Quality
and the Sunk Cost Fallacy: When Your Organization Keeps Throwing Good
Money After Bad — and the Investment You’re Trying to Protect Becomes
the Anchor That Sinks You
The boardroom was quiet. Not the comfortable kind of quiet — the kind
that settles in when everyone in the room knows the answer but nobody
wants to say it out loud.
On the table sat the numbers for Project Horizon, the quality
management system they had been implementing for eighteen months. The
original budget was €400,000. They had spent €1.2 million so far. The
go-live date had been pushed back four times. The supplier they had
chosen was clearly not delivering. The system was unstable in testing.
Three of the five production plants had gone back to their old processes
because the new one kept crashing during critical operations.
And the Quality Director, a man named Tomas who had staked his
reputation on this project, was asking for another €350,000 and six more
months.
“We’ve already invested so much,” he said. “We can’t just walk away
from that.”
And there it was. The five words that have destroyed more quality
initiatives than incompetence, apathy, and bad luck combined.
We’ve already invested so much.
The Psychology of Escalation
The sunk cost fallacy is one of the most well-documented cognitive
biases in behavioral science, and one of the most dangerous forces
acting on quality organizations. At its core, it’s simple: once you’ve
invested time, money, or reputation into something, you feel compelled
to continue — even when the rational choice is to stop.
The bias was first formally described by researchers Hal Arkes and
Catherine Blumer in 1985, though the phenomenon had been observed for
centuries. Their experiments showed that people consistently make worse
decisions when they factor in costs they’ve already incurred — costs
that, by definition, cannot be recovered.
In a classic study, participants were told they had bought a €100
ticket for a ski trip to Michigan and later found a €50 ticket for a ski
trip to Wisconsin on the same weekend. The Wisconsin trip was described
as better in every way. When told they would lose the €100 either way,
most participants still chose the Michigan trip. The money was gone
regardless. The rational choice was the better trip. But the pain of
“wasting” the larger investment pulled them toward the inferior
option.
Now scale that up from a ski trip to a million-dollar quality system
implementation. From a €50 dinner to a three-year process improvement
initiative. From a personal decision to one that involves teams,
departments, reputations, and careers. The fallacy doesn’t get smaller
at scale — it gets far, far more powerful.
Why Quality
Organizations Are Especially Vulnerable
Quality functions face a unique confluence of factors that make them
ground zero for the sunk cost fallacy.
Reputation is intertwined with decisions. When a
Quality Director champions a new system, a new methodology, or a new
supplier, their professional credibility gets bundled into the choice.
Abandoning the initiative doesn’t just mean acknowledging a mistake — it
means admitting it publicly, to superiors, to peers, to the shop floor
that watched the rollout. The personal cost of reversal is enormous, and
it creates a powerful incentive to keep pushing forward.
Quality investments are hard to evaluate mid-stream.
Unlike a manufacturing line where you can count units per hour, quality
improvement projects often have ambiguous metrics during implementation.
“Are we making progress?” becomes a matter of interpretation. This
ambiguity provides cover for escalation — you can always argue that
you’re almost there, that the next phase will turn things around, that
the worst is behind you.
Sunk costs are often invisible. The €1.2 million in
the Project Horizon example is the obvious cost. But what about the
4,200 hours of shop-floor operator time spent in training sessions for a
system they eventually abandoned? What about the opportunity cost of
three other improvement initiatives that were deferred because resources
were tied up? What about the erosion of trust in the quality function
when the team sees leadership doubling down on a failing project? These
costs rarely appear on any dashboard, but they compound silently.
Organizations reward persistence. Corporate culture
celebrates resilience. “Don’t give up.” “See it through.” “Champions
finish what they start.” These are beautiful sentiments when you’re
running a marathon. They are catastrophic when you’re implementing the
wrong quality system. The cultural framing of quitting as failure and
persistence as virtue creates an environment where the sunk cost fallacy
doesn’t just survive — it’s actively rewarded.
The Escalation Dynamic
What makes the sunk cost fallacy so insidious in quality contexts is
that it doesn’t arrive as a single bad decision. It arrives as a series
of small, individually defensible choices that compound into
disaster.
Here’s how it typically unfolds:
Month 0: The organization selects a quality system
or methodology. The decision is made in good faith with available
information. The budget is approved. The team is assembled. Everything
feels right.
Month 6: Early warning signs appear. Implementation
is slower than expected. The supplier misses a deadline. User acceptance
testing reveals more issues than anticipated. But the team has already
invested six months of work and significant budget. “It’s normal for
implementations to have hiccups,” the project sponsor says. This is
often true. The decision to continue is defensible.
Month 12: The problems are no longer hiccups.
They’re systemic. The budget is 60% over. The timeline has doubled. But
now the investment is substantial, and the people who championed the
project have a personal stake in its success. Walking away means writing
off a year of work and explaining to the board why €750,000 produced
nothing. The decision to continue is driven by loss aversion — the pain
of losing what was invested outweighs the rational assessment of whether
the project can still succeed.
Month 18: We’re in full escalation territory. The
project has become too big to fail — not because it’s strategically
important, but because the cost of admitting failure has become
unacceptable. Every additional investment is justified not by the
probability of success but by the desire to make the previous
investments meaningful. The organization is no longer investing in a
solution. It’s investing in a narrative.
Month 24: The project either launches in a
compromised state that delivers a fraction of its promised value, or
it’s finally killed in a painful, acrimonious process that leaves scars
on the organization for years. The total cost is often three to five
times the original budget, and the opportunity cost of everything that
wasn’t done during those two years is incalculable.
I’ve seen this movie play out in automotive plants in Slovakia,
pharmaceutical facilities in Switzerland, aerospace suppliers in the UK,
and electronics manufacturers in Southeast Asia. The names change. The
numbers change. The dynamic is always the same.
The Anatomy of a Sunk Cost
Trap
Not every struggling project is a sunk cost trap. Some projects
genuinely do need more time and more investment. The challenge is
distinguishing between legitimate perseverance and irrational
escalation. Here are the signals that separate the two:
You’re justifying the investment, not the outcome.
When someone asks whether the project will succeed and your first
response involves how much has already been spent, you’re in the trap. A
rational assessment of a project’s viability should be based on future
costs and future benefits — not past expenditures. The money is gone
regardless of what you do next.
The justification has shifted. Projects that start
with clear business cases — “this will reduce defect rates by 40%” —
often morph into different justifications when they struggle. “It will
improve our digital maturity.” “It positions us for future
requirements.” “The lessons we’re learning are valuable.” When the
original rationale collapses but the project continues with a new,
vaguer justification, the sunk cost fallacy is driving the bus.
Nobody wants to write the report. If the most
compelling reason to continue is that killing the project would require
explaining the failure to senior leadership, the board, or the customer,
you’re not making a quality decision. You’re making an emotional
one.
The advocates have become the audience. In healthy
projects, the project team is constantly testing assumptions and looking
for disconfirming evidence. In sunk cost traps, the team is performing
for stakeholders — presenting progress, managing perceptions, and
controlling the narrative. The question shifts from “Is this working?”
to “How do we show this is working?”
The alternatives keep getting better. Here’s a
painful one: sometimes the reason a project is struggling is that better
alternatives have emerged since it started. The supplier landscape
shifted. A new technology became viable. A competitor solved the same
problem for a tenth of the cost. Continuing with the original plan in
the face of better options is the sunk cost fallacy at its most
expensive.
The Quality-Specific Damage
The sunk cost fallacy doesn’t just waste money. In quality contexts,
it causes specific, lasting damage that extends far beyond the project
itself.
Trust erosion. When the shop floor watches
leadership pour resources into a failing initiative for months or years,
they learn something about the organization. They learn that admitting
mistakes is harder than enduring them. They learn that saving face
matters more than getting it right. And they carry that lesson into
every future initiative — which means the next quality improvement
effort, even a good one, starts with a trust deficit.
Opportunity destruction. Every euro and every hour
spent on a sunk cost trap is a euro and an hour not spent on something
that could actually improve quality. The visible cost of a failed
project is the budget that was spent. The invisible cost is the
improvements that were never made because the resources were tied up. In
my experience, the invisible cost is always larger.
Methodology contamination. When an organization
forces a struggling methodology to completion and then declares victory,
the methodology itself gets contaminated. People don’t experience it as
a genuine improvement — they experience it as a management exercise. The
next time you try to implement something similar, even if it’s the right
approach for that situation, you’re fighting the memory of the last
failed rollout.
Talent loss. The best quality professionals — the
ones you most need to retain — are the ones most frustrated by
irrational escalation. They can see the trap forming. They raise
concerns. They’re told to be team players. Eventually, some of them
leave. And the organization loses not just their skills but their
judgment — exactly the judgment that would have prevented the next sunk
cost trap.
Breaking the Cycle
So how do you protect your quality organization from the sunk cost
fallacy? The answer is not simply “be more rational.” The fallacy is a
cognitive bias, which means it operates below the level of conscious
awareness. Fighting it requires structural interventions, not just
individual willpower.
Build exit criteria into every project at the start.
Before a single euro is spent, define the conditions under which the
project will be stopped. Not paused. Not restructured. Stopped. Write
these criteria down, get them approved by the sponsor, and review them
at every milestone. The key is to make these decisions before the
emotional investment clouds judgment — and then to hold yourself to them
even when every instinct screams to continue.
This is psychologically similar to the premortem technique: by
imagining failure before it happens and agreeing on what failure looks
like in advance, you create a shared framework for stopping that doesn’t
require anyone to admit personal failure. The criteria triggered. It’s
not about blame — it’s about discipline.
Separate the decision-maker from the advocate. The
person who championed a project should not be the sole person deciding
whether to continue it. This isn’t about distrust — it’s about
acknowledging that cognitive biases are universal and that proximity to
a decision creates attachment. Create a review mechanism where someone
without personal investment in the original choice can evaluate the
project on its current merits.
In practice, this might mean a quarterly portfolio review where an
independent panel evaluates all active projects against their original
business cases. The questions should be simple: Given what we know
today, would we start this project from scratch? If not, why are we
continuing it?
Track decision quality, not just project progress.
Most organizations track whether projects are on time and on budget. Few
track whether the original decisions to undertake them were sound.
Adding a decision-quality metric — are we making decisions based on
future expected value or past incurred cost? — creates awareness of the
fallacy itself. What gets measured gets managed.
Normalize project termination. This is perhaps the
most important intervention. If the only way a project ends is success
or scandal, you’ve created an environment where every struggling project
will be dragged to completion. But if project termination is a normal,
expected part of portfolio management — if killing a project is seen as
a sign of disciplined thinking rather than failure — then the emotional
cost of stopping drops dramatically, and the sunk cost fallacy loses its
power.
Some of the best organizations I’ve worked with celebrate killed
projects. They hold retrospectives focused on what was learned. They
recognize the courage it takes to stop something you’ve invested in.
They explicitly tell their teams: “We will start ten initiatives this
year. We expect three to succeed, four to pivot, and three to be
stopped. That’s not failure. That’s strategy.”
Use reference class forecasting. When evaluating
whether to continue a struggling project, don’t just look at that
project in isolation. Look at similar projects across the industry. What
percentage of quality system implementations with these characteristics
actually succeed? What was the typical cost overrun? Reference class
forecasting — a technique developed by psychologists Daniel Kahneman and
Amos Tversky — forces you to ground your expectations in reality rather
than hope.
If 70% of projects like yours fail, and yours is showing the same
warning signs, the rational assessment is that yours will probably fail
too. The €800,000 you’ve already spent doesn’t change that probability.
It just makes it more painful to accept.
The Personal Dimension
I need to be honest about something: writing about the sunk cost
fallacy is easy. Living it is brutally hard.
I’ve been the person in the boardroom arguing for more time and more
budget for a project I knew, deep down, was failing. I’ve felt the knot
in my stomach when someone asked a question I didn’t want to answer.
I’ve experienced the strange relief of being told to continue — the
reprieve, the temporary escape from the conversation I was dreading.
And I’ve also been on the other side: the person who killed a project
that a colleague had poured two years of their life into. The
conversation where you look someone in the eye and say, “This isn’t
working, and we need to stop.” The look on their face. The silence
afterward.
Neither role is comfortable. But I’ve learned that the discomfort of
stopping is always less than the damage of continuing. Always. The
organizations I’ve seen thrive — the ones that sustain quality
excellence over decades, not just quarters — are the ones that have
built the muscle to stop. They don’t enjoy it. They don’t celebrate it.
But they do it, reliably, because they understand that the cost of a
wrong decision is always less than the cost of a wrong decision
defended.
The Framework: Four Questions
When you suspect you’re in a sunk cost trap — with a quality system,
a methodology, a supplier relationship, a training program, any
investment — ask yourself these four questions:
1. If we hadn’t already invested anything, would we start
this project today? This is the purest test. It strips away the
emotional weight of past investment and forces you to evaluate the
project on its future merits. If the answer is no, you have your answer.
The money you’ve spent is gone regardless. The money you haven’t spent
yet can still be redirected.
2. What could we do with the resources if we
stopped? This question reframes stopping as a positive choice
rather than a loss. You’re not “giving up” — you’re freeing up €350,000
and six months of team capacity for something better. The opportunity
cost of continuing is the best alternative you could pursue instead.
Make that alternative visible and concrete.
3. Who is advocating for continuation, and what do they have
at stake? This isn’t about questioning anyone’s integrity. It’s
about recognizing that personal stakes create unconscious bias. If the
strongest advocates for continuing are the people whose reputations are
most tied to the project’s success, their judgment is compromised — not
because they’re dishonest, but because they’re human.
4. What would I advise a peer in the same situation?
Distance improves judgment. When you imagine giving advice to a friend
or colleague facing the same dilemma, the sunk cost fallacy often
evaporates. You’d tell them to stop. You’d tell them the money is gone.
You’d tell them to cut their losses and move on. The question is whether
you’re willing to take your own advice.
The Real Cost of Sunk Costs
Let me return to Project Horizon for a moment.
The organization did eventually kill the project — at the
twenty-six-month mark, after spending €1.6 million. The decision came
only because a new COO joined the company, reviewed the portfolio with
fresh eyes, and asked the simple question: “If we were starting from
scratch today, would we choose this system?”
The answer was obviously no. The system was discontinued. The
supplier relationship was terminated. The team was reassigned.
But here’s what the COO also discovered: during those twenty-six
months, three other quality improvement initiatives had been shelved
because resources were tied up in Horizon. One was a statistical process
control upgrade that would have reduced a critical defect rate by an
estimated 35%. Another was a supplier quality program that would have
prevented two customer complaints that later turned into lost contracts
worth €2.3 million annually. The third was a training program that would
have addressed a skills gap that contributed to a regulatory audit
finding.
The €1.6 million spent on Project Horizon was visible. The €2.3
million in lost contracts, the regulatory risk, and the unresolved
defect rate were invisible — until they weren’t.
That’s the real cost of the sunk cost fallacy in quality. It’s not
just the money you waste. It’s the excellence you never achieve because
you were too busy defending a decision that was wrong.
A Final Word
There is a concept in economics called the contrite
pragmatist — someone who acknowledges past mistakes honestly,
adjusts course without drama, and focuses relentlessly on future value.
It’s not a glamorous archetype. There’s no hero’s journey in cutting
your losses. Nobody writes case studies about the project you wisely
decided not to complete.
But in twenty-five years of quality work across automotive,
aerospace, and pharmaceutical industries, I can tell you this: the
organizations that last are not the ones that never fail. They’re the
ones that fail fast, learn honestly, and redirect their resources toward
what actually works.
The sunk cost fallacy tells you to keep going because you’ve already
come so far.
Quality tells you to stop going in the wrong direction, no matter how
far you’ve come.
Listen to quality.
Peter Stasko is a Quality Architect with 25+ years
of experience transforming organizations across automotive, aerospace,
and pharmaceutical industries. He has led quality system
implementations, process improvement initiatives, and cultural
transformations on three continents — and has killed more failing
projects than he cares to count, each time more grateful that he
did.
