The Paradox
at the Heart of Every Quality Program
In colonial Delhi, the British government grew concerned about the
population of venomous cobras. Their solution was elegantly simple:
offer a cash bounty for every dead cobra. At first, it worked. Thousands
of snakes were killed, and the administrators congratulated themselves
on their clever policy.
Then something unexpected happened. Enterprising citizens began
breeding cobras in their homes, killing them, and collecting the bounty.
When the government discovered the fraud and cancelled the program, the
breeders released their now-worthless snakes into the streets. Delhi
ended up with more cobras than before the intervention.
This is the Cobra Effect — a perverse incentive that produces the
opposite of its intended result. And it is alive and well in
manufacturing quality departments around the world, quietly dismantling
the very systems organizations build to improve.
How
Perverse Incentives Destroy Quality From the Inside
Every manufacturing organization has incentive systems. Some are
formal — bonuses tied to scrap rates, performance reviews linked to OEE
targets, promotions based on audit scores. Others are informal — the
shift lead who looks the other way when operators skip steps to hit
production numbers, the plant manager who praises speed over accuracy,
the unwritten rule that you never shut down a line for a quality issue
unless you want a reputation as “difficult.”
The formal incentives are the ones everyone sees. The informal ones
are the ones that actually drive behavior.
Consider the plant that ties operator bonuses to units produced per
shift. The message is clear: quantity matters more than quality.
Operators learn, quickly and without being told, that catching a defect
means stopping the line, which means losing their bonus. The “quality
culture” posters on the break room wall become irrelevant the moment the
paycheck tells a different story.
Or the quality department that is measured by the number of
corrective actions closed per quarter. Close actions fast, get rewarded.
Never mind whether the root causes were actually addressed. Never mind
whether the corrective actions were real solutions or just paperwork
exercises designed to check a box. The metric says “closed,” so the
metric says “success.” The defects, meanwhile, keep coming back.
The Most
Common Cobra Effects in Manufacturing Quality
The Scrap Rate Target
A plant sets a target to reduce scrap to below two percent. Managers
bonuses are tied to hitting this number. What happens? The definition of
“scrap” begins to change. Rework gets reclassified as “reprocessing.”
Material that should be scrapped gets downgraded and sold to a secondary
market at a loss — a loss that appears in a different budget line,
invisible to the scrap metric. Operators learn to hide defective parts
in rework bins rather than scrapping them.
The scrap rate hits 1.8 percent. The manager gets a bonus. The actual
waste in the plant has increased. Nobody notices because the metric says
improvement.
The Audit Score Obsession
An automotive supplier prepares for its annual IATF 16949
surveillance audit. The quality team spends three months updating
procedures, training records, and process flow diagrams. Mock audits are
conducted. Nonconformances from the previous audit are verified as
closed. The auditor arrives, finds everything in order, and issues a
clean report.
Between audits, nothing has changed. The procedures that were updated
for the audit sit in binders collecting dust while the actual work on
the floor follows a completely different set of unwritten rules. The
audit score says the quality system is excellent. The customer returns
say otherwise.
The Zero Defect Month
A plant manager announces a “Zero Defect Month” competition. The
shift with the fewest recorded defects wins a pizza party and
recognition at the all-hands meeting. The unintended consequence is
immediate and predictable: defects stop being recorded. Operators
negotiate with inspectors. Inspectors negotiate with their supervisors.
The quality system becomes a mechanism for hiding problems rather than
exposing them.
The plant records zero defects for the month. The manager celebrates.
Three months later, a customer finds a defect that should have been
caught during that “perfect” month. The warranty claim costs more than a
year of pizza parties.
The Efficiency Trap
OEE is the holy metric of lean manufacturing. Plants obsess over it.
But OEE measures availability, performance, and quality — and most
plants weight availability and performance far more heavily than
quality. An operator who keeps the line running at full speed while
letting marginal parts pass is rewarded for high OEE. An operator who
stops the line to investigate a quality concern is punished with lower
OEE.
The message is not subtle. The organization says it values quality.
Its measurement system says it values throughput. Operators are not
stupid. They follow the measurement system.
The Cost of Quality Budget
Some organizations treat the cost of quality as a budget line that
needs to be reduced. Prevention costs — training, process improvement,
measurement system analysis — get cut because they look like overhead.
Appraisal costs — inspection, testing, audits — get cut because they
look like non-value-added activities. Failure costs — scrap, rework,
warranty claims — remain because they are the consequence of the cuts to
prevention and appraisal.
The quality budget goes down. The quality manager gets praised for
“efficiency.” The actual cost of poor quality goes up, but it is
distributed across warranty, customer service, lost business, and
reputation — budget lines that nobody connects back to the quality
cuts.
Why Smart
Organizations Create Dumb Incentives
The Cobra Effect is not the result of stupid people making stupid
decisions. It is the result of intelligent people designing systems
without understanding the second-order consequences of those
systems.
When a vice president of operations sets a scrap reduction target,
they are not trying to encourage people to hide defects. They genuinely
want less scrap. But they have designed a system that rewards the
appearance of less scrap rather than the reality of less scrap. The
difference between appearance and reality is where the cobras breed.
This happens because most incentive systems are designed around what
is easy to measure rather than what is meaningful to measure. Scrap rate
is easy to measure. Actual waste — including hidden waste, rework,
downgraded material, and customer returns — is hard to measure. OEE is
easy to calculate. The true cost of defects that escape to the customer
is difficult to calculate, often invisible until a catastrophic failure
occurs.
The gap between what is easy to measure and what matters is where
every perverse incentive lives.
The Psychology of Gaming
the System
It is tempting to blame the people who game the system. The operator
who hides defects, the inspector who passes marginal parts, the manager
who reclassifies scrap — they are all making rational decisions within
the incentive structure they have been given.
Consider the operator who discovers a defect mid-shift. Reporting it
means stopping the line. Stopping the line means the shift will not hit
its production target. Missing the target means the team loses its
bonus. The operator’s rent is due next week. The operator quietly puts
the defective part in the rework bin and keeps running.
Is this operator dishonest? Or is this operator responding rationally
to an incentive system that punishes honesty? When the cost of doing the
right thing falls on the individual but the benefit accrues to the
organization, most people will protect themselves. This is not a
character flaw. It is a system design flaw.
The same dynamic plays out at every level. The quality engineer who
knows a corrective action is superficial but closes it anyway because
the metric demands closure. The plant manager who knows the real scrap
rate is higher than reported but defends the number because corporate
will punish honesty with a resource cut. The supply chain director who
knows a supplier’s quality is deteriorating but continues buying from
them because the cost savings metric looks good this quarter.
At every level, the system rewards compliance with the metric and
punishes alignment with the truth.
How
to Design Incentive Systems That Actually Improve Quality
Measure Outcomes, Not
Activities
Stop counting corrective actions closed. Start measuring whether the
problems actually stopped recurring. Stop tracking audit scores. Start
tracking customer returns, warranty claims, and field failures — the
outcomes that audits are supposed to prevent.
This is harder. It requires patience, because outcomes take time to
manifest. It requires honesty, because outcome metrics are often uglier
than activity metrics. But it eliminates the Cobra Effect at its source:
if you measure what you actually want, people will optimize for what you
actually want.
Separate Reporting From
Consequences
One of the most effective structural changes an organization can make
is to decouple the reporting of quality problems from the punishment of
the people who report them. This is the principle behind no-blame
cultures, but it needs to go deeper than culture — it needs to be
structural.
If a shift reports a spike in defects and the consequence is that the
shift loses its bonus, defects will not be reported. If a shift reports
a spike in defects and the consequence is that additional resources are
deployed to investigate the root cause, defects will be reported. The
behavior follows the consequence, not the poster on the wall.
Reward Problem
Discovery, Not Problem Absence
Most quality incentive systems reward the absence of problems. Zero
defect months, scrap reduction targets, clean audit scores — all reward
the appearance of having no problems. But every manufacturing process
generates problems. The healthy organization is not the one with no
problems; it is the one that finds and fixes them fastest.
Flip the incentive. Reward the team that discovers and permanently
eliminates a root cause. Reward the operator who stops the line to
prevent a defect from escaping. Reward the inspector who rejects a
marginal part. Make problem discovery a source of status and reward, not
a source of blame and punishment.
Use Balanced Metrics,
Not Single Metrics
No single metric captures quality performance. Scrap rate, OEE, audit
scores, customer returns, warranty costs, first-pass yield — each
illuminates one facet of a complex system. When an organization
optimizes for a single metric, it creates the conditions for the Cobra
Effect.
Use a balanced scorecard that weights multiple quality indicators.
Include leading indicators (process capability, measurement system
performance, training completion) and lagging indicators (customer
returns, warranty claims, cost of poor quality). When one metric is
gamed, the others reveal the truth.
Include the
People Who Do the Work in System Design
The people who will game an incentive system are the same people who
can tell you how it will be gamed before it is even implemented. Ask
operators how they would achieve a scrap reduction target. Ask quality
engineers what would happen if corrective action closure were tied to
bonuses. Ask inspectors what would change if defect counts became a
performance metric for their department.
The answers will be uncomfortable. They will reveal every loophole,
every workaround, every perverse incentive before it has a chance to
cause damage. Including frontline workers in system design does not
guarantee perfect incentives, but it prevents the most obvious and
damaging mistakes.
The
Deeper Lesson: Trust the System, But Verify the System
The Cobra Effect teaches us that systems shape behavior more
powerfully than exhortations, training, or culture initiatives. If your
quality system rewards the wrong behavior, no amount of quality culture
training will override it. People optimize for what is measured and
rewarded, not for what is posted on the break room wall.
The solution is not to abandon metrics and incentives. It is to
design them carefully, monitor their effects honestly, and adjust them
when they produce unintended consequences. This requires humility — the
willingness to admit that a well-intentioned system might be making
things worse. It requires vigilance — the discipline to look at
outcomes, not just at the metrics that were supposed to predict
outcomes.
Most of all, it requires understanding that the cobras are always
there. Every incentive system, no matter how well-designed, will produce
some unintended behavior. The question is not whether perverse
incentives exist in your organization. They do. The question is whether
you are looking for them — or whether you are so busy celebrating your
metrics that you have not noticed the snakes breeding in the
basement.
Peter Stasko is a Quality Architect with over 25
years of experience in manufacturing quality management, process
improvement, and quality system design across automotive, aerospace, and
industrial manufacturing sectors. He specializes in helping
organizations align their measurement systems with their actual quality
goals — and in finding the cobras before they find their customers.
