Quality
and the Normalization of Deviance: When Your Organization’s Accepted
Exceptions Become Its Unwritten Standards — and the Violations Everyone
Tolerated Became the Catastrophe Nobody Predicted
The Disaster That Wasn’t an
Accident
On January 28, 1986, the Space Shuttle Challenger disintegrated 73
seconds after launch, killing all seven crew members. The technical
cause was clear: an O-ring seal failed in the right solid rocket
booster. But the organizational cause was far more
insidious.
The O-rings had been showing signs of erosion on earlier flights.
Engineers noticed it. Management noticed it. NASA noticed it. And over
successive missions, what began as an alarming anomaly gradually became
an accepted condition. The reasoning went like this: We saw erosion
on the last flight, and nothing catastrophic happened. So erosion must
be within acceptable limits. Each successful flight that included
O-ring erosion reinforced the belief that erosion was tolerable. The
boundary of “acceptable” moved, flight by flight, until it had moved so
far that seven people died.
Diane Vaughan, the sociologist who studied the Challenger disaster
extensively, gave this phenomenon a name: the normalization of
deviance. She defined it as the process by which individuals or
groups become so accustomed to deviant behavior — behavior that violates
their own standards — that they no longer consider it deviant at
all.
The Challenger was not a quality failure caused by ignorance. It was
a quality failure caused by familiarity. And that makes it the
most dangerous kind.
Why This
Is a Quality Problem, Not Just a Space Problem
You might be thinking: That’s NASA. We don’t launch space
shuttles. Our organization follows ISO 9001. We have
procedures.
Let me ask you a few questions.
Have you ever signed off on a deviation that you would have flagged
as unacceptable two years ago? Have you ever watched an operator skip a
step in a work instruction and thought, It’s probably fine, they’ve
done it a thousand times? Have you ever approved a supplier
shipment that failed one minor specification because we’ve never had
a problem with them before? Have you ever let a calibration slide a
few weeks past its due date because the instrument has been stable
for years?
If you’ve been in quality long enough, you’ve seen it. You may have
even done it. And each time, the reasoning was perfectly rational in the
moment. The deviation was small. Nothing bad happened last time.
Production needs to keep moving. The customer won’t notice.
This is normalization of deviance. And it is happening in your
organization right now.
The Mechanism: How
Excellence Erodes
Normalization of deviance follows a predictable, almost mechanical
pattern. Understanding it is the first step to interrupting it.
Stage 1: The First Deviation
Something unusual happens. A process parameter drifts outside its
control limit. A skipped inspection catches nothing. A supplier ships
material that’s slightly out of spec. The deviation is noticed,
discussed, and — here’s the critical moment — accepted, usually
with good reason. There’s a deadline. The risk seems low. The business
case is compelling.
The first time, there’s usually discomfort. People talk about it.
Someone raises a concern. But the deviation produces no negative
consequence, and the concern fades.
Stage 2: Accumulation
The same deviation occurs again. And again. Each occurrence without
consequence makes the next one easier to accept. What was once an
exception becomes a pattern. What was once discussed becomes routine.
What was once flagged becomes invisible.
This is where the danger compounds. The organization isn’t just
accepting the deviation — it’s building institutional memory
that the deviation is safe. New employees learn from watching veterans.
Veterans learn from watching each other. Nobody formally changes the
procedure, but everyone understands the real procedure.
Stage 3: Redefinition
At some point, the accumulated deviations become the de facto
standard. The original standard still exists on paper, but the
operational standard — what people actually do — has shifted. The gap
between documented procedure and actual practice grows wider, and nobody
acknowledges it because acknowledging it would mean admitting that the
organization has been operating outside its own rules.
This is the stage where auditors find “minor nonconformities” that
are actually symptoms of a much deeper cultural drift. This is where
quality professionals write corrective actions that address the symptom
while the root cause — the gradual acceptance of lower standards —
remains untouched.
Stage 4: Catastrophe
Eventually, the accumulated deviations create conditions for a
significant failure. A product recall. A customer complaint that
escalates. An audit finding that shuts down a line. An actual safety
incident.
And the post-mortem always sounds the same: How did we not see
this coming? The answer is: you did see it coming. You saw it every
day. You just stopped recognizing it as a problem.
Where It Hides in Your
Organization
Normalization of deviance doesn’t announce itself. It hides in the
spaces between your procedures and your practices. Here are the places
it lives:
Calibration and Measurement
Your calibration schedule says every 12 months. Equipment has been
stable for years. You push it to 13 months, then 14. The measurement
data looks fine — until it doesn’t. By the time you discover the drift,
you’ve shipped months of product measured with instruments that were
outside their calibration window, and you have no idea which
measurements were affected.
Incoming Inspection
Your sampling plan says inspect 10% of incoming lots. The supplier
has a good track record. You drop to 5%, then to “visual only,” then to
“we trust them.” One day, a lot arrives with a systematic defect that
your reduced inspection doesn’t catch, and it reaches your production
line.
Process Controls
Your control plan calls for checking a critical parameter every hour.
Production is running smoothly. The operator checks every two hours,
then every four. The parameter drifts during the overnight shift, and by
morning, you’ve produced eight hours of nonconforming material.
Document Control
Your procedures specify that changes require review and approval. But
the “quick change” — the one everyone agrees is minor — gets implemented
without going through the formal process. Over time, the documented
procedure diverges so far from actual practice that an auditor can’t
reconcile the two.
Training and Competence
Your training matrix requires specific qualifications for certain
tasks. But when someone’s out sick and production needs to keep running,
an unqualified person steps in. “Just this once” becomes “they’ve done
it before.” Competence requirements become suggestions.
The Role of Metrics in
Enabling Deviance
One of the most dangerous enablers of normalized deviance is the
metrics system itself. When organizations measure the wrong things — or
measure the right things in the wrong way — they create incentives that
accelerate the normalization process.
Consider the organization that tracks “number of deviations” as a
quality metric. When deviations are penalized, people stop documenting
them. The metric improves while actual performance degrades. The
organization congratulates itself on reducing deviations while the real
deviations continue unabated, now invisible.
Or the organization that rewards on-time delivery above all else.
When every missed shipment triggers a management review but every
shipped defect triggers only a corrective action (if it’s caught at
all), the message is clear: Ship on time. Quality is secondary.
The deviation from quality standards isn’t just tolerated — it’s
incentivized.
The most insidious metric-driven normalization happens when
organizations set targets that are impossible to achieve without cutting
corners. When the target is 100% on-time delivery but the process is
only capable of 95%, people learn to game the system. They ship early,
they reclassify orders, they negotiate with customers. And each gaming
behavior becomes normal, expected, even rewarded.
How to Detect It Before
It’s Too Late
Detecting normalization of deviance requires looking at your
organization with fresh eyes — or, more practically, finding someone who
has them.
Gemba Walks With New Eyes
Go to the production floor with a simple question: What are we
doing differently from what our procedures say? Not as an auditor
looking for nonconformities, but as a curious observer trying to
understand the gap between the documented process and the actual
process. If operators immediately volunteer the differences, you have
transparency. If they hesitate or look nervous, the deviations have
already been normalized.
New Employee Observations
Your newest employees are your best auditors for normalized deviance.
They haven’t been conditioned to accept the current practices. Listen
when they ask questions like Why do we do it this way when the
procedure says to do it differently? That question is worth more
than any external audit.
If new employees stop asking those questions within their first
month, you’ve successfully normalized your deviance into them. That’s
not onboarding success — that’s cultural failure.
Trend Analysis on
Near-Misses
Near-misses are the early warning system for normalized deviance.
When near-miss reports suddenly drop without a corresponding improvement
in process capability, it doesn’t mean things are getting better. It
means people have stopped reporting — because they’ve accepted the
conditions that produce near-misses as normal.
Track near-miss reporting rates alongside actual incident rates. If
reporting drops while incidents hold steady or increase, you have a
normalization problem.
The “Would
We Have Accepted This Five Years Ago?” Test
Take a sample of current process parameters, incoming material
acceptances, and deviation closures. Compare them to the same data from
five years ago. If today’s accepted values would have been rejected five
years ago — and the process hasn’t been formally changed — you’ve
experienced normalization of deviance.
What to Do About It
Interrupting normalization of deviance requires both systemic and
cultural interventions. Neither alone is sufficient.
Rebuild the Standards
Go back to your original standards. Not the ones that have been
“adjusted” over the years — the ones that were set based on actual
engineering analysis, customer requirements, and risk assessment.
Compare them to your current operating standards. The gap between them
is your accumulated normalization.
This doesn’t mean reverting to the original standards in every case.
Some adjustments may be justified by improved process understanding,
better materials, or changed customer requirements. But each adjustment
should be conscious, documented, and justified — not the result
of gradual, invisible drift.
Separate
Deviation Acceptance From Production Pressure
One of the primary drivers of normalization is the conflict between
quality standards and production pressure. When the same person who
approves a deviation is also accountable for meeting production targets,
the decision is compromised by design.
Create a clear separation. The person who accepts or rejects a
deviation should have no production delivery targets in their
performance metrics. This isn’t bureaucratic overhead — it’s structural
integrity.
Conduct “Deviance Audits”
Traditional audits check compliance with documented procedures.
Deviance audits check the gap between documented procedures and actual
practice — not to assign blame, but to understand where and why the
drift has occurred.
Frame the audit as a learning exercise, not a compliance exercise. If
people feel threatened, they’ll hide the deviations more carefully. If
they feel heard, they’ll tell you exactly where the standards have
drifted and why.
Establish a “Fresh Eyes”
Program
Rotate people through roles and departments regularly. Bring in
external perspectives — not just consultants, but people from other
industries, other facilities, even other companies through benchmarking
visits. Fresh eyes see what familiar eyes have stopped seeing.
Create Consequence-Free
Reporting
The single most effective countermeasure to normalization of deviance
is a culture where people can report deviations without fear of
punishment. This doesn’t mean accepting deviations — it means separating
the act of reporting from the act of disciplining. Report freely.
Investigate thoroughly. Address systemically. And if the investigation
reveals that the deviation was caused by an impossible production target
or an inadequate process, fix the system, not the person.
The Paradox of Experience
There’s a bitter irony at the heart of normalization of deviance: the
more experienced your people are, the more vulnerable your organization
is.
Experienced operators have seen deviations produce no consequences —
not because the deviations were safe, but because they were lucky. Over
time, they accumulate a personal database of “things that went wrong but
didn’t cause problems,” and they use that database to make risk
judgments. Their confidence is genuine. Their judgment is
compromised.
This is why the Challenger disaster wasn’t caused by inexperienced
engineers. It was caused by experienced engineers who had seen O-ring
erosion so many times that it no longer alarmed them. Their experience
didn’t protect them from the normalization — their experience
was the normalization.
The antidote isn’t to replace experienced people. It’s to ensure that
experience is paired with continuous challenge — with people who ask
why are we accepting this? and have the institutional support
to keep asking until they get a real answer.
The Cost of Vigilance
I won’t pretend that fighting normalization of deviance is easy or
free. It requires constant attention. It means sometimes stopping
production when “everyone knows it’ll be fine.” It means rejecting
material when “we’ve never had a problem with this supplier.” It means
enforcing standards when the business case for relaxing them is
compelling and immediate.
But the cost of vigilance is infinitesimal compared to the cost of
catastrophe. The Challenger O-rings were a component that cost a few
dollars. The failure cost seven lives, billions of dollars, and the
credibility of an entire space program.
In your organization, the cost of maintaining your standards — of
refusing to accept the incremental, invisible erosion of excellence — is
measured in extra inspections, delayed shipments, and uncomfortable
conversations. The cost of not maintaining them is measured in recalls,
lawsuits, lost customers, and the slow, irreversible destruction of your
quality culture.
The Standard You Keep
Every organization has two sets of standards: the ones it writes down
and the ones it actually follows. The gap between them is where
normalization of deviance lives. Closing that gap isn’t a one-time
project — it’s a permanent discipline.
The organizations that survive and thrive aren’t the ones that never
experience deviations. They’re the ones that refuse to normalize them.
They treat every deviation as new information, not as confirmation that
the deviation is safe. They maintain the discomfort that the first
deviation produced, even after the hundredth occurrence.
Your quality system is not defined by what’s in your procedures. It’s
defined by what your people do when nobody is watching. And if what they
do differs from what your procedures say — if they’ve learned to accept
what you once found unacceptable — then your quality system has already
been compromised.
The question isn’t whether normalization of deviance is happening in
your organization. It is. The question is whether you’ll have the
courage to find it before it finds you.
Peter Stasko is a Quality Architect with 25+ years
of experience transforming organizations across automotive, aerospace,
and pharmaceutical industries. He specializes in building quality
cultures that don’t just meet standards — they refuse to compromise
them, even when compromise seems like the rational choice.
