The Complete Diagnostic Taxonomy
Recurring failure is not random. It belongs to a class. The class names the correction. The sub-type determines the correction sequence.
This is the classification standard for structural failure in organisations, platforms, and institutions. If your system keeps producing the same failure despite correction effort — this is the reference that names what kind of failure it is, what qualifies as that failure, and what the correction path looks like.
Nine failure classes. Fifty-four sub-types. Formal inclusion and exclusion criteria, false positive tests, refusal conditions, and corrective directions for each. Built from eleven papers of structural diagnostic research.
Start with the decision tree below — nine questions that route you to a failure class. Then open the relevant class section and check your failure against the inclusion, exclusion, and refusal criteria. A failure only belongs to a class when it survives all four tests.
Every recurring structural failure belongs to one of these nine classes. Click any class to jump directly to its criteria, sub-types, and corrective direction.
Start at Q1. Follow the path. Most real failures trigger multiple branches — the branch with the strongest structural signal is the primary failure class.
Most real failures are compounds. A compound failure contains a primary governing break, secondary consequences, and tertiary adaptations. Do not diagnose every symptom as a separate primary failure. The diagnostic question is always: which break, if corrected, would change the behaviour of the others? That is the primary class. The others are secondary.
The primary class governs the correction sequence. Fix the governing break first, then address secondary failures in dependency order. When two classes are triggered simultaneously, ask which one is upstream. The upstream class is primary.
Misclassification rules — do not classify by:
A company is forced to operate through a system it does not control — a host platform, regulator, marketplace, payment authority, or infrastructure provider — while still carrying the user-facing consequence of any failure or mismatch that system produces.
Core diagnostic question: What external system governs the condition of success while this organisation carries the consequence of failure?
Boundary with Class 3: Remove the external constraint mentally. If the company's model would work without it, the failure is Class 1. If the model is structurally broken regardless, the failure is Class 3.
An external system governs a condition of operating success; the company cannot change the rules of that external system; and the company carries the user-facing consequence when that system fails or misaligns.
The company voluntarily chose the dependency and can exit it without structural consequence; the external system's rules are compatible with the company's promise; or the failure is in the company's own response to the constraint rather than in the constraint itself.
The product must pass through another company's rules, payment grammar, or distribution logic to reach its users, and those rules are incompatible with the product's core promise.
The company owns the customer relationship but not the payment authority. A third party controls refunds, payouts, entitlement logic, or subscription state.
Discovery, ranking, pricing, or access is governed externally with no structural alternative the company controls.
External regulation reshapes the operating model faster than the architecture can follow, creating compliance regimes that grow without resolving the underlying structural mismatch.
The product's reliability has a ceiling set by a provider the company does not control. When the upstream provider degrades, the product degrades with it.
Users now structurally expect something the architecture was never built to provide — continuity, ownership, personalisation — and every product improvement falls short because the underlying architecture cannot satisfy the expectation.
The organisation was designed for a world that no longer exists. Reform programmes stall at the same architectural boundary each cycle because the founding structure cannot be redesigned from inside it.
The failure is not in the people or the execution — it is in the mismatch between the design's load assumptions and the current load.
Core diagnostic question: What scale condition did the original design never have to survive?
Boundary with Class 5: Ask whether the authority problem would exist at the original scale. If yes, the failure is Class 5 — it was always a governance problem, scale merely made it visible. If no, the failure is Class 2.
The system functioned adequately at a prior scale; the failure emerged or worsened significantly as the system scaled; and the design was not updated to match the new operating conditions.
The authority to govern the scaled system is missing (Class 5); humans absorb scale load because the system was designed around impossible human behaviour (Class 7); or the failure exists at original scale as well as current scale.
The same process runs on more inputs than it was designed for. Quality, accuracy, or reliability degrades proportionally to volume.
Edge cases grow faster than the rules designed to handle them. At founding scale the common cases dominate. At current scale the edge cases multiply until they become the majority.
Informal coordination that worked at small scale has no structural replacement at large scale. Nobody knows who owns the decision when twelve teams need to make it simultaneously.
The platform was built on human reputation and relationships. At scale, those mechanisms stop working — the signals are too thin, the relationships too distant — and fraud, abuse, and quality collapse fill the gap.
The system was never designed for the maintenance load it now carries. Technical debt, documentation decay, and dependency accumulation create a structural maintenance burden that exceeds the organisation's capacity to manage it.
The founding premise — about what users want, what technology can do, what the market looks like, or what the competitive environment allows — was wrong or has become wrong. The organisation continues to operate as if the premise holds.
Core diagnostic question: Which founding assumption is the organisation still executing against that has been materially disproved?
A core premise on which the operating model was built has been materially disproved; the organisation continues to execute against the original model; and the recurring failure traces back to that original mismatch.
The failure is caused by an external constraint the organisation cannot remove (Class 1); or the failure is in execution of a sound model rather than in the model itself.
The organisation was built for a market that does not exist at the assumed size, shape, or willingness to pay.
The product's core value depended on a technology capability that does not exist, cannot be achieved at the assumed cost, or has been commoditised.
The unit economics the model was built on cannot be achieved at the operating scale required for sustainability.
The competitive position the model assumed — first-mover advantage, network effects, structural moat — has been eroded or never materialised.
Two or more systems are expected to interoperate, but their boundaries — around data, authority, timing, vocabulary, or ownership — were designed independently and are now structurally incompatible.
Core diagnostic question: Where do two systems meet, and what does each system assume about the other that the other cannot satisfy?
A failure lives at the interface between two systems; both systems function adequately in isolation; and the failure recurs at the boundary regardless of changes made within either system.
The failure is within a single system rather than at an interface; or the boundary is shared with an external party the organisation cannot negotiate with (Class 1).
Two systems expect to share data that is defined, formatted, or updated differently in each. The mismatch is structural, not a communication failure.
Two systems each believe they have authority over the same decision. The conflict is structural: both systems were designed to own the same thing.
Two systems operate at different time scales and the mismatch creates structural failure at the handoff point.
Two systems use the same word to mean structurally different things, creating failures that appear to be communication problems but are actually definitional conflicts.
Nobody owns the space between two systems. The interface is everyone's problem and therefore nobody's problem.
A failure in one system triggers failures in dependent systems in a sequence that amplifies faster than any single system can absorb or correct.
The failure is not in the technology or the people — it is in the absence of a structural decision-making, enforcement, memory, or correction authority for a specific class of decisions.
Core diagnostic question: What class of decision has no structural owner — no person or system with the authority, information, and accountability to make, enforce, remember, or correct it?
A specific class of decision recurs; the decision has no structural owner; and the failure repeats because the absence of authority is not remedied by adding more people, process, or technology within the current structure.
Authority exists but is poorly exercised — that is an execution failure. Exclude if the governance gap is caused by scaling (Class 2) rather than by an inherent absence of authority design.
No person or system has the structural authority to make a specific class of decision. Decisions are deferred, escalated, or made inconsistently because ownership is absent.
Decisions are made but there is no structural authority to enforce them. Rules exist on paper; nobody is structurally empowered to require compliance.
Decisions are made but not recorded in a form that persists across personnel changes. The same decision is remade repeatedly because the prior decision left no structural trace.
When decisions produce wrong outcomes, there is no structural authority to recognise the error and correct it. Errors persist because the correction path does not exist.
The system has no structural mechanism for converting outcomes into decision inputs. Decisions are made without access to what previous decisions produced.
The incentive structure produces outcomes that diverge from organisational goals. Nobody is acting irrationally or maliciously. The structure itself is the problem.
Core diagnostic question: What rational behaviour does the current incentive structure reward that produces the outcome the organisation does not want?
The failure produces outcomes the organisation does not want; the people producing those outcomes are following rational incentives; and changing the people without changing the incentive structure would not change the outcome.
The failure is caused by deliberate bad behaviour rather than rational incentive-following; or the incentive structure is sound and the failure is in execution or measurement.
The metric being optimised is not the outcome the organisation actually needs. People maximise the measured thing at the expense of the unmeasured thing that matters.
The reward structure compensates behaviour that produces the wrong outcome. People are rewarded for activity that looks right but produces structural harm.
Decision-makers are rewarded on short-term outcomes; the structural costs of those decisions are paid on long-term timescales they will not personally experience.
The agent is structurally positioned to benefit from outcomes that harm the principal. The relationship is structurally conflicted, not a matter of individual integrity.
The failure is not in the humans — it is in the system's design assumption that humans will reliably do something they structurally cannot sustain. The system places cognitive, procedural, or interpretive demands that exceed what humans can reliably carry under real operating conditions.
Core diagnostic question: What does this system require humans to reliably do that humans cannot reliably do under actual conditions?
The failure traces to a structural assumption about human behaviour; that assumption is violated under real operating conditions of load, ambiguity, or complexity; and the failure would disappear if the system stopped making that assumption.
The failure is caused by individual bad behaviour or deliberate violation. Exclude if the human load has simply grown beyond design capacity — that is Class 2 (Volume Scaling Failure).
The system requires humans to hold too much in working memory simultaneously. Errors are produced not by incompetence but by structural overload.
The system produces ambiguous outputs or instructions and relies on humans to resolve the ambiguity consistently. The resolution is inconsistent because the ambiguity is structural.
The documented procedure and the real operating practice have diverged. Humans work around the procedure to get the job done; the procedure is nominal.
The system produces more signals than humans can process, or produces signals that cannot be distinguished from noise. Critical signals are missed because the signal architecture is structurally unusable.
The system makes decisions humans cannot understand, override, or correct. Accountability is structurally absent because the decision process is opaque.
Humans have built informal systems — spreadsheets, workarounds, offline processes — to do the work the formal system cannot do. The shadow systems have become load-bearing.
The failure is not in the origin state or the target state — it is in the transition itself. The movement between states has not been designed as a governed sequence, and failure occurs in the space between.
Core diagnostic question: What is the system in movement between, and where does the transition itself produce failure that neither the origin nor the target state contains?
The system is currently in transition between states; the failure occurs in the transition itself rather than in the origin or target state; and the failure would not occur if the system were stable at either end.
The failure is in the target state design rather than the transition; or the failure is a steady-state failure that was present before the transition began.
The old system and the new system are running simultaneously, creating conflicting states, duplicate processes, and accountability gaps in the overlap zone.
The transition is being executed in the wrong order. Changes are being made to dependent systems before the systems they depend on have been changed.
During a merger, acquisition, or reorganisation, the new entity cannot establish a stable identity — who owns what, who has authority, whose processes govern — and the resulting ambiguity produces recurring failure.
An external deadline has compressed the transition timeline to the point where structural dependencies cannot be resolved in order. The deadline produces a structurally incomplete transition.
The institution compresses its own governance obligations onto the individual. The individual must carry administrative, evidential, reputational, or relational burden that the institution was designed to carry — and cannot carry it without losing access to what the institution controls.
Core diagnostic question: What institutional obligation has been transferred to the individual — and what does the individual lose when they cannot carry it?
An individual is absorbing burden that a governing institution should carry; that burden recurs structurally rather than as a one-off administrative friction; and the individual loses access or entitlement when they cannot carry the burden.
The administrative burden is genuinely reasonable given the nature of what is being accessed; or the failure is a one-off error rather than a structural pattern.
Access to a resource or entitlement is allocated through a process the individual cannot navigate without specialist knowledge, persistent effort, or insider information. The allocation mechanism compresses the institution's obligation to distribute access fairly onto the individual's capacity to navigate complexity.
The individual needs records to access entitlements, but the institution holds those records and will not release them without prior authorisation from the entitlement the individual cannot yet access. A closed loop where access requires proof and proof requires access.
Consent is obtained in a form — volume, complexity, or frequency — that makes genuine informed consent structurally impossible. The individual must consent without understanding, or lose access.
The institution places the burden of identity or eligibility verification on the individual in a form that creates structural exclusion for those who cannot provide the required documentation.
The individual's access or treatment is determined by a reputational signal they cannot understand, contest, or correct. The reputation system is opaque and the individual cannot navigate it.
A relationship that requires continuity — care, support, professional relationship — is interrupted by institutional structure, and the individual must rebuild it from scratch each time.
The institution makes commitments it cannot keep, and the individual must absorb the gap between the commitment and the reality without recourse.
Care — medical, social, or professional — is fragmented by institutional structure so that no single actor carries the individual's full picture. The individual must carry their own care continuity across a system designed to provide it.
An entitlement the individual had built up is reset — by system change, institutional restructuring, or policy shift — without acknowledgment or compensation. The individual must restart from zero.
The institution reduces the individual to a category or data point that does not capture their actual situation. The individual must carry the burden of being misrepresented in a system that makes decisions based on that misrepresentation.
The institution retains a record of the individual's past failure — a penalty, a debt, a classification — beyond the point where that record serves a structural purpose. The individual cannot escape the record.
The individual has been placed in the wrong category by the institution — the wrong risk band, the wrong eligibility tier, the wrong classification — and cannot correct the assignment through normal channels.
Ongoing treatment — medical, therapeutic, legal, or administrative — is interrupted by institutional structure rather than by the individual's situation. The interruption creates harm that the institution is not designed to track or repair.
Run the 60-Second Scan first. Four questions that surface your system's purpose, primitive, and failure signal — so you know which class to test before entering the criteria.
RUN THE 60-SECOND SCAN →The taxonomy names the class. The Diagnostic Artifact names the governing break inside your specific system and produces the correction sequence your builders execute against.
EMAIL FIVE-LINE BRIEFSend the repeated failure. Five lines and any relevant material.