Behavioral Science. Cognitive biases example: my battle with the sunk cost trap

The embarrassing part is not that people make the mistake. The embarrassing part is how intelligent the mistake feels from the inside. The brain puts on a little accountant’s visor, opens a ledger of time, money, effort, reputation, and emotional pain, then reaches the worst possible conclusion: since we have already lost something, we should risk losing more.

That is the sunk cost fallacy. It happens when prior investment drives current commitment, even when continuing has worse expected value than stopping. Rationally, unrecoverable costs should be ignored. Behaviorally, they sit in the room like an unpaid debt collector.

And yes, I call this “my battle” with the sunk cost trap because I have lost to it in precisely the ordinary ways that make the bias so durable: finishing work that should have been killed, defending tools that had become friction machines, staying with plans long after the plan had stopped returning calls. Not heroic failures. Administrative failures. The kind that consume calendars.

The psychology of irrational persistence: why we stay the course

Most people do not persist in bad decisions because they are foolish. They persist because quitting feels like turning a temporary loss into a confirmed loss.

That distinction matters. A failing project still has narrative wiggle room. It can be “early.” It can be “under-optimized.” It can be “one iteration away.” The moment you stop, the story hardens. The money is gone. The time is gone. The effort did not become wisdom by magic. It became a receipt.

The sunk cost trap feeds on this discomfort. It converts stopping into failure and continuing into virtue. Suddenly the person making the bad decision is not being stubborn; they are being committed. They are not escalating; they are honoring the investment. They are not avoiding reality; they are “giving it a fair shot.”

This is where productivity culture often makes the problem worse. It praises persistence as if persistence were always morally superior to revision. But grit and sunk cost persistence are not the same animal.

Grit is staying with a long-term goal despite predictable difficulty because the goal still matters and the path still has plausible returns. Sunk cost persistence is staying with a course of action because leaving would make past waste emotionally undeniable.

Different mechanisms. Different outcomes.

A useful way to separate them:

The trouble is that the brain is not an impartial committee. It is a prediction machine with an ego problem. It wants coherence. It wants past decisions to look sensible. It wants the self-image of “I make good choices” to survive contact with evidence.

So it performs a small trick: it changes the question.

The rational question is: “Given what I know now, would I choose this again?”

The biased question is: “How can I make the earlier choice look less wrong?”

That second question is poison. It does not sound irrational. It sounds responsible. It sounds mature. It sounds like a person taking ownership. In reality, it is cognitive load dressed as loyalty.

The sunk cost trap is not a failure to think. It is thinking with the wrong timestamp.

Loss aversion and the 2:1 engine behind the bias

The sunk cost fallacy gets much of its horsepower from loss aversion, a central idea in Prospect Theory, introduced by Daniel Kahneman and Amos Tversky in 1979. The rough behavioral finding is blunt: losses hurt about twice as much as equivalent gains feel good.

That 2:1 ratio is not a law of physics. Humans are not vending machines. But as a heuristic, it explains why people do bizarre things to avoid accepting a loss that has already happened.

Lose $100 and it stings. Gain $100 and it helps, but it usually does not produce equal and opposite joy. The pain has sharper teeth. So when a decision threatens to label previous effort as wasted, the brain scrambles to avoid that label.

This is why a person can spend three more hours watching a terrible movie trilogy because they already watched the first installment. Why a company can keep funding a product that customers politely ignore. Why a student can stay in a major they dislike because switching would mean “wasting” two years, even though staying may waste four.

Loss aversion makes the old investment feel alive. But sunk costs are dead. That is the whole point. They cannot be recovered by continuing. You can only decide whether to add new costs on top.

This sounds obvious on paper. In practice, the mind refuses to treat the past as past. It drags old costs into current decisions because old costs carry emotional weight. The ledger may be economically irrelevant, but psychologically it is glowing red.

Here is the nasty little sequence:

1. You invest resources. Time, money, social capital, attention, identity. The category barely matters.

2. The expected return deteriorates. The project stops working, the relationship becomes corrosive, the habit becomes maintenance without benefit.

3. Stopping would create a visible loss. Not a theoretical one. A real one you have to explain to yourself and possibly to other people.

4. Continuing offers emotional delay. The loss remains unbooked. Hope stays technically alive.

5. The next investment gets justified by the previous one. This is the trap door. At this point, the past has hijacked the future.

Notice what is missing from that sequence: evidence that continuing is the best option.

That is why sunk cost decisions often have a strange emotional signature. They feel tense, defensive, and oddly humorless. Ask someone why they are continuing and they will tell you what they have already sacrificed. They will not describe the future payoff in crisp terms. The argument points backward.

That backward-pointing argument is the tell.

From Concorde to daily life: recognizing the sunk cost trap

The classic example is the Concorde, the supersonic passenger jet backed by the British and French governments. The project became so associated with irrational persistence that “Concorde fallacy” is now a common name for the sunk cost trap. Even after it became clear the aircraft would not be commercially viable, funding continued because the investment already made was enormous.

This example is useful, but also slightly dangerous. It lets ordinary people keep the bias at museum distance. Governments, prestige projects, aviation budgets — very grand, very expensive, very easy to judge from a chair.

Daily sunk costs are smaller and therefore better camouflaged.

A sunk cost fallacy real life example rarely announces itself with a boardroom memo. It sounds more like:

• “I already bought the annual subscription.”
• “I’ve been doing this for six months; stopping now would be stupid.”
• “We can’t change systems after training everyone.”
• “I waited this long, so I might as well stay.”
• “I know it is not working yet, but we’ve invested too much to quit.”

That last sentence is practically the bias wearing a name tag.

One of the most common cognitive bias examples in daily life is the bad tool that survives because it was hard to implement. A team adopts a project management platform. Migration takes weeks. Everyone hates it. The tool adds friction instead of removing it. Tasks now require more clicks, more meetings, more performative status updates. But abandoning it would mean admitting the migration was a waste, so the team creates training sessions, process documents, and “best practices” to compensate for the fact that the default environment is badly designed.

This is not optimization. It is a rescue mission for a prior decision.

The same pattern appears in personal habits. Someone buys an expensive course and continues pushing through low-quality lessons because quitting would mean admitting the purchase was poor. A runner keeps using shoes that cause pain because they were not cheap. A reader finishes a book they resent because the bookmark is too far from the front cover. These are not catastrophic decisions, but they train the same machinery: past cost becomes a command.

And then there are identity sunk costs, the most expensive kind because no receipt exists.

A person stays in a career path because they have spent a decade becoming “the kind of person who does this.” A founder keeps a company alive because closing it would threaten the self-concept of being resilient. A professional clings to a method they mastered years ago because replacing it would demote their expertise.

Identity raises the switching cost. Not technically. Psychologically.

When the decision is tangled with who you think you are, the sunk cost trap gains body armor.

If the strongest argument for continuing is how painful quitting would feel, you are not making a forecast. You are negotiating with embarrassment.

Evolutionary roots: why our brains struggle to let go

The sunk cost effect is not just a quirky human spreadsheet error. Research has found sunk cost-like behavior in non-human animals, including rats and mice. That should make us a little less smug and a little more careful.

If animals show versions of this behavior, the mechanism is probably not just cultural pride or MBA overconfidence. It may reflect deeper decision rules that were useful often enough to survive: do not abandon food patches too quickly, do not waste effort already spent searching, do not constantly reset when persistence might pay off.

In unstable environments, persistence can be adaptive. If every obstacle caused immediate abandonment, nothing difficult would ever get done. The organism that quits too early may starve. The person who quits every difficult project becomes a tourist in their own life.

So the problem is not persistence itself. The problem is an outdated heuristic applied without context.

A heuristic is a shortcut. Shortcuts are not dumb. They are efficient. They reduce cognitive load when the world moves faster than careful analysis. But every shortcut has a failure mode, and the sunk cost heuristic fails when it treats past investment as evidence of future value.

Modern environments intensify the problem because they manufacture partial commitments everywhere. Apps use streaks. Platforms sell annual plans. Careers reward specialization. Social networks make public declarations cheap and reversals embarrassing. Organizations build processes that become harder to question the longer they exist.

The design of the environment keeps asking: “Wouldn’t it be a shame to stop now?”

That question is manipulation-friendly. It transforms friction into commitment. It turns accumulated inconvenience into proof that the thing must matter.

A better question is colder: “What is the expected return from the next unit of investment?”

Not the last unit. Not the first unit. The next one.

That is the rational actor’s threshold. Costs that cannot be recovered should not control the decision. Of course, humans are not rational actors in the clean textbook sense. We are animals with calendars, reputations, inboxes, and a talent for self-justification. So we need decision environments that assume bias will appear.

Not pep talks. Guardrails.

How to identify cognitive biases before they become expensive

“How to identify cognitive biases” is usually framed as an introspection problem: look within, examine your thoughts, become more aware. Fine. Awareness helps. But awareness alone is a weak control system. The mind that created the justification should not be the only mind auditing it.

A practical system needs external prompts, pre-set thresholds, and a way to reduce the shame of stopping. Shame is fuel for sunk costs. Lower the shame, and the decision gets less sticky.

Start by listening for backward logic. Sunk cost reasoning has a grammatical pattern. It leans on phrases like “after all this,” “we already,” “too late,” and “can’t waste.” These phrases are not always wrong, but they are smoke. Follow the smoke.

Then force a present-tense decision. Ask: “If I were not already in this, would I choose it today?” This question is annoying because it strips away the emotional subsidy from past effort. Good. Annoyance is sometimes the sound of a bias losing its grip.

For higher-stakes decisions, use a small decision audit:

1. Write the next investment in concrete terms. Not “keep going.” Say “spend three more months,” “allocate another $20,000,” “give this another five hours every week,” or “stay in this role through the next review cycle.” Vague continuation is where bad decisions breed.

2. Name the opportunity cost. What else could those resources do? A different project, rest, training, a better hire, a cleaner system, a serious conversation. Opportunity cost is the antidote because it makes the invisible alternative visible.

3. Separate learning value from continuation value. A failed project may have taught you something. That does not mean it deserves more funding. Education and escalation are different line items.

4. Ask an outsider question. If a competent stranger inherited this situation today, with no need to defend the original choice, what would they do?

5. Set a kill criterion before the next investment. Decide what evidence would make you stop. If no possible evidence would change your mind, stop pretending this is analysis.

That last point is where many people discover the truth. They are not evaluating. They are protecting.

A cognitive bias journal can help here, if it is used as a decision tool rather than a feelings scrapbook. The point is not to write elegant reflections about human imperfection. The point is to build a searchable record of bad defaults.

A useful entry is short:

This is not glamorous. That is the advantage. Glamour attracts narrative. A table attracts accuracy.

The journal also exposes patterns. Some people sink costs into relationships because conflict avoidance is their default. Others sink costs into tools because they hate migration. Others sink costs into goals because public identity is involved. The bias has a general mechanism, but the doorway is personal.

Once you know your doorway, you can put friction there.

Reframing decisions: from past costs to future utility

The central move is simple and hard: shift attention from past costs to future utility.

Past costs ask, “How do I make this feel less wasted?”

Future utility asks, “What choice has the best expected value now?”

The first question protects the ego. The second protects the calendar.

This reframing works best when it is embedded in defaults. If you rely on heroic clarity in the exact moment when embarrassment is high, good luck. That is not a system. That is a motivational poster with a scheduling problem.

Use pre-commitments instead. Before starting a project, define what success would look like, what failure would look like, and when the review happens. The review date matters. Without it, evaluation gets postponed by optimism, busyness, and the conveniently urgent nonsense that fills adult life.

For example:

• A new software tool gets a 60-day trial with three metrics: time saved, adoption rate, and error reduction. If it fails two of three, the default is removal, not another committee.
• A personal learning project gets ten sessions. After ten, the decision is based on retention, usefulness, and continued interest — not the price of the course.
• A business experiment gets a budget ceiling and a stop rule. If demand does not appear under defined conditions, the team captures the lesson and reallocates resources.
• A book gets 50 pages unless there is a specific reason to continue. Life is too short to let completionism cosplay as integrity.

The purpose is not to become a quitter. The purpose is to stop making quitting emotionally impossible.

There is also a language change worth making. Instead of saying “I wasted six months,” say “I bought information with six months.” Sometimes that information was overpriced. Fine. Still, once the information has arrived, refusing to act on it is not maturity. It is paying twice.

This distinction is especially useful in organizations, where sunk costs hide inside status games. Nobody wants to be the person who says the initiative failed. So meetings become theater. Metrics get softened. Deadlines become “phases.” The project continues because stopping would require someone to absorb blame.

A sane organization makes reversals less humiliating. It treats clean exits as competence, not betrayal. It rewards people for updating their beliefs when evidence changes. This sounds obvious. It is rare.

At the individual level, you can do the same thing by designing a personal fail-safe: a rule that removes some discretion when your bias is likely to be loud.

Try this:

• If the only reason to continue is previous investment, stop.
• If the next investment would not be attractive as a fresh decision, stop.
• If you cannot describe the future payoff without mentioning the past cost, pause.
• If the opportunity cost is clearly better and the current path survives only through guilt, switch.
• If stopping teaches you something and continuing only delays the lesson, take the lesson.

Blunt rules are useful because biased thinking loves nuance. It will negotiate forever if you let it.

The uncomfortable win: abandoning the wrong thing on purpose

The sunk cost trap survives because it offers emotional credit. Keep going and you get to postpone grief, embarrassment, and the administrative burden of changing course. Stop and you have to face the bill.

But the bill does not disappear because you refuse to open it.

A good decision system does not assume you will be wise at the exact moment your ego is threatened. It assumes you will be human: loss-averse, status-sensitive, allergic to waste, and capable of building elaborate arguments around a bad default. Then it narrows the path. It makes the better decision easier to take and the biased decision harder to justify.

That is the practical lesson hiding inside this cognitive biases example. The enemy is not investment. The enemy is letting yesterday’s investment dictate tomorrow’s allocation.

Persistence is valuable when the future still deserves funding. When it does not, persistence becomes a very expensive way to avoid saying “I was wrong.”

Say it earlier. Spend less. Move the resources. Let the dead cost stay dead.