Why Platforms Rot Like Epidemics: The Hidden Reproduction Number of Enshittification

What if “getting worse” is not a moral failure, but a contagion?

Why do some platforms begin as elegant, useful tools and end as cluttered, manipulative wrecks that users tolerate only because escape is too painful? The easy answer is greed. The more useful answer is structure. A platform does not become miserable simply because someone wakes up and decides to be evil. It becomes miserable when the system around it makes extraction easier than service, and when the cost of leaving is higher than the cost of enduring abuse.

That sounds abstract until you notice how closely it resembles the logic of an epidemic. A virus spreads when its reproduction number stays above 1. If each infected person passes it to more than one other person, the outbreak grows. If that number drops below 1, the chain burns out. Now translate that into digital life: a degrading platform persists when its bad practices spread faster than users can coordinate a response, and when each new layer of lock in makes resistance harder. What looks like a product problem is often a transmission problem.

Enshittification is not merely decline. It is a self-reinforcing epidemic of control.

The deeper question is not why platforms get worse. It is why some environments let degradation reproduce so efficiently. Once you see that, the story changes. The fight is not just for better companies. It is for lowering the reproduction number of bad business models.

The three hosts of platform decay

A platform usually begins by courting users. It offers convenience, novelty, and often a genuine improvement over the incumbent. Facebook promised privacy and a cleaner feed. MySpace looked noisy and compromised. The first phase is not a lie in the narrow sense. It is an investment in trust. The company is buying time to become indispensable.

Then comes the second phase: lock in. The platform realizes that users are not really attached to the platform itself. They are attached to the people, groups, archives, routines, and workflows sitting inside it. Leaving is no longer a simple consumer choice. It becomes a coordination problem. If your friends are there, your community is there, your clients are there, your family chat is there, your local league schedule is there, then your exit is not an exit at all. It is an attempt to drag a whole social ecosystem across a moat.

This is where switching costs and collective action problems do the real work. A single user can leave a diner, a bank, or a shoe brand with relative ease. But leaving a social platform often means asking dozens or hundreds of other people to move too, even when each of them has their own reasons to stay. That is why platform dominance becomes sticky: not because people love the service, but because they cannot get enough other people to hate it together.

The third phase is extraction. Once users, publishers, advertisers, suppliers, or creators are locked in, the platform begins to claw back the surplus it once spent to win loyalty. Feeds fill with slop. Search quality deteriorates. Ads become more intrusive and less useful. Publishers lose traffic unless they give away more of their work. Advertisers pay more for less certainty. Everyone remains on the platform, but nobody is really being served.

This is the key insight: the platform does not need to make every participant miserable equally. It only needs to preserve a brittle equilibrium in which no one group can leave without hurting itself more than the platform hurts them. That is why decay can be profitable. A platform can maximize extraction while remaining just tolerable enough to prevent coordinated revolt.

Why the system is brittle, and why brittleness matters

A brittle equilibrium is one that looks stable right up until it breaks suddenly. It is held together by fear of mutual loss. Users stay because their friends are there. Publishers stay because the audience is there. Advertisers stay because the buyers are there. Each side tolerates the platform because abandoning it feels like jumping into a void.

But brittleness is not strength. It is fragility disguised as dominance. A scandal, a policy change, a new competitor, or even a shift in investor expectations can tip the whole arrangement. That is because the platform has spent years converting trust into dependency and dependency into resentment. The moment enough participants decide that hate now exceeds attachment, the story changes from growth to decline.

There is a useful way to think about this. In epidemiology, the reproduction number is not fixed by nature alone. It depends on behavior, immunity, contact patterns, and intervention. Likewise, the reproduction number of platform rot depends on policy choices, law, competition, and technical openness. If those conditions favor lock in, degradation spreads. If they favor exit, interoperability, and substitution, decay loses momentum.

That changes the diagnosis. We are not dealing with an inevitable law of nature. We are dealing with a policy shaped contagion.

When a system makes bad behavior profitable and escape illegal or impractical, decay becomes self replicating.

This is why the usual comforting story, that “all big systems get worse,” is only half true. Size alone is not the poison. The poison is size plus insulation. A large firm can remain decent if it remains vulnerable to real competition, enforceable rules, worker power, and user exit. Remove those constraints, and scale becomes a machine for extracting value faster than alternatives can emerge.

The secret weapon is not just competition. It is escape

Most debates about monopolies stop at market concentration. That is important, but incomplete. The deeper problem is not simply that one firm is large. It is that rivals cannot easily interoperate with the incumbent, and users cannot easily take their data, communities, or workflows with them.

This is why adversarial interoperability matters so much. Digital systems are unusual because they are built on general purpose computers, meaning they can run counter technologies. If a platform erects a wall, someone can, in principle, build a ladder. Ad blockers, privacy blockers, scraper based migration tools, alternative clients, printer firmware fixes, and protocol bridges all exploit the same fact: software can talk to software.

But the law has increasingly treated that possibility as a threat. Anti circumvention rules turned what had been a technical possibility into a legal liability. If you can no longer legally build the ladder, the wall becomes much taller in practice. The platform’s power no longer rests only on engineering or network effects. It rests on a legal regime that forbids escape.

This is where the article’s most important synthesis lies. Platform decay is not just about the internal incentives of the firm. It is about the political decision to outlaw the user’s off ramp. A platform that can trap you is a platform that can rot you slowly. A platform that can be bypassed must compete with the possibility of replacement.

Consider how different the web would feel if leaving a service did not mean losing your history, contacts, reputation, or ability to communicate. Migration paths are the political equivalent of fire exits. Nobody expects to use them constantly. Their presence changes everything.

Privacy, labor, and pricing are one system, not three separate scandals

It is tempting to treat privacy, labor exploitation, and price manipulation as unrelated problems. They are not. They are different faces of the same architecture: information asymmetry under conditions of lock in.

A nurse hired through a staffing app can be priced according to debt profile. A breakfast burrito can be repriced based on payday timing. A publisher can be squeezed into giving away more content because traffic depends on a platform’s recommendation feed. A social feed can be reshaped to show more ads and less of what people wanted. These are not isolated abuses. They are examples of what happens when a platform can observe, segment, and exploit users with little fear of defection.

That is why privacy is a keystone issue. Without privacy, downstream harms multiply. Without enforcement, the surveillance economy becomes a general purpose extraction machine. And when enforcement is weakened by regulatory capture or forum shopping, the machine keeps running.

The same dynamic shows up in trade and statecraft. If your critical infrastructure depends on a foreign vendor that can cut you off, censor you, or reorder your incentives, then your vulnerability is political, not merely technical. That is why digital sovereignty is not a nationalist slogan. It is a demand for resilience. But resilience cannot be achieved by building shiny local substitutes alone. There must be a way to move from the old to the new without asking permission from the old gatekeeper.

That is the hidden lesson of migration path design. A new stack without migration is just another island. A new stack with migration is an escape route.

What actually lowers the reproduction number of enshittification?

If platform decay behaves like a contagion, then the goal is not abstract outrage. The goal is intervention at the transmission points. The most effective interventions are boring, practical, and structural.

Think of them as the four conditions that keep bad platform dynamics from reproducing:

1. Competition: users and business partners need credible alternatives.
2. Regulation: bad behavior must carry consequences that are not optional.
3. Worker power: the people inside the firm must have some ability to resist harmful incentives.
4. Interoperability: exit must be technically and legally possible.

Remove any one of these, and the reproduction number of platform rot rises. Remove all four, and decay becomes almost guaranteed.

This is why “move fast and break things” is not automatically a vice. It depends on what is being broken. Breaking a dominant platform’s wall can restore competition. Breaking ordinary people’s housing supply or civic infrastructure is another matter entirely. The moral axis is not speed. It is power direction.

A good rule of thumb: if breaking the thing increases user choice, it may be liberation; if it increases dependency, it is probably predation.

Key Takeaways

• Treat platform decay as a systems problem, not a personality problem. Bad managers matter, but the bigger issue is a structure that rewards extraction and punishes restraint.
• Look for lock in, not just monopoly. The most durable power comes from controlling the user, the supplier, and the buyer at the same time.
• Defend interoperability as a civil liberty. If people can build bridges, scrapers, adapters, and alternative clients, platforms must compete instead of coast.
• Prioritize migration paths. New systems fail when they demand a complete reset. They succeed when users can leave without losing their lives, contacts, or work.
• Treat privacy enforcement as anti extraction policy. Privacy is not a niche concern. It is the brake on surveillance pricing, labor discrimination, and behavioral manipulation.

The real opposite of enshittification is not goodness. It is permeability

We often talk as if the cure for bad platforms is better ethics, better founders, or better vibes. That is too soft and too vague. The real opposite of enshittification is not virtue. It is permeability: the ability of people, data, tools, and institutions to move across boundaries without asking permission from the very systems trying to trap them.

A permeable market cannot so easily become a hostage situation. A permeable network cannot so easily harden into a toll road. A permeable institution can replace bad tools without losing itself. That is why the fight is not merely against monopoly, or against surveillance, or against bad product design. It is against the legal and technical walls that let each of those forces feed the others.

If you want a single reframing to keep, it is this: platforms do not usually become terrible all at once. They become less escapable. And once escape is hard enough, quality stops mattering.

That is the hidden reproduction number of enshittification. The question for the next decade is not whether the internet can be made better in the abstract. It is whether we can make bad systems fail to reproduce themselves faster than they can rot the spaces we depend on.