Quality
and the Sunk Cost Fallacy: When Your Organization Keeps Throwing Good
Money After Bad — and the Investment You’ve Already Made Becomes the
Reason You Refuse to Abandon the Process That’s Failing
The Meeting
That Should Have Ended Six Months Ago
It was a Tuesday morning in March when the quality director at a
mid-size automotive supplier finally said what everyone in the room had
been thinking for half a year.
“We need to kill the project.”
The silence that followed was not disagreement. It was relief. Six
people around the table — engineers, managers, a VP of operations — had
spent months defending a new inspection system that was supposed to
revolutionize their defect detection. They had poured $2.3 million into
it. They had reassigned their best people. They had told the CEO it
would reduce customer complaints by 40 percent.
It had reduced nothing. Customer complaints were up 12 percent since
installation. The system produced false positives at a rate that
paralyzed production. And every month, the team met to discuss
“optimization” — a word that meant “we’re going to keep trying because
we’ve already spent too much to stop.”
That $2.3 million was gone whether they continued or quit. But the
money wasn’t the real trap. The trap was the story they had told
themselves about what that money meant. It meant commitment. It meant
they were serious. It meant they couldn’t possibly walk away now.
That’s the sunk cost fallacy. And it is one of the most expensive
cognitive biases in quality management.
What the Sunk Cost
Fallacy Actually Is
The sunk cost fallacy is the tendency to continue investing in a
decision — a process, a tool, a strategy, a supplier relationship — not
because the future returns justify it, but because you’ve already
invested resources you can’t recover.
The logic is seductive and broken: “We’ve come too far to stop now.”
“We’ve invested too much to switch.” “After all this effort, it has to
work.” These sentences feel like determination. They are actually
rationalization.
In personal life, this looks like staying in a bad movie because you
paid for the ticket. In quality management, it looks like maintaining a
failing quality system because you spent eighteen months implementing
it. The scale is different. The psychology is identical.
The economist’s definition is clean: a sunk cost is a cost that has
already been incurred and cannot be recovered. Rational decision-making
says these costs should have zero influence on future choices. Only
future costs and future benefits should matter. But human beings are not
rational decision-makers, and organizations — which are collections of
human beings operating under shared narratives — are even less rational
than the individuals within them.
Why Quality
Organizations Are Especially Vulnerable
Quality departments are uniquely susceptible to the sunk cost fallacy
for several structural reasons.
First, quality investments are often invisible. When you buy a new
CNC machine, you can see it on the floor. When you implement a new SPC
system, the value is in the data and the decisions it enables — things
that are harder to point to and say “this is what we got for our money.”
This ambiguity makes it easier to tell yourself the investment is about
to pay off, even when the evidence says otherwise.
Second, quality improvements have long feedback loops. Unlike a
marketing campaign where you see results in weeks, a quality culture
initiative might take years to show measurable impact. This long horizon
creates a natural narrative: “It hasn’t worked yet, but we just need
more time.” Sometimes that’s true. Sometimes it’s the sunk cost fallacy
wearing a patient expression.
Third, quality professionals are, by training and temperament, people
who don’t give up. They are problem-solvers. They believe that with
enough analysis, enough data, enough effort, every problem has a
solution. This is admirable. It is also exactly the psychological
profile that the sunk cost fallacy exploits.
Fourth, and perhaps most importantly, quality investments are often
tied to organizational identity. When a company rolls out a new quality
management system, it’s not just buying software. It’s making a
statement about who it is and what it values. Abandoning that system
isn’t just a financial decision — it’s an admission that the statement
was wrong. The ego cost of admitting a mistake amplifies the sunk cost
effect dramatically.
Five Forms
the Sunk Cost Fallacy Takes in Quality
1. The Failing Technology
Investment
This is the most obvious form. An organization invests in a new
quality platform — a QMS, an automated inspection system, a predictive
analytics tool — and it doesn’t deliver. But instead of evaluating
whether the technology is the right fit, the organization doubles down.
More training. More customization. More integration work. Each
additional investment makes it harder to walk away, because walking away
means admitting that all of it — the original purchase and every dollar
and hour since — was wasted.
I watched a pharmaceutical company spend three years trying to make a
cloud-based QMS work for their GMP environment. The system was designed
for a different regulatory framework. It couldn’t handle their batch
record requirements without extensive customization, and each
customization broke something else. After $4 million and countless
frustrated man-hours, they finally replaced it with a system that should
have been chosen in the first place. The total cost of the wrong
decision, including the remediation, was close to $9 million.
2. The Failed Process Redesign
A company redesigns its manufacturing process to achieve tighter
tolerances. The new process looks elegant in theory. In practice, it
produces more variation than the old one. But the old process has
already been dismantled, the operators have already been retrained, and
the documentation has already been rewritten. Going back feels like
regression, not correction.
This is particularly common in lean transformations. An organization
maps its value stream, identifies waste, redesigns the flow — and the
new flow has problems the old one didn’t. Maybe the cells are too small.
Maybe the single-piece flow creates bottlenecks that batch processing
hid. The response is often to “lean harder” rather than to honestly
assess whether the specific lean tool was appropriate for this specific
situation.
3. The
Underperforming Supplier Relationship
You’ve qualified a supplier. You’ve audited them, trained them, sent
your engineers to their facility, integrated their components into your
designs. Their quality is mediocre. Their delivery is inconsistent. But
you’ve invested so much in the relationship that switching feels
impossible. The qualifying effort for a new supplier looms as a cost,
while the cost of staying — the defects, the delays, the fire-fighting —
is distributed across dozens of small incidents that never seem serious
enough individually to trigger a change.
This is one of the most insidious forms because the switching costs
are real. But the sunk cost fallacy makes you overestimate them. You
focus on what you’ve already invested in the relationship and
underestimate both the cost of staying and the real effort required to
switch. Often, a new supplier can be qualified in a fraction of the time
you imagine, because your organization has improved its qualification
processes since the last time it went through them.
4. The Outdated Standard
An organization built its quality management system around an older
version of a standard — say, ISO 9001:2008. The new version (ISO
9001:2015) introduces risk-based thinking, leadership engagement, and
organizational context in ways that require fundamentally different
approaches. But the organization has years of documentation, procedures,
audit records, and training built around the old framework. Instead of
rethinking the system from first principles, they do the minimum
necessary to “tick the boxes” for the new standard, preserving the
structure they’ve already invested in rather than building what would
actually work best.
The result is a quality system that technically meets the
requirements of the new standard but captures none of its intent. It’s
the quality equivalent of painting a new facade on a crumbling
building.
5. The Failing Quality
Culture Initiative
This is perhaps the most painful form. An organization launches a
quality culture transformation — new values, new behaviors, new reward
systems, new communication strategies. Six months in, engagement scores
haven’t moved. Defect rates haven’t changed. But the initiative has
become part of the company’s identity. Leadership has staked their
credibility on it. Walking away would mean admitting that the approach
was wrong.
So they rebrand it. Same initiative, new name, refreshed materials.
The sunk cost isn’t just money — it’s political capital, organizational
energy, and leadership face. These are currencies that are even harder
to write off than dollars.
The Mathematics of Letting
Go
Here is a simple framework for evaluating whether you’re being
influenced by sunk costs. Ask yourself two questions:
-
If I had not already invested anything in this decision, would I
choose to invest in it today based on what I know now? -
Will the future benefits of continuing exceed the future costs of
continuing?
If the answer to the first question is no and the answer to the
second question is no, you should stop. Full stop. The money, time, and
effort you’ve already spent are gone regardless of what you do next. The
only question is whether you’re going to waste more.
This is easy to understand intellectually and nearly impossible to
execute emotionally. That’s why organizations need structural mechanisms
— not just individual willpower — to overcome the sunk cost fallacy.
Structural
Defenses Against Sunk Cost Thinking
Independent Review Gates
Every significant quality investment should have predefined review
points where an independent team evaluates whether to continue.
“Independent” is the critical word. The people who championed the
investment, who implemented it, who have their professional reputations
tied to its success — those people cannot be objective. You need fresh
eyes with no emotional investment in the original decision.
These gates should be scheduled before the investment begins, not
triggered by failure. If the review is triggered by failure, the meeting
becomes about explaining the failure rather than evaluating the
future.
Precommitment to Exit
Criteria
Before starting any major quality initiative, define the conditions
under which you will stop. Write them down. Get leadership sign-off.
Make them specific and measurable: “If the defect rate has not decreased
by at least 15 percent within twelve months, we will evaluate
alternative approaches.”
This sounds simple. It is rarely done. The reason is that nobody
wants to plan for failure at the beginning of an exciting initiative.
But precommitment to exit criteria is not planning for failure — it’s
planning for rationality. It’s acknowledging that you don’t know the
future and building a structure that protects you from your future
self’s tendency to justify past decisions.
Separate the Decision-Makers
The people who decide to continue an investment should not be the
same people who made the original decision. This is a governance
principle, not a personal judgment. It’s the same reason financial
auditors are independent of the departments they audit. When your
reputation is tied to a decision, you cannot evaluate that decision
objectively.
In practice, this might mean that the quality council that approves a
new SPC system is different from the steering committee that evaluates
its performance after implementation.
Regular Zero-Based Reviews
Once a year, evaluate every significant quality investment as if you
were making the decision for the first time. Don’t ask “Is it working?”
Ask “If we didn’t have this, would we choose to acquire it?” This is
zero-based budgeting applied to quality systems, and it forces you to
distinguish between value and familiarity.
Many organizations discover, through this exercise, that they’re
maintaining systems, processes, and supplier relationships that survive
purely on institutional momentum. The reason they exist is not that
they’re the best option. The reason they exist is that they already
exist.
The Emotional Dimension
I want to be honest about something. Overcoming the sunk cost fallacy
is not primarily an intellectual challenge. You already know,
rationally, that past investments shouldn’t influence future decisions.
That knowledge doesn’t help, because the sunk cost fallacy operates
through emotions — pride, embarrassment, loyalty, fear of admitting
mistakes.
The quality director who finally said “We need to kill the project”
in that Tuesday morning meeting wasn’t the smartest person in the room.
Everyone in that room knew the project should end. He was simply the
first person willing to absorb the emotional cost of saying it out
loud.
Organizations that handle sunk costs well create cultures where
admitting a wrong decision is not punished — it’s respected. Where “we
tried this and it didn’t work” is treated as valuable information rather
than failure. Where leaders model the willingness to change course
publicly and without shame.
If your culture punishes people for abandoning failed initiatives,
you will never overcome the sunk cost fallacy. People will continue
throwing good money after bad not because they’re stupid, but because
the alternative — admitting that the money was wasted — is culturally
more expensive than wasting more.
The Paradox of Persistence
There is a tension here that must be acknowledged. Sometimes
persistence is exactly the right strategy. Quality improvement is
genuinely difficult. New systems genuinely take time to show results.
The organization that gives up too early on every initiative never
builds momentum.
How do you distinguish productive persistence from sunk cost
rationalization? Look for three signals:
Signal 1: The evidence is getting worse, not better.
If the metrics are moving in the wrong direction and the trend is
accelerating, persistence is probably denial. If the metrics are slowly
improving or the rate of improvement is increasing, persistence is
probably justified.
Signal 2: The explanations keep changing. When a
project is genuinely on track, the explanation for delays is usually
consistent. When a project is failing, the explanation shifts every
month — it was a training issue, then a data issue, then a vendor issue,
then an integration issue. Shifting explanations are a sign that you’re
generating narratives to protect the decision rather than diagnosing the
real problem.
Signal 3: Nobody can articulate the success criteria
anymore. At the beginning of a well-planned initiative, you can
state clearly what success looks like. As the sunk cost fallacy takes
hold, success gets redefined downward or becomes vague. “We’re learning
a lot” or “The foundation is in place” are phrases that often indicate
the original success criteria have been quietly abandoned.
A Final Thought
The $2.3 million inspection system at that automotive supplier? It
was decommissioned within a month of the Tuesday meeting. The company
replaced it with a simpler, targeted solution that cost $400,000 and
actually worked. The quality director who spoke up was later
promoted.
The money was gone regardless. The question was never about the $2.3
million. The question was about the next $2.3 million — and whether the
organization was going to spend it on a solution or on a story.
In quality management, the most expensive thing you can do is
continue doing the wrong thing. The second most expensive thing is
pretending that the time and money you’ve already spent gives you any
obligation to continue.
It doesn’t. The best time to correct a mistake is now. The second
best time is right now.
Peter Stasko is a Quality Architect with 25+ years
of experience transforming organizations across automotive, aerospace,
and pharmaceutical industries. He specializes in bridging the gap
between behavioral science and quality management, helping leaders
understand why their best processes fail — and how human psychology, not
just engineering, holds the key to building systems that actually
work.
