor once money, effort, or time has been invested in it. Their account of the cause is the useful part for practice: the effect comes partly from a wish not to appear wasteful, so the pull attaches to money that is already gone.
Catherine Blumer Catherine Blumer, a psychologist and Arkes's coauthor on The Psychology of Sunk Cost (1985), ran its best-known field study with him. Theater patrons who had paid full price for season tickets attended more plays than patrons who had been given discounts, though every one of them held tickets to the same plays. The only difference was how much each had paid, which is the sunk cost effect in its plainest form.
How to run it
- Restate the decision as of today. Write the choice as if the project were a new proposal arriving this week, with today's team, budget, and facts.
- Take the spent money off the page. Note past spending as context, and leave it out of every column the decision compares.
- Price finishing from here. Estimate what is still to spend and what the finished work will return, counting from today.
- Price the best alternative the same way. Name the strongest other use of the same money and people, stopping included, and estimate its net value from today.
- Choose on the forward comparison alone. If the alternative wins, stop or pivot; if finishing wins, continue and record the forward case as the reason.
- Keep what carries over. Finished parts, data, and trained people that help the chosen path count at what they are worth from here.
- Set the next checkpoint. Agree in advance on the date and the numbers at which the forward case will be run again.
Where it goes wrong
Sunk costs hide inside forward estimates. A team asked only for the cost to finish shades it down, or the value up, because stopping feels like losing; spot it when the cost to finish shrinks at every review while the date slips. Have someone outside the project produce the forward numbers.
The opposite error treats everything built so far as worthless. Working modules, clean data, and people who learned the domain can still have value from here, and that value belongs in the forward case.
The third is a review chaired by the person who approved the spending, for whom stopping reads as a confession. Give the forward comparison to someone who did not make the original call, and make the checkpoint routine so stopping is an ordinary outcome.
When to use something else
- Opportunity Cost Reasoning. When the live question is what the same money and people could do elsewhere, with no project yet in trouble.
- Premortem. When the forward case looks strong and needs testing by imagining it has already failed.
- Decision Tree Analysis. When finishing has several uncertain outcomes and the comparison needs a probability on each branch.
Where it sits in SOLVE eX
Go to the source
- Toward a Positive Theory of Consumer Choice, Journal of Economic Behavior and Organization, 1980
- The Psychology of Sunk Cost, Organizational Behavior and Human Decision Processes, 1985