Every Exception Left a Rule.

Search Info

Primary topic: Organisational complexity created by rules, controls and process additions introduced in response to individual exceptions

Central question: Why did organisations repeatedly respond to exceptional incidents by adding permanent rules, and how did this gradually increase process complexity?

Key concepts: organisational complexity, process design, exception management, organisational learning, bureaucracy, risk management, standardisation

Core insight: Organisations often treated exceptional incidents as evidence of missing rules rather than asking whether the incident justified permanent structural change, causing rare events to accumulate as lasting process requirements.

The Incident That Changed the Process

Looking back from 2049, many organisational rules of the 2020s had surprisingly specific biographies. An invoice had once been approved incorrectly, so another person was added to the approval process. A customer complained that important information had not been provided, so a mandatory information step was introduced. A project failed because one department had been consulted too late, so future projects required an additional coordination meeting. A document was sent with an error, so a second review became standard procedure. None of these responses was inherently unreasonable. Organisations need to learn from mistakes, and recurring failures should alter processes. Risk management depends precisely on the ability to convert experience into more reliable structures rather than repeatedly exposing an organisation to preventable errors.

Yet from the later R2049 perspective, another pattern became visible. Organisations were often much better at converting incidents into new requirements than at determining whether those requirements should remain permanent. The exceptional event disappeared into history while the organisational response became part of normal work. Over time, the process remembered more exceptions than the people operating it did.

Why Adding a Rule Felt Rational

The logic was powerful because incidents created an immediate demand for explanation. When something went wrong, managers needed to understand why it had happened and how recurrence could be prevented. Doing nothing after a visible failure could appear negligent, particularly when the event involved customers, compliance, financial loss or reputational risk. A new rule offered something valuable: observable evidence that the organisation had learned. An additional approval demonstrated control, a checklist showed that the missing step had been formalised, and a new reporting requirement made future oversight possible. The organisation could point to a concrete modification and say that the problem had been addressed.

Removing a rule produced no comparable signal. If an organisation deleted an approval because the underlying risk had disappeared, nothing visible happened when the decision proved correct; the process merely became slightly simpler. If the risk later reappeared, however, the removal could be reconstructed as a mistake. This produced an important asymmetry in organisational learning: adding control was easier to justify than removing it. The later Struction literature called this control asymmetry, the structural tendency for organisations to accumulate safeguards because the consequences of insufficient control are more visible than the consequences of excessive control.

The Difference Between an Error and a Pattern

The deeper problem concerned interpretation. An incident contains information, but it does not automatically reveal what kind of information it contains. A mistake may indicate a systematic weakness, an unusual combination of circumstances, insufficient training, unclear responsibility, poor information transfer or simply ordinary human error. These conditions require different responses, yet organisations under pressure often moved rapidly from event to intervention. Something had happened and therefore something had to change. The visible incident became evidence that the existing process was insufficient, even when the frequency and structural significance of the event had not yet been established.

From 2049, this appeared as a failure to distinguish event information from pattern information. Event information tells an organisation that something occurred; pattern information establishes whether the occurrence reveals a recurring structural condition that justifies redesign. The distinction sounds elementary, but it becomes difficult immediately after a consequential incident. The more uncomfortable the event, the stronger the pressure to demonstrate that it cannot happen again. Organisations therefore sometimes designed permanent structures around events whose defining characteristic was precisely that they were exceptional.

How Exceptions Entered the Standard Process

Once introduced, the new control usually ceased to look exceptional. Employees learned the additional step, software incorporated it, templates reflected it and new colleagues encountered it as part of the established workflow. The historical reason for the rule gradually disappeared while the rule itself became increasingly normal. Someone joining the organisation three years later might discover that every contract required four approvals without knowing that the fourth had been introduced after one unusual incident in 2023. The approval no longer appeared to be an organisational response; it simply appeared to be the process.

This transformation mattered because rules whose origins remain visible can still be reconsidered against those origins. Once the causal history disappears, the existing structure begins to justify itself through existence. People assume that there must be a reason, and usually there was. The more difficult question is whether the reason still exists. Later Struction research called this transition rationale decay: the gradual disappearance of the original justification for a rule while the operational requirement itself remains intact.

Exception Sedimentation

The cumulative result was more consequential than any individual rule. One additional approval might consume only two minutes, one additional field in a form might seem trivial, one additional meeting could be absorbed into the week, and one more person copied into an email might appear almost costless. Over years, however, these additions formed layers. Struction later described this process as exception sedimentation: the accumulation of permanent organisational requirements created by temporary, unusual or historically specific events.

The geological metaphor was deliberate. Sediment rarely arrives as a dramatic mass; it accumulates in thin layers until the underlying structure is covered by material deposited at different moments for different reasons. Organisational processes could develop in much the same way. A workflow that originally contained five necessary steps might eventually contain eleven, each supported by a plausible historical explanation. No single addition had made the process irrational. The complexity existed in the accumulation.

This was why conventional process improvement could struggle to identify it. Each individual requirement could still be defended. The structural question concerned not whether every rule once had a reason, but whether the complete set of historical reasons still described the organisation’s present risk environment.

The Organisation Remembered Through Procedure

Seen from this perspective, organisational bureaucracy acquired a more interesting meaning. Rules and procedures were not merely administrative restrictions; they were a form of institutional memory. A checklist remembered an error that employees might never have witnessed, an approval remembered a risk that had once materialised, a reporting requirement remembered a period when management had lacked information, and a mandatory handover remembered the consequences of a previous communication failure. This memory could be extremely valuable because organisations should not depend on each generation of employees rediscovering the same vulnerabilities through fresh mistakes. Procedures allow experience to survive personnel turnover.

The problem arose when procedural memory lacked forgetting. Human memory is selective partly because retaining everything would be dysfunctional. Organisational memory, by contrast, could become increasingly additive: lessons entered the system, but there was often no equivalent mechanism for determining when a lesson had ceased to require its original procedural form. The organisation therefore became capable of remembering risks after the environment that produced them had changed.

Software might remove the original source of error, responsibilities might be reorganised, regulation might change or a former dependency might disappear. Yet the procedure created around the earlier condition could remain because nothing explicitly connected its continued existence to the persistence of its original cause. The organisation remembered the lesson but gradually lost the reason why that lesson still had to be remembered in precisely this procedural form.

When Safety Accumulated Into Friction

This helps explain a familiar organisational paradox. Employees could encounter a process that everybody considered unnecessarily complicated while every individual step remained defensible. The first approval protected financial control, the second ensured technical accuracy, the third involved compliance, the fourth had been introduced after a previous incident, the documentation requirement preserved traceability, and the meeting ensured coordination. Removing any one element appeared to sacrifice something valuable; keeping all of them produced the complexity everybody complained about.

The problem could therefore not be solved simply by asking which rules were unnecessary, because few rules describe themselves that way. A more useful question was whether several controls were protecting against the same underlying risk or whether the risk had already been reduced elsewhere in the system. This became known in later Struction analysis as control overlap: the persistence of multiple safeguards whose individual rationales remain plausible even though their combined protective value is substantially lower than their combined organisational cost. Complexity was therefore not always the result of pointless bureaucracy. Sometimes it was the accumulated residue of reasonable caution.

The Hidden Cost of Preventing Everything Twice

Organisations were often good at measuring the direct time required by a control. An approval might take only a minute and completing an additional field might require seconds; from this perspective, retaining a precaution appeared inexpensive. Yet the later reconstruction showed that the relevant cost was not simply execution time. Every additional rule introduced another condition that work had to satisfy before it could continue. It created another possible waiting point, another dependency, another source of exceptions and another element employees had to understand.

A two-minute approval could therefore generate much more than two minutes of structural cost if the approver was unavailable, responsibility was unclear or the request returned for clarification. Small procedural requirements multiplied coordination dependencies. This distinction became central to Struction: task cost describes the effort required to perform an activity, whereas structural cost includes the dependencies, handovers, decisions and waiting conditions created because the activity exists within a sequence. Organisations of the 2020s often optimised the first while underestimating the second.

Why Digitalisation Could Preserve the Rules

Digital transformation sometimes intensified this effect. Workflow software made additional controls inexpensive to administer. Fields could become mandatory, approval stages inserted automatically and notifications triggered without manual coordination. Processes that would once have generated enough friction to provoke reconsideration could now operate quietly. This was genuine operational improvement because automation removed repetitive administrative effort and increased consistency, but it also changed the survival conditions of organisational rules.

A cumbersome manual control attracted complaints because people experienced its cost directly. An automated control could remain almost invisible even when its original purpose had weakened. Digitalisation therefore made it possible for organisations to carry more procedural history at lower visible operating cost. The rule survived because technology had made it cheap enough not to question. Complexity did not disappear; it became easier to operate.

The Missing Expiry Date

From 2049, one design omission appeared repeatedly: organisational rules rarely contained an expiry condition. Projects had deadlines, contracts had renewal dates, certifications required periodic review and temporary employees had defined terms. Yet a process step introduced after an incident could enter organisational life without any corresponding date on which its continued necessity had to be reconsidered.

The asymmetry was striking. A rule needed a reason to be introduced, but it often needed no new reason to remain. Later organisations began treating selected controls more like hypotheses than permanent truths. A new requirement could be linked explicitly to the risk it was intended to address, the evidence that justified it and the conditions under which it should be reviewed. This did not mean automatically deleting rules after a fixed period. It meant preserving the connection between the intervention and its rationale. The structural innovation was simple: the organisation stopped assuming that yesterday’s lesson automatically required tomorrow’s procedure.

What Mature Organisational Learning Required

By the 2040s, organisational learning was increasingly understood as a two-directional process. Systems needed to acquire structure when recurring problems revealed genuine weaknesses, but they also needed to release structure when risks changed, controls became redundant or the original problem had been solved elsewhere. Learning without subtraction was no longer considered complete learning.

This changed the interpretation of simplification. Removing a rule was not necessarily deregulation, cost cutting or a retreat from quality. It could represent evidence that an organisation understood its own control architecture well enough to distinguish active safeguards from historical residue. The question therefore shifted from Why do we have this rule? to the more demanding formulation: Which present condition still requires this rule to exist? Historical explanation could justify why a procedure had entered the organisation; it could not by itself establish why the procedure should remain.

The R2049 Reconstruction

From the distance of 2049, the organisations of the 2020s did not appear irrational for accumulating rules. Their behaviour followed a coherent logic: visible failures demanded responses, new controls demonstrated learning and the cost of individual additions often seemed negligible. The misreading lay in assuming that organisational learning consisted primarily of retaining lessons. Every retained lesson occupied structural space by adding conditions, dependencies and expectations to future work. Without mechanisms for reassessment, yesterday’s exceptions could gradually become today’s normal operating environment.

This was how organisations became complicated without anyone deciding that they should become complicated. The process required no bureaucracy for its own sake, only repeated rational responses combined with insufficient structural forgetting. An error occurred, a rule was added and the error did not recur, making the rule appear successful. Years later, nobody could say with certainty whether the absence of the error proved that the rule remained necessary or whether the condition that had once produced the error had disappeared. The rule nevertheless remained because persistence required no new decision. That was exception sedimentation.

From 2049, one of the simplest principles of organisational design therefore appeared surprisingly late in management practice: a system that knows how to learn from exceptions must also know how to stop organising itself around exceptions that no longer define its reality.

They added a rule whenever something went wrong. What they forgot to add was a rule for deciding when the rule could disappear.

Summary

From the perspective of 2049, much of the organisational complexity of the 2020s was not created through deliberate redesign but through repeated responses to exceptions. An error produced an additional approval, a misunderstanding generated another documentation requirement, and an unusual customer case became a new procedural rule. Each intervention could be rational in isolation, but additions were rarely matched by equivalent removal mechanisms. Struction later described this as exception sedimentation: exceptional events disappeared, while the rules created in response to them remained embedded in everyday work.

Transparency

This article was developed within the framework of the concept The Second Thinking Space with the support of generative artificial intelligence. AI is used to explore questions, broaden perspectives, generate alternative formulations, identify patterns and facilitate the critical examination of ideas and assumptions. The article has been substantively reviewed, editorially revised and approved by the author. All editorial decisions, evaluations, interpretations and conclusions are the sole responsibility of the author.