A startup with ten people ships code twice a day. Decisions happen in hallway conversations. Everyone knows what everyone else is working on. The founder spends 90% of her time building. Two years later, the company has fifty people. There are four approval layers for a feature release. A Slack message takes two days to get a clear answer. Three teams are accidentally building the same thing. The founder spends 90% of her time in meetings. Nobody made a catastrophic decision. No villain torpedoed the culture. The company simply grew, and somewhere in that growth, everything got harder, slower, and more tangled. A physicist would use one word: entropy.
Entropy is the most underrated concept in business thinking. Most people vaguely associate it with chaos or disorder, then file it away as a physics thing. That's a mistake. Entropy is not chaos. It is the natural tendency of any system to drift from order toward disorder unless energy is continuously invested to maintain structure. It explains why your email inbox gets worse, why companies calcify into bureaucracies, why your morning routine dissolves into scrolling. Once you understand the mechanism, you can fight it.
What Is Entropy, Really?
In physics, entropy is a measure of the number of possible arrangements a system can take. A low-entropy state has few possible arrangements (everything in its place, highly ordered). A high-entropy state has many possible arrangements (everything scattered, disordered). The second law of thermodynamics says that in any closed system, entropy never decreases. It either stays the same or increases. Always. No exceptions.
Think of a desk. A perfectly organized desk has one arrangement: every item in its designated spot. That is low entropy. There is only one way for it to be "perfectly organized" and thousands of ways for it to be messy. Without deliberate effort (you putting things back), the desk drifts toward mess. Not because a gremlin rearranges your pens. Because there are overwhelmingly more disordered states than ordered ones, and every small disturbance nudges the system toward the statistical majority. Drop an egg and it shatters. You will never see a shattered egg spontaneously reassemble itself. Entropy is the reason things break and don't unbreak.
Here is the critical detail most people miss: the second law applies to closed systems. A closed system exchanges no energy or matter with its surroundings. In a closed system, entropy always increases. But open systems can decrease their internal entropy by importing energy from outside. Your refrigerator keeps food cold (local entropy decrease) by pumping heat outward and consuming electricity (total entropy of the larger system still increases). A living cell maintains its complex internal structure by consuming food and expelling waste.
This is the single most important insight entropy offers to business: if your organization behaves like a closed system, entropy wins. Always. The only way to fight entropy is to keep the system open, to continuously inject energy in the form of attention, communication, pruning, and deliberate restructuring. Stop investing that energy, and decay is not a possibility. It is a certainty.
Entropy in Physics vs. Entropy in Business
The physics metaphor maps onto organizations with uncomfortable precision. Here is where the concept gets practical.
| Concept | In Physics | In Business |
|---|---|---|
| Entropy | Measure of disorder in a system | Accumulated complexity, misalignment, and friction in an organization |
| Second Law | Entropy always increases in closed systems | Complexity always grows unless actively managed |
| Closed System | No energy or matter exchanged with surroundings | Organization that stops reflecting, restructuring, and pruning |
| Energy Input | Work done to reduce local entropy (e.g., refrigerator) | Leadership attention, clear communication, active simplification |
| Heat Death | Maximum entropy; no usable energy remains | Total bureaucratic gridlock; the organization produces nothing but process |
| Low Entropy | Highly ordered, few possible states | Clear roles, aligned goals, minimal redundant process |
| High Entropy | Highly disordered, many possible states | Confused ownership, competing priorities, process for the sake of process |
The core mechanism transfers cleanly. Systems drift toward disorder, not because of bad actors, but because disordered states vastly outnumber ordered ones. Every new hire, product line, and communication channel adds possible arrangements. Without deliberate effort, the system migrates toward the most probable state: a mess.
Organizational Entropy: How Companies Calcify
Every company starts lean. The early team has shared context. Communication is direct. Decisions are fast because everyone sees the same information and trusts each other's judgment. This is a low-entropy state, and it feels effortless because the system is small enough to be naturally ordered.
Then the company grows. Each new person adds communication pathways. In a team of n people, there are n(n-1)/2 possible pairwise connections. A team of 10 has 45 channels. A team of 50 has 1,225. A team of 200 has 19,900. The information that used to flow naturally through hallway conversations now needs meetings, documents, Slack channels, wikis, and status reports. This routing infrastructure is overhead, the organizational equivalent of heat loss.
But the real entropy kicks in through process accumulation. Every problem generates a process designed to prevent recurrence. Someone shipped a bug? Add a code review requirement. A customer got wrong information? Add an approval workflow. A team overspent? Add a procurement sign-off. Each process is individually reasonable. Collectively, they are entropy. After five years, the company has hundreds of processes, many addressing problems that no longer exist, and nobody remembers why half of them were created.
This is how a startup that shipped a feature in a day becomes a company that needs six weeks and four sign-offs for the same feature. Not because anyone decided to slow things down. Because entropy happened, one reasonable process at a time. If you study operations and process optimization, you learn that the hardest part is not designing good processes. It is killing bad ones that outlived their purpose.
Communication Entropy: The Signal Degrades
Information is subject to its own form of entropy. Every time a message passes from one person to another, it loses fidelity. Details get dropped. Nuance gets flattened. Context that was obvious to the sender is invisible to the receiver.
Think of it as a game of telephone at organizational scale. The CEO explains a strategic pivot to the VP. The VP explains it to the directors. The directors explain it to the managers. At each hop, the message degrades. Not because anyone is incompetent, but because each person translates the message through their own context and priorities. The result is four different versions of the strategy being executed simultaneously, with nobody realizing they are out of alignment until the project fails.
Communication entropy increases with organizational distance. Two people sitting next to each other share context and can ask clarifying questions instantly. Two people in different departments, time zones, and reporting structures lose signal on every exchange. This is why co-located teams tend to outperform distributed teams on tightly coupled projects. Not because remote work is inherently inferior, but because the communication entropy is lower when the distance is shorter.
A VP of Product tells her three directors: "We need to focus on enterprise customers this quarter." Director A interprets this as "build the features our biggest enterprise prospect asked for." Director B interprets it as "stop spending time on our self-serve product." Director C interprets it as "hire enterprise sales reps." All three are reasonable interpretations. All three lead to different resource allocation. By the time the misalignment surfaces in a quarterly review, the team has spent three months pulling in three directions. The message was not unclear. It was under-specified, and each hop added entropy.
The antidote to communication entropy is redundancy and specificity. Say the same thing in multiple formats. Be specific enough that misinterpretation is difficult. Check understanding by asking people to repeat back what they heard. This feels like over-communicating. It is the energy input required to keep entropy from eating your alignment.
Product Entropy: Tech Debt, Feature Bloat, and Process Creep
Products are entropy machines. Every feature added creates maintenance surface area. Every shortcut taken during a deadline creates technical debt. Every customer request accommodated adds complexity to the codebase, the UI, and the support burden. Over time, a product that started as a clean, focused tool becomes a sprawling platform that does many things adequately and nothing brilliantly.
Technical debt is entropy made concrete. It is the gap between the code you have and the code you would write if you could start over with everything you know now. Small amounts are inevitable and even useful (shipping fast sometimes matters more than shipping clean). But technical debt compounds. Each shortcut makes the next feature harder to build, which creates pressure to take more shortcuts. The system accelerates toward disorder.
Feature bloat follows the same pattern. A product manager adds a feature because a big customer asked for it. Then another for another customer. Then a settings toggle because different customers want different behaviors. Each addition is individually justified. Collectively, they transform the product from a scalpel into a Swiss Army knife that nobody can figure out without documentation.
Process creep is the operational version. A workflow that started as "designer hands mockup to engineer" becomes "designer creates mockup, posts in Figma, tags PM for review, PM approves, PM creates Jira ticket, engineer picks up ticket, engineer posts PR, PR gets two reviews, QA tests in staging, PM approves staging build, deploys to production." Each step was added for a reason. The workflow now takes ten times longer than it originally did.
Fighting Entropy: The Energy You Must Invest
The physics is clear: you cannot eliminate entropy. You can only manage it by continuously investing energy. In business terms, that energy takes specific forms.
Clear, redundant communication. Entropy degrades signals. Fight back by over-communicating key messages in multiple channels and formats. The leader who says the same strategic priority in every all-hands, every team meeting, and every one-on-one is not being repetitive. They are doing the thermodynamic work required to maintain alignment.
Active pruning. Most organizations only add. They add processes, tools, meetings, reports. Very few subtract. The entropic cost of addition without subtraction is guaranteed complexity growth. Kill processes that no longer serve their original purpose. Sunset features that fewer than 5% of users touch. Cancel recurring meetings that have become status theater. Pruning is not downsizing. It is maintenance. A gardener who only plants and never prunes ends up with a tangled mess, not a garden.
Simplification rituals. Schedule regular reviews aimed at reducing complexity. Once a quarter, look at your processes, your product, your communication structures, and ask: what can we remove, merge, or simplify without losing the thing it was protecting? If the answer is "nothing," you are not looking hard enough. There is always entropy to burn off.
Small teams, clear ownership. Entropy increases with the number of connections in a system. Smaller teams have fewer connections and therefore lower entropy. Amazon's two-pizza team rule is an entropy management strategy, whether Bezos framed it that way or not. Small teams with clear ownership of specific outcomes have less coordination overhead and less process accumulation than large, amorphous groups.
Feedback loops. In physics, open systems fight entropy by exchanging energy with their environment. In organizations, the equivalent is feedback: customer feedback, employee feedback, performance data, market signals. A company that stops listening to external input becomes a closed system. And in closed systems, entropy always wins.
Personal Entropy: Why Your Productivity Decays
Entropy does not only operate at the organizational level. Your personal systems are subject to the same physics.
Think about the last time you built a productivity system. Morning routine, task manager, filing system, workout schedule. For the first two weeks, everything hummed. Clear inputs, clear outputs, low friction.
Then entropy started. You skipped one workout because you slept poorly. Your task manager accumulated items you never reviewed. Your filing system got a few documents in the wrong folder, and you stopped trusting it. Within two months, the system had degraded back to something resembling your pre-system state. You did not decide to abandon it. Entropy happened.
This is why most productivity advice fails. It tells you to build a system but never mentions that every system requires continuous energy to maintain. A personal operating system is not something you set up once and forget. It is something you invest energy in every week: reviewing, pruning, adjusting. The maintenance is not a failure of the system. It is the system.
Habit slip follows the same curve. Missing one day barely matters. Missing two days cuts the probability of continuation significantly. Missing three essentially resets you to zero. The habit is a low-entropy state (consistent behavior, one arrangement). Breaking it opens up a high-entropy space (many possible behaviors, no default). Getting back requires the same investment of energy as building the habit originally.
Clear morning routine. Weekly review of priorities. Inbox at zero (or close). Task system trusted and current. Calendar reflects actual priorities. Decisions feel easy because frameworks are in place. Energy invested in maintenance: 2-3 hours per week.
No consistent routine. 347 unread emails. Task list is a graveyard of stale items. Calendar is a patchwork of other people's priorities. Every decision requires full deliberation because no frameworks exist. Constant low-grade anxiety about forgotten commitments.
The Entropy Audit: A Practical Framework
You can measure and manage entropy in your organization (or your personal life) with a structured audit. Run this quarterly, or whenever things start feeling "heavier" than they should.
List every recurring process, meeting, report, and approval workflow. For each one, write down: (a) what problem it was created to solve, (b) whether that problem still exists, and (c) when was the last time this process actually prevented the problem. If you cannot answer (a) or the answer to (b) is "no," you have found entropy. Mark it for removal or simplification.
Identify how information flows from leadership to execution. Count the hops. Each hop is an entropy point where signal degrades. If a strategic decision passes through more than two layers before reaching the people who execute it, you have a communication entropy problem. Consider how to shorten the path: skip-level meetings, written memos that go direct, or recorded video explanations that everyone receives unfiltered.
How long does it take to go from "idea approved" to "live in production" (or the equivalent in your field)? Compare this to how long it took one year ago. Two years ago. If the timeline has grown, identify where the new time is being spent. Each added step is accumulated entropy. Calculate the ratio of time spent building vs. time spent coordinating, approving, and waiting. If coordination time exceeds building time, entropy is winning.
Armed with your audit data, cut. Kill processes whose original problems no longer exist. Merge redundant meetings. Remove approval layers that add time but not quality. Sunset product features that fewer than 5% of users engage with. The default organizational instinct is to keep everything and add more. Overrule that instinct deliberately. If something survives the audit without a clear, current justification, it goes.
Entropy does not take quarters off. If you run this audit once and never again, the system will re-accumulate complexity within months. Put the next audit on the calendar before you finish the current one. Quarterly works for fast-moving organizations. Biannually works for more stable ones. The rhythm matters more than the frequency.
The Entropy Curve: Why Older Organizations Struggle More
There is a predictable relationship between organizational age, size, and process complexity. It is exponential, because entropy compounds.
A 2-year-old company with 20 people might have 5 formal processes. A 5-year-old with 100 people might have 60. A 10-year-old with 500 people might have 400. Processes do not scale linearly with headcount. They scale with headcount multiplied by time, because each year adds new layers of accumulated solutions to problems, many of which are no longer active.
This is why large, old companies feel slow. It is not that the people are less talented. It is that accumulated entropy (process layers, approval chains, communication overhead, legacy systems) consumes an increasingly large share of the organization's total energy. At some point, the majority of effort goes to maintaining internal structure rather than producing external value. That is organizational heat death: maximum entropy, no usable energy left for actual work.
The companies that stay productive at scale invest disproportionate energy in fighting entropy. Amazon's leadership principles, Spotify's squad model, Apple's small-team obsession: all entropy management strategies designed to counteract the natural drift toward complexity.
Entropy as a Thinking Tool
Once you internalize entropy as a mental model, you start seeing it everywhere. And that is the point. It is a genuine thinking tool that changes how you diagnose problems.
When a project feels stuck, ask: where is the entropy? Communication entropy (signal degradation between teams)? Process entropy (too many steps between idea and execution)? Product entropy (codebase has become unmanageable)? Personal entropy (your own systems have degraded)? The diagnosis determines the fix. More process does not solve communication entropy. Pruning does not solve a problem that actually needs a new process. Entropy thinking gives you a vocabulary for distinguishing between different types of decay.
It also recalibrates your expectations. If you understand entropy, you stop being surprised when systems degrade. You stop blaming people for natural drift. The maintenance is not a sign that the system is broken. The maintenance is the system working as designed. A refrigerator that needs electricity to keep running is not defective. It is doing exactly what physics requires.
Entropy is not your enemy. Ignoring entropy is your enemy. Every system you build, whether it is a company, a product, a team, or a daily routine, will drift toward disorder unless you invest continuous energy to maintain it. That is not pessimism. That is physics. The second law does not care about your intentions, your culture deck, or your project management tool. It cares about whether you are running an open system (with energy flowing in through communication, pruning, feedback, and simplification) or a closed one (coasting on momentum and hoping structure maintains itself). The startup that was lean at ten and drowning at fifty did not fail because someone made a bad call. It failed because nobody was doing the thermodynamic work. Run the entropy audit. Prune what no longer serves. Invest energy where signal is degrading. And accept that this work never finishes, because entropy never stops. That acceptance is not a burden. It is the beginning of building things that last.
