pinkaku 組織病理学研究所

現場から生まれた「社腸」という組織論で、会社の詰まりを言語化する

タグ: performance metrics

Case 49: When Feedback Loops Reinforce the Wrong Signals
Concept Inversion
Feedback loops are assumed to correct errors.
They do not.
Feedback loops amplify whatever signal they are connected to.

Structural Decomposition
Systems rely on feedback loops to adjust behavior.
Outputs are measured.
Signals are generated.
Adjustments are made.
When signals are accurate, correction occurs.
However, when signals are distorted, feedback loops do not detect the error.
They reinforce it.
The system optimizes toward the signal.
The signal drifts further from reality.
Each cycle increases alignment with the wrong target.
Correction becomes amplification.

Pathology Progression
Signals are established.
Feedback loops are connected.
Initial adjustments improve performance.
Signal distortion begins.
Feedback reinforces the distortion.
Performance appears stable or improving.
Reality deteriorates.
Discrepancies grow.
The system becomes locked into misalignment.

Cold Diagnosis
An organization that cannot verify the validity of its signals will reinforce its own errors.
It becomes self-reinforcing in the wrong direction.

Structural Definition
This case defines a condition where feedback loops amplify incorrect or distorted signals instead of correcting them.

One-Line Summary
This case describes how feedback loops can reinforce error when connected to invalid signals.

Explore the full case index
This article is part of the Organizational Pathology case archive.
All published cases can be found here:
Organizational Pathology — Case Index
2026-07-25
Case 46: When Systems Adapt to Evaluation Instead of Reality
Concept Inversion
Adaptation is assumed to improve performance.
It does not.
Adaptation to evaluation criteria can detach a system from reality.

Structural Decomposition
Systems are exposed to external evaluation.
They begin to adapt.
Behavior changes.
Output changes.
Priorities shift.
However, adaptation is directed toward evaluation metrics, not actual conditions.
Reality remains unchanged.
The system becomes optimized for being judged, not for functioning.

Pathology Progression
Evaluation pressure increases.
The system adapts.
Performance indicators improve.
Real conditions stagnate.
Mismatch grows.
The system becomes dependent on evaluation signals.
Reality is no longer the reference point.

Cold Diagnosis
An organization that adapts to evaluation systems instead of real conditions loses functional integrity.
It performs well in metrics but poorly in reality.

Structural Definition
This case defines a condition where systems adapt to external evaluation criteria rather than actual operational reality.

One-Line Summary
This case describes how adaptation to evaluation systems detaches performance from reality.

Explore the full case index
This article is part of the Organizational Pathology case archive.
All published cases can be found here:
Organizational Pathology — Case Index
2026-07-12
Case 44: When Systems Optimize for Incompatible Metrics
Concept Inversion
Optimization is assumed to improve outcomes.
It does not.
Optimization improves alignment with a metric, not necessarily with value.

Structural Decomposition
Multiple systems evaluate the same output using different metrics.
Search systems prioritize relevance and structure.
Social systems amplify engagement signals.
Monetization systems enforce compliance and advertiser safety.
Each system optimizes for its own objective.
Metrics are not shared.
Objectives are not aligned.
Trade-offs are not resolved.
Improvement in one metric can degrade performance in another.
Optimization becomes fragmentation.

Pathology Progression
A system selects a primary metric.
Optimization begins.
Performance improves in that metric.
Other systems react negatively.
Visibility changes.
Engagement shifts.
Monetization declines.
Further optimization is applied.
Conflicts intensify.
The system loses coherence.

Cold Diagnosis
An organization that optimizes for multiple incompatible metrics without hierarchy or integration cannot stabilize its performance.
It fragments its own structure.

Structural Definition
This case defines a condition where multiple systems optimize for incompatible metrics, resulting in structural fragmentation.

One-Line Summary
This case describes how optimization across conflicting metrics fragments system performance instead of improving it.

Explore the full case index
This article is part of the Organizational Pathology case archive.
All published cases can be found here:
Organizational Pathology — Case Index
2026-07-05