Exception Inflation
Definition
Temporary exceptions created to handle edge cases gradually expand until they become the dominant operating mode.
What begins as a narrowly scoped allowance ("just this once," "only for this case") accumulates over time. Each exception adds conditional logic, special handling, or informal override paths. The system becomes harder to reason about, harder to automate, and increasingly dependent on human judgment to navigate its own rules.
Eventually, the exception becomes the system.
What Fails
Rule clarity, predictability, and system coherence.
Why It Emerges
- Systems are optimized for speed, not integrity
- Redesign is costlier than approval in the short term
- Authority is centralized while responsibility remains distributed
- Metrics reward throughput, not rule quality
How It Hides
- Exceptions are framed as customer advocacy
- Each exception seems reasonable in isolation
- Success depends on "knowing who to ask"
- Failures are absorbed by experienced operators
What It Gets Mistaken For
- Empowerment
- Customer obsession
- Mature judgment
- Real-world pragmatism
Fatal misread: Flexibility is mistaken for resilience.
Early Warning Signals
- "This is an exception, but it happens a lot"
- Approval queues grow faster than throughput
- Different teams apply the same rule differently
- Automation fails disproportionately on "special cases"
- Leaders rely on judgment calls instead of rules
- Exception logic outnumbers core rules
If these appear, the system is already fragmenting.
Common Misdiagnoses
- "We just need a little more flexibility"
- "Policy can't cover every situation"
- "This is a customer-first exception"
- "It's faster to approve than redesign"
- "Edge cases are unavoidable at scale"
What Actually Interrupts It
- Hard limits on exception volume
- Mandatory root-cause conversion of recurring exceptions
- Sunset clauses with enforcement
- Redesigning the rule instead of approving around it
- Making exception cost visible to leadership
Recovery Condition
Exception volume trends downward and policy is updated or enforced.
Fail-Safe Default
Default to denial. Exceptions require explicit cost and sunset date.
Cascade Relationships
Upstream: FM-01 Responsibility Compression, FM-03 Responsibility Without Authority
Downstream: FM-02 Escalation Inversion, FM-05 Normalized Workarounds