Leave a sand castle on the beach. Wind, rain, and footsteps will reduce it to a formless mound. No wind, rain, or footsteps will spontaneously assemble a sand castle from a formless mound. This asymmetry — the universe's persistent tendency toward disorder rather than order — is entropy.
The concept originates in thermodynamics. In 1865, Rudolf Clausius formalised what steam engineers had observed for decades: every time energy is converted from one form to another, some portion becomes unavailable for useful work. Heat flows from hot objects to cold objects, never the reverse. A burning log converts chemical energy into heat and light, but the dispersed heat cannot spontaneously reconcentrate into a log. Clausius named this irreversible dissipation entropy, from the Greek entropia — a turning toward. The Second Law of Thermodynamics states that in any isolated system, entropy never decreases. It can remain constant in idealised reversible processes. In every real process, it increases.
Ludwig Boltzmann provided the statistical foundation in the 1870s. He demonstrated that entropy is not a mysterious tendency but a mathematical inevitability arising from probability. A system of gas molecules in a container has an astronomically larger number of disordered arrangements than ordered ones. If you shuffle the molecules randomly, they will almost certainly land in one of the vastly more numerous disordered states. The Second Law is not a force pushing systems toward disorder — it is the statistical certainty that disorder has overwhelmingly more ways to exist than order does. Boltzmann's equation, S = k log W, relates entropy (S) to the number of accessible microstates (W). Order is improbable. Disorder is not merely probable — it is practically inevitable.
Claude Shannon extended the concept to information theory in 1948. Shannon's entropy measures uncertainty — the average amount of information needed to describe the outcome of a random process. A message composed of the same letter repeated has zero entropy; it is perfectly predictable. A message composed of randomly selected letters has maximum entropy; each character is maximally surprising. Shannon showed that the thermodynamic concept and the information-theoretic concept share identical mathematical structure. Disorder in physics and uncertainty in communication are the same phenomenon viewed through different lenses.
For founders, investors, and decision-makers, entropy carries a specific and consequential implication: every system you build, maintain, or invest in is under constant entropic pressure. Organisations lose alignment. Codebases accumulate technical debt. Cultures drift from their founding values. Customer relationships erode. Competitive advantages decay. Brand meaning dilutes. None of this requires an external attack or a strategic mistake. Entropy does not need a cause. It is the default state of the universe. Order is the anomaly that requires continuous energy to maintain.
The entropic frame reorients the fundamental question of management from "how do we grow?" to "how do we prevent decay?" Growth is the visible, exciting work. Maintenance is the invisible, essential work. The organisations that endure are not the ones that build the most elaborate structures — they are the ones that invest the most consistent energy in preventing those structures from collapsing under their own complexity. Every meeting that drifts without an agenda, every process that persists without review, every hire made without rigour, every quarter where culture is assumed rather than reinforced — each is a micro-concession to entropy. Individually, each is trivial. Cumulatively, they are fatal.
The concept applies with particular force to information systems and organisations because these are high-complexity structures with enormous numbers of interacting components. A ten-person startup has forty-five pairwise relationships between team members. A hundred-person company has 4,950. A thousand-person company has nearly half a million. Each relationship is a channel through which misalignment, miscommunication, and conflicting assumptions can propagate. The entropy surface area of an organisation grows combinatorially with headcount — which is why companies that were coherent at twenty people become incoherent at two hundred without deliberate, energy-intensive efforts to maintain alignment. The organisational charts and all-hands meetings and written memos and cultural documents that large companies produce are not bureaucratic overhead. They are anti-entropy systems — structured investments of energy designed to counteract the combinatorial explosion of disorder that accompanies growth.
The deepest insight is temporal. Entropy is not a one-time threat you can address and move past. It is a continuous force operating on every ordered system at every moment. The founder who builds an extraordinary culture in year one and assumes it will persist is making the same error as the physicist who expects a hot cup of coffee to stay hot in a cold room. The energy required to maintain order is not a one-time investment — it is a perpetual tax levied by the universe on every structure that resists the drift toward randomness. The organisations that understand this allocate resources to maintenance, reinforcement, and renewal not as overhead but as the primary cost of existence.
This is why the longest-surviving institutions — religions, militaries, universities — all share a common structural feature: elaborate, repeating rituals designed to maintain internal order. The Catholic Mass has been performed billions of times over two millennia. Military drill is repeated daily, not because soldiers forget how to march but because unit cohesion decays without constant reinforcement. University commencement ceremonies, faculty governance procedures, and tenure reviews are entropy-management systems disguised as tradition. The rituals look ceremonial. They are thermodynamic. They are the continuous energy investment that prevents the institution's internal order from reverting to the disorder that the Second Law guarantees is always waiting.
Section 2
How to See It
Entropy is visible everywhere systems lose coherence without any single identifiable cause. The signature is gradual, undirected degradation — not the sharp decline caused by a bad decision or an external shock, but the slow erosion caused by the accumulation of small, unmanaged deviations from the original ordered state. The system does not break; it decays. The process is so gradual that participants inside the system often cannot perceive it until the cumulative damage becomes irreversible.
The diagnostic is straightforward: compare the current state of any system to its original design intent, then ask whether the gap is explained by deliberate decisions or by accumulated drift. If the gap is explained by drift — if nobody chose the current state, it just happened — entropy is the operating cause.
A second diagnostic is the energy audit: how much active effort is being directed at maintaining the system's current level of order? If the answer is "none" or "very little," and the system appears to be functioning, one of two things is true: either the system is genuinely low-entropy (simple, robust, with few components that can misalign), or the system is decaying and the decay has not yet become visible. In complex systems — organisations, codebases, relationships, markets — the second explanation is almost always correct.
A third diagnostic is the reversibility test: can the system be restored to its original ordered state with reasonable effort? A freshly painted house that has weathered for two years can be repainted. A culture that has drifted for a decade often cannot be restored, because the people who embodied the original culture have left, the institutional memory has degraded, and the new norms have calcified into the assumed way things work. The difficulty of restoration reveals the depth of entropic decay — and the exponential cost of deferred maintenance.
Organisations
You're seeing Entropy when a company's employee handbook describes a culture of "radical transparency and fast decision-making," but the lived reality involves three layers of approval for any initiative, meetings that exist because they've always existed, and a pervasive sense that "things used to move faster." No single policy change caused the drift. No individual is responsible. The organisation accumulated procedural complexity one reasonable-seeming addition at a time — each new approval step, each new standing meeting, each new reporting requirement — until the aggregate weight of accumulated process smothered the speed it was designed to protect.
Technology
You're seeing Entropy when a software system that launched clean and fast three years ago now takes forty-five seconds to load, crashes intermittently, and requires a dedicated team just to keep running. The original architecture was coherent. But three years of patches, workarounds, feature additions built on assumptions that are no longer true, deprecated dependencies that nobody removed, and undocumented modifications by engineers who have since left have transformed a designed system into an archaeological site. The codebase did not fail — it decayed. Each individual change was locally rational. The cumulative effect was systemic disorder.
Markets
You're seeing Entropy when a brand that once commanded premium pricing and fierce loyalty finds its market position eroding despite no obvious competitive attack. The product quality hasn't declined. The competitors haven't innovated dramatically. But the brand's meaning — the specific associations, emotional resonance, and trust that justified the premium — has gradually diluted through inconsistent messaging, line extensions that blurred the core identity, and a slow drift from the original value proposition. Brand entropy operates through the accumulation of small incoherences, not through a single catastrophic misstep.
Relationships
You're seeing Entropy when a founding team that once operated with seamless coordination and shared vision finds itself in recurring conflicts about strategy, spending, and priorities — not because anyone changed their values, but because the increasing complexity of the business created ambiguity that the original alignment could not cover. The founders' relationship did not break. It decayed, as the energy required to maintain alignment grew faster than the energy invested in maintaining it. Every unresolved disagreement, every assumption left unspoken, every difficult conversation deferred added a small increment of disorder to the partnership.
Section 3
How to Use It
Decision filter
"Before evaluating any system's growth potential, first evaluate its decay rate. Ask: what is actively preventing this system from degrading? If the answer is 'nothing specific,' the system is already degrading — you just haven't measured it yet."
As a founder
Treat entropy as your primary operational adversary. Every system you build — culture, codebase, team alignment, customer experience, operational processes — begins decaying the moment you stop actively maintaining it. The practical discipline is scheduled renewal: quarterly reviews of every standing process, annual reassessment of cultural norms, continuous refactoring of technical systems, and regular recalibration of team alignment. The cost of this maintenance feels like overhead. It is not overhead. It is the energy required to prevent the ordered system you built from reverting to the disordered state that requires no energy at all.
The entropic frame also transforms how you think about scaling. Every additional employee, product line, market, and process increases the surface area exposed to entropy. A ten-person startup with one product has a manageable entropy budget — the energy required to keep everything aligned is small relative to the energy available. A thousand-person company with twelve product lines in eight markets has an entropy budget that can consume the entire organisation's energy if not managed deliberately. The companies that scale successfully are the ones that scale their entropy-management systems — hiring practices, communication rituals, architectural standards, cultural reinforcement — at the same rate they scale their complexity. The companies that fail at scale are the ones that scale complexity while assuming entropy management will take care of itself.
As an investor
Use entropy as a diagnostic for organisational health. When evaluating a company, ask not just "how good is this business?" but "how fast is this business decaying, and how much energy is being directed at preventing decay?" The signals are specific and measurable: employee tenure and voluntary turnover trends, technical debt ratios, customer retention rates over multi-year horizons, the ratio of maintenance engineering to feature engineering, the frequency and rigour of process reviews.
A company with strong revenue growth but rising technical debt, increasing employee turnover, and declining customer satisfaction scores is a company where growth is outrunning entropy management. The financial statements show strength. The entropy indicators show a structure that is accumulating disorder faster than it is being repaired. The eventual outcome — a period of stagnation or decline during which the accumulated disorder must be addressed — is not a surprise. It is a thermodynamic inevitability that the financial statements were too slow to capture.
As a decision-maker
Apply entropy thinking to every system under your stewardship by distinguishing between creation energy and maintenance energy. Creation energy builds new structures — launches products, hires teams, enters markets, designs processes. Maintenance energy preserves existing structures — refactors code, reinforces culture, renews relationships, reviews processes. Most organisations dramatically overallocate to creation and underallocate to maintenance, because creation is visible and celebrated while maintenance is invisible and unrewarded.
The entropy-aware decision-maker budgets maintenance energy explicitly. For every new system created, the question is not just "what will this cost to build?" but "what will this cost to maintain indefinitely, and where will that energy come from?" A process that costs one hour to design and requires one hour per week to maintain has a four-week payback period on the design cost — and then becomes a permanent weekly energy drain. An organisation that creates ten such processes per quarter without retiring any is adding forty hours per week of maintenance load per year. Within three years, the maintenance burden consumes more energy than the organisation has available, and the system begins decaying faster than it can be repaired.
Common misapplication: Using entropy as an argument against building anything complex.
Entropy does not argue against complexity. It argues against unmanaged complexity. A jet engine is extraordinarily complex and operates with extraordinary reliability — because the maintenance regime is as sophisticated as the engineering. The problem is not complexity per se but complexity that exceeds the organisation's capacity to maintain it. An organisation that builds systems commensurate with its maintenance capacity can sustain high complexity indefinitely. An organisation that builds systems beyond its maintenance capacity will watch those systems decay regardless of how brilliantly they were designed.
A second misapplication is treating entropy as a metaphor rather than a structural force. When a founder says "entropy is eating our culture," the statement should prompt the same analytical rigour as "our servers are failing." What specifically is decaying? At what rate? What maintenance energy is currently being applied? What additional energy is required to arrest the decay? Entropy is not a vague concern — it is a measurable phenomenon with identifiable causes and quantifiable remedies. Treating it as a metaphor allows organisations to acknowledge the problem without committing to the specific, often unglamorous, maintenance work required to address it.
A third misapplication is assuming that entropy operates at a constant rate. It does not. Entropy accelerates with complexity. A system with ten components has forty-five possible pairwise interactions. A system with a hundred components has 4,950. The maintenance energy required to prevent disorder grows faster than linearly with system size — which is why organisations that were manageable at fifty people become ungovernable at five hundred without a fundamental rearchitecture of their entropy-management systems.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The founders who build durable organisations share a common structural insight that is rarely stated explicitly: the primary threat to any ordered system is not external attack but internal decay. These operators invest disproportionate energy in maintenance, renewal, and the prevention of disorder — not because they are conservative, but because they understand that the universe's default trajectory is toward randomness, and that every ordered system requires continuous energy to resist that trajectory.
The cases below share a diagnostic signature: each operator recognised that the system they had built was under constant entropic pressure and designed specific, repeating mechanisms to counteract it. The mechanisms differ — cultural rituals, architectural principles, organisational structures, communication practices — but the underlying logic is identical: order does not persist without energy, and the energy must be directed, sustained, and never assumed.
What separates these operators from their peers is not that they avoided entropy — nobody avoids entropy — but that they treated entropy management as a core executive function rather than a delegated operational concern. They invested personal energy, institutional resources, and strategic attention in the maintenance work that most leaders dismiss as beneath their role. The unglamorous, repetitive, never-finished work of keeping ordered systems ordered was, for each of them, the primary job.
Bezos's "Day 1" philosophy is the most explicit anti-entropy framework in modern corporate leadership. In his 2016 shareholder letter, Bezos defined "Day 2" as "stasis, followed by irrelevance, followed by excruciating, painful decline, followed by death." When asked how to fend off Day 2, his answer was a list of entropy-fighting mechanisms: obsessive customer focus, resisting proxies (substituting process for judgment), eagerly adopting external trends, and high-velocity decision-making.
Each mechanism targets a specific entropy pathway. Customer obsession prevents the organisation from optimising for internal convenience — the entropic drift that converts customer-serving processes into self-serving bureaucracies. Resisting proxies prevents the organisation from substituting measurable outputs (the process was followed) for meaningful outcomes (the customer was served). Adopting external trends prevents the organisation from calcifying around assumptions that the environment has already invalidated. High-velocity decision-making prevents the accumulation of deferred choices — each unresolved decision adding a quantum of disorder to the organisation's strategic coherence.
The six-page memo requirement for senior meetings is an anti-entropy tool disguised as a communication format. PowerPoint slides — with their bullet points, animations, and sequential reveals — allow presenters to obscure fuzzy thinking behind polished production. A six-page narrative forces the writer to construct a coherent argument with complete sentences, logical transitions, and explicit assumptions. The format is a forcing function for intellectual order — a direct investment of energy against the entropy that accumulates when organisations allow imprecise thinking to masquerade as strategic clarity.
Jobs's return to Apple in 1997 was one of the most dramatic entropy reversals in corporate history. In his absence, Apple had drifted from a focused producer of distinctive personal computers into a sprawling, incoherent company shipping dozens of products across overlapping categories — printers, PDAs, game consoles, a television appliance, multiple indistinguishable computer lines. The product line was the organisational expression of maximum entropy: complexity without coherence, activity without direction.
Jobs's first act was radical simplification. He drew a two-by-two matrix — consumer/professional on one axis, desktop/portable on the other — and declared that Apple would make four products, one for each quadrant. Everything else was cancelled. The act was not merely strategic. It was thermodynamic. Jobs was injecting massive energy into the system to reduce its entropy — collapsing the disordered product proliferation back into a coherent, low-entropy structure that the organisation could maintain.
The subsequent two decades at Apple reflected the same anti-entropy discipline applied continuously. The product line remained ruthlessly constrained. Materials, finishes, and design language were unified across categories. The retail stores were controlled environments designed to prevent the brand-entropy that occurs when third-party retailers present your product alongside competitors in undifferentiated settings. Jobs understood — viscerally if not in thermodynamic language — that the default state of any product line is proliferation, the default state of any brand is dilution, and the default state of any design language is incoherence. Maintaining simplicity requires more energy than allowing complexity. He spent that energy relentlessly.
Charlie MungerVice Chairman, Berkshire Hathaway, 1978–2023
Munger's mental model framework is an anti-entropy system for cognition. The human mind, left to its default operations, accumulates cognitive disorder — biases compound, assumptions calcify, mental models narrow, and the gap between the map and the territory widens silently. Munger's prescription — building a "latticework of mental models" drawn from multiple disciplines — is a structured investment of energy against intellectual entropy.
The discipline is maintenance, not acquisition. Munger did not merely collect models — he continuously tested them against new evidence, calibrated their domains of applicability, and identified where they conflicted. The Annual Meeting of Berkshire Hathaway was, among other things, a public entropy audit: Munger and Buffett subjecting their own reasoning to hours of unscripted questioning, identifying where their thinking had drifted, and recalibrating in real time.
Munger's emphasis on avoiding stupidity rather than seeking brilliance is entropy thinking applied to decision quality. Brilliance is a one-time energy injection — a single great decision. Stupidity avoidance is continuous maintenance — a perpetual practice of reviewing, questioning, and correcting the cognitive decay that accumulates when assumptions go unexamined. "It is remarkable how much long-term advantage people like us have gotten by trying to be consistently not stupid, instead of trying to be very intelligent." The insight is thermodynamic: the energy required to maintain consistent decision quality (low cognitive entropy) is smaller and more sustainable than the energy required to produce occasional genius.
Grove's management philosophy was built on the recognition that organisational entropy is the default state and that managerial energy must be continuously directed against it. His concept of "constructive confrontation" — institutionalised debate as a standard operating procedure — was designed to prevent the specific form of entropy that accumulates when disagreements go unspoken and decisions go unchallenged.
In most organisations, the path of least resistance is agreement. Challenging a colleague's proposal costs social capital. Accepting it costs nothing immediately. Over time, this asymmetry produces organisational entropy: decisions are made without rigorous examination, assumptions propagate unchecked, and the organisation's strategic coherence decays as uncommitted initiatives accumulate. Grove's solution was to make confrontation the default — to invest energy in surfacing disagreement as a prophylactic against the slow decay of decision quality.
Grove's "strategic inflection point" framework is entropy thinking applied to competitive positioning. A strategic inflection point is the moment when the accumulated changes in a company's environment — each individually small — cross a threshold and the existing strategy becomes fundamentally misaligned. The insight is that competitive entropy accumulates continuously and invisibly, then manifests suddenly. By the time the inflection point is obvious, the decay is often too advanced to reverse. Grove's paranoia was not personality — it was the operational recognition that order in a competitive environment is always temporary and that the energy required to detect and respond to decay must be invested before the decay becomes visible.
His insistence on measurable objectives and regular one-on-one meetings — documented in High Output Management — was an anti-entropy maintenance system for organisational alignment. Each meeting was an energy injection into the manager-report relationship, preventing the drift of priorities and the accumulation of unspoken disagreements that silently degrade an organisation's coherence. Grove understood that alignment is not a state to be achieved but a process to be maintained — and that the maintenance cadence must match the rate at which the organisation's internal entropy naturally accumulates.
Shannon's contribution to entropy thinking is foundational in a way that transcends his information theory. By demonstrating the mathematical equivalence between thermodynamic entropy and information entropy, Shannon revealed that disorder and uncertainty are the same phenomenon — and that the cost of maintaining order in any system, whether physical or informational, is governed by the same quantitative laws.
Shannon's practical insight for organisational thinkers is the concept of channel capacity — the maximum rate at which information can be transmitted reliably through a noisy channel. Every communication system in an organisation — from Slack channels to all-hands meetings to one-on-ones — is a noisy channel with finite capacity. When the information that needs to be transmitted (strategic priorities, cultural norms, operational decisions) exceeds the channel capacity, errors accumulate. Misunderstandings propagate. Decisions are made on incomplete or incorrect information. The organisation's information entropy increases — not because anyone failed, but because the communication infrastructure could not carry the informational load that the organisation's complexity required.
Shannon's framework explains why scaling organisations is so difficult. Each additional employee increases both the information that must be transmitted (more people need to know more things) and the noise in the system (more channels, more intermediaries, more opportunities for signal degradation). The organisation that does not systematically invest in expanding its communication channel capacity — through better documentation, clearer decision frameworks, more structured meeting rhythms, and deliberate redundancy in critical information flows — will experience information entropy that grows faster than its headcount. Shannon proved mathematically what every founder discovers empirically: clarity at scale is not free, and the cost increases nonlinearly with complexity.
Section 6
Visual Explanation
Section 7
Connected Models
Entropy is the foundational force that explains why every other defensive and structural model in the lattice exists. Margins of safety exist because systems decay. Technical debt accumulates because code entropy is real. Creative destruction occurs because organisational entropy eventually makes incumbents fragile enough to displace. The models below represent the frameworks that either implement entropy management, create productive tension with entropy's implications, or describe the downstream consequences that entropy inevitably produces.
The most durable strategic positions are built by operators who understand entropy's interaction with these adjacent frameworks — using maintenance energy to slow decay while investing in the structural mechanisms that channel entropy's inevitable effects toward productive outcomes rather than catastrophic ones. The six connections below map how entropy interacts with frameworks that implement its maintenance logic, create productive friction with its conservative implications, or describe the downstream consequences that unmanaged entropy inevitably produces in competitive and organisational systems.
Reinforces
Technical Debt
Technical debt is entropy made concrete in software systems. Every shortcut, workaround, and deferred refactoring decision adds disorder to a codebase — increasing the number of implicit dependencies, undocumented assumptions, and fragile coupling points that make the system harder to understand, modify, and maintain. Entropy predicts that this accumulation is not an aberration but the default trajectory of any codebase that receives new features without corresponding maintenance energy. The reinforcement is bidirectional: entropy theory explains why technical debt accumulates (disorder is the default state of complex systems), and technical debt provides the most measurable, tangible instance of entropy operating in an organisational context. The teams that manage technical debt effectively are performing entropy maintenance — investing energy to preserve the codebase's structural order against the constant drift toward disorder.
Reinforces
[Compounding](/mental-models/compounding)
Compounding and entropy are mirror forces operating in opposite directions. Compounding describes how small, consistent inputs of energy accumulate into exponentially growing order — wealth, knowledge, reputation, capability. Entropy describes how the absence of energy inputs produces exponentially accelerating disorder. The reinforcement is structural: compounding only operates on systems that entropy has not yet destroyed. An investment that compounds at 10% annually for thirty years produces a 17x return — but only if the investor survives every intervening year without a ruin event. The most powerful compounders are the systems with the most effective entropy management: Berkshire Hathaway's returns compound because Buffett's survival-first framework prevents the entropy of catastrophic loss. The two models are inseparable — compounding is the reward for winning the war against entropy.
Tension
Section 8
One Key Quote
"The entropy of the universe tends to a maximum."
— Rudolf Clausius, The Mechanical Theory of Heat (1865)
The statement is the most concise expression of the most consequential asymmetry in physics. It explains why time has a direction, why broken cups do not reassemble, why organisations decay without maintenance, and why the default trajectory of every ordered system — from a star to a startup — is toward disorder. Six words that describe the universe's deepest structural bias.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Entropy is the mental model that explains why most organisations die — not from a single catastrophic failure, but from the slow, invisible accumulation of disorder that eventually makes the structure too fragile to withstand even a modest external shock. The postmortem always identifies a proximate cause: a competitive disruption, a leadership failure, a market shift. But the real cause, in most cases, is the decade of deferred maintenance that preceded the shock — the entropic decay that made the organisation brittle enough for the shock to be fatal.
The reason this model belongs in Tier 1 is that it reframes the fundamental task of leadership. Most management literature focuses on creation — building products, entering markets, designing strategies, inspiring teams. Entropy says that creation is the easy part. The hard part — the part that determines whether the created thing survives — is maintenance. The energy required to prevent decay is chronic, unglamorous, and never finished. It does not produce quarterly earnings surprises or feature launches. It produces the absence of failure, which is cognitively invisible and organisationally unrewarded. The leaders who build durable institutions are the ones who understand that the absence of visible failure is not evidence of a healthy system — it may be evidence of decay that has not yet manifested.
The most actionable insight is the entropy budget. Every organisation has a finite amount of energy — measured in human attention, capital, and time. Every system the organisation maintains — its codebase, its culture, its processes, its relationships, its brand — consumes a share of that energy budget just to remain at its current level of order. When the total maintenance cost of all systems exceeds the organisation's energy budget, some systems begin decaying unmanaged. The decay is silent at first. It manifests as "things feel slower," "quality isn't what it used to be," "we've lost our edge." These are not cultural complaints. They are thermodynamic reports. The organisation's entropy production has exceeded its entropy management capacity.
The scaling trap is the most dangerous expression of this dynamic. Every new product line, market, process, team, and tool increases the organisation's entropy surface area — the total volume of systems that require maintenance energy to remain ordered. If the organisation scales its complexity faster than it scales its maintenance capacity, a gap opens between entropy production and entropy management. The gap compounds. Each unmanaged system generates secondary disorder — the decaying codebase creates bugs that require emergency fixes that pull engineers from maintenance of other systems that then begin decaying themselves. The cascade is thermodynamic: entropy in one subsystem increases the entropy of connected subsystems, producing a positive feedback loop that can consume the entire organisation.
Section 10
Test Yourself
Entropy hides wherever systems appear to be functioning but are actually decaying beneath the surface. The diagnostic question is always: is this system being actively maintained, or is it coasting on the order that was created during its initial construction? In complex systems — organisations, technologies, relationships, strategies — the answer determines whether the system will endure or quietly collapse when the accumulated decay finally exceeds the structure's load-bearing capacity.
The most common analytical error is attributing entropic decay to a specific cause — a bad hire, a wrong decision, a competitive attack — when the real cause is the absence of maintenance energy applied to a system that was slowly degrading over a long period. The second most common error is assuming that current performance indicates current health. Entropy produces a lag between the onset of decay and the appearance of symptoms — the system continues functioning on the residual order from its initial construction long after maintenance has ceased. By the time the symptoms are visible, the decay is often advanced, and the energy required to restore order is exponentially greater than the energy that would have been needed to maintain it continuously.
A third common error is confusing entropy with external disruption. Entropy operates from within, through the accumulation of small internal deviations from the designed state. External disruption operates from without, through a competitive or environmental shock. The two interact — entropy makes systems fragile, and disruption exploits that fragility — but they are distinct forces with distinct remedies. Entropy requires maintenance. Disruption requires adaptation. Confusing them leads to organisations that invest in competitive response when the real problem is internal decay, or invest in internal maintenance when the real problem is environmental change.
Is Entropy at work here?
Scenario 1
A SaaS company's deployment pipeline that once took 15 minutes now takes 3 hours. No single change caused the slowdown. Over two years, hundreds of small additions — extra test suites, configuration checks, monitoring hooks, legacy compatibility layers — accumulated without anyone auditing the pipeline's total complexity. Each addition was individually reasonable.
Scenario 2
A restaurant chain opens 200 locations in five years. The original locations maintain high food quality and consistent customer experience. The newer locations show declining review scores, higher staff turnover, and increasing customer complaints about inconsistency. The menu, suppliers, and training materials are identical across all locations.
Scenario 3
A cybersecurity firm experiences a major data breach. Investigation reveals that the breach exploited a vulnerability in a third-party library that had been flagged for update eighteen months earlier. The update was deprioritised in favour of feature development in six consecutive sprint cycles. The security team's warnings were acknowledged but never acted on.
Section 11
Top Resources
The entropy framework spans thermodynamics, information theory, and systems thinking. Clausius and Boltzmann provide the physical foundation — the laws that govern energy dissipation and the statistical mechanics that explain why disorder is overwhelmingly more probable than order. Shannon extends the concept to information and communication. The applied literature — from software engineering to organisational theory — demonstrates how the same thermodynamic principles operate on the systems that founders and investors build and maintain.
The strongest foundation combines the physical theory (understanding why entropy operates) with the applied frameworks (understanding how it manifests in organisations, code, and markets). Avoid treatments that reduce entropy to a vague metaphor for "things fall apart." The concept's power lies in its specificity: entropy is measurable, its mechanisms are identifiable, and its remedies — energy investment in maintenance — are concrete and actionable. The best resources combine rigorous theoretical grounding with practical frameworks for diagnosing and counteracting entropic decay in the systems that founders, investors, and decision-makers actually build and maintain.
The paper that extended entropy from physics to information. Shannon's framework provides the mathematical tools for measuring disorder in any system that processes information — which includes every organisation, market, and communication channel. The paper demonstrates that the entropy of a message source determines the minimum resources needed to transmit information reliably, establishing the theoretical foundation for understanding why maintaining clarity in complex systems requires continuous energy investment. Technically dense but foundational.
The original formulation of entropy and the Second Law of Thermodynamics. Clausius's work establishes the physical foundation that every subsequent application of entropy — from information theory to organisational decay — builds upon. The core insight — that every real process increases the total entropy of the universe, making energy progressively less available for useful work — provides the thermodynamic grounding for understanding why maintenance is not optional but existential.
Norman's treatise on human-centred design is, beneath the surface, a book about fighting entropy in product systems. Every design principle Norman advocates — visibility, feedback, constraints, mapping — is a mechanism for maintaining order in the interaction between humans and objects. Poor design is high-entropy design: ambiguous affordances, unclear feedback, and unconstrained actions that allow the system to drift into states the designer never intended. The book provides the practical framework for understanding how entropy operates on user experience and how design energy can counteract it.
Meadows provides the systems-thinking framework that translates entropy from a physics concept into an organisational diagnostic. Her analysis of feedback loops, stocks and flows, and system archetypes explains the mechanisms through which entropy operates on complex adaptive systems — and the leverage points where targeted energy investment can most effectively counteract decay. The chapter on system traps describes the specific pathways through which unmanaged entropy destroys organisations.
Taleb's framework extends beyond entropy resistance (robustness) to entropy exploitation (antifragility). While entropy describes the universal tendency toward disorder, Taleb demonstrates that certain systems are structured to benefit from disorder — gaining strength from stressors that would destroy fragile systems. The book provides the conceptual bridge between understanding entropy as a threat and designing systems that harness it as a source of improvement. The most durable organisations are not merely entropy-resistant; they are antifragile — converting the disorder that destroys their competitors into fuel for their own adaptation.
Leaders who apply this model
Playbooks and public thinking from people closely associated with this idea.
Entropy — All ordered systems tend toward disorder unless energy is continuously invested to maintain structure. Order is the anomaly; decay is the default.
Creative Destruction
Creative destruction is what happens when entropy wins at the organisational level — when an incumbent's accumulated disorder creates enough fragility that a new entrant can displace it. The tension is perspective-dependent. From the incumbent's viewpoint, entropy is the enemy: the internal decay that made the organisation vulnerable to disruption. From the ecosystem's viewpoint, entropy is the mechanism of renewal: the force that clears away calcified structures and creates space for new ones. Schumpeter celebrated creative destruction as capitalism's engine of progress. Entropy explains why the engine never stops: every new structure begins decaying from the moment of its creation, guaranteeing that today's disruptor becomes tomorrow's incumbent becomes the next cycle's displaced firm.
Tension
[Ergodicity](/mental-models/ergodicity)
Entropy and ergodicity create a productive tension around the question of whether average outcomes describe individual experience. Entropy says that all ordered systems decay toward disorder over time — implying that the long-run average is worse than any single snapshot might suggest. Ergodicity analysis distinguishes between the ensemble average (the average across many systems at one moment) and the time average (one system's experience across many moments). In entropic systems, the time average is systematically worse than the ensemble average, because entropy compounds through sequential periods in ways that ensemble snapshots do not capture. A portfolio of companies may show strong average returns while individual companies within it are decaying at rates the average conceals. The tension forces the analyst to ask: is this reported average describing a population snapshot or a time trajectory — and is entropy eroding the trajectory faster than the snapshot reveals?
Leads-to
Margin of Safety
Entropy leads directly to the demand for margin of safety, because any system subject to entropic decay will eventually perform worse than its current state suggests. Graham's margin of safety — buying assets at a discount to intrinsic value — is an implicit acknowledgement that the intrinsic value estimate is itself subject to entropy: the assumptions will degrade, the competitive position will erode, the management will lose alignment, the market will shift. The margin absorbs the decay that entropy guarantees will occur. An investor who pays exactly intrinsic value is betting that no entropy will operate on the asset. An investor who demands a 40% discount is budgeting for the entropic decay that thermodynamics guarantees. The margin of safety is not conservatism — it is entropy insurance.
Leads-to
Adaptation & Red Queen Effect
Entropy leads to the Red Queen because the continuous decay of competitive advantages forces continuous adaptation. A company's current advantage — its technology lead, brand equity, cost structure, talent base — is subject to entropy from the moment it is established. Competitors do not need to be brilliant; they merely need to match the decaying advantage faster than the incumbent can maintain it. The Red Queen race is what happens when entropy operates simultaneously on all competitors: each must invest continuous energy merely to maintain their current position relative to rivals who are also decaying and also investing energy to counteract their own decay. The Red Queen's "running in place" is the competitive expression of entropy's thermodynamic law: maintaining the current state of order requires continuous energy, and the moment that energy is withdrawn, decay begins.
The most common strategic error I see is treating maintenance as optional overhead rather than existential infrastructure. The startup that stops investing in cultural reinforcement once the culture "feels right." The software company that stops refactoring once the code "works." The brand that stops investing in positioning once the market share "is established." Each is making the same thermodynamic error: assuming that an ordered system will remain ordered without continuous energy input. The coffee will cool. The garden will grow weeds. The culture will drift. The code will rot. The brand will dilute. The question is not whether entropy will operate. The question is whether you are investing enough energy to counteract it.
Claude Shannon's extension of entropy to information provides the deepest insight for knowledge workers. In information-theoretic terms, entropy measures the uncertainty in a system. An organisation with high information entropy is one where employees are uncertain about priorities, strategies are ambiguous, roles are unclear, and communication is noisy. Reducing information entropy — creating clarity about what matters, why it matters, and what each person should be doing — requires the same continuous energy input that reducing physical entropy requires. A single all-hands meeting does not permanently reduce information entropy any more than a single cleaning permanently reduces the entropy of a room. The clarity decays the moment people leave the room and encounter the ambiguity of daily operations. Information entropy must be fought continuously, through repeated communication, consistent reinforcement, and structured review.
The investor's application is equally direct. When evaluating a company's durability, the entropy lens asks: what is this company's maintenance energy budget, and how does it compare to the entropy its complexity generates? A company with a pristine balance sheet, a focused product line, a small team, and a narrow market has low entropy exposure — the maintenance cost is manageable. A company with a sprawling product portfolio, thousands of employees across dozens of markets, a fragmented technology stack, and a culture sustained by momentum rather than deliberate reinforcement has massive entropy exposure — and the question is whether its maintenance investment matches that exposure. The companies that report "unexpected" declines are almost never surprising to an entropy-aware observer. The decay was visible in the gap between complexity and maintenance long before the financial statements registered the consequences.
My operational framework: for every unit of complexity I add to a system, I ask what maintenance regime will prevent it from decaying, and whether I have the energy budget to sustain that regime indefinitely. If the answer is no, I either simplify the addition or expand the maintenance capacity before adding the complexity. The organisations I've seen fail most catastrophically were the ones that added complexity enthusiastically and addressed maintenance reluctantly — building elaborate structures on foundations that were quietly decaying. The organisations I've seen endure were the ones that treated maintenance as a first-class activity: budgeted, staffed, measured, and rewarded with the same seriousness as creation. They understood what Clausius formalised a century and a half ago: the entropy of the universe tends to a maximum. The only question is whether your particular corner of it will resist that tendency long enough to matter.
Scenario 4
A venture-backed startup pivots from B2B to B2C after its Series A. The new strategy, product, and target market are entirely different from the original plan. The previous B2B product is shut down. Customer acquisition begins from zero.
