Enshittification Is Not a Bug in SaaS. It Is the Model.

In this article:

• What Incentive Structure Makes SaaS Decay Inevitable
• How Predictable Is the Enshittification Timeline for Any Platform
• What Questions Should You Ask Before Adopting Any SaaS Tool
• What Ownership Models Produce Software That Does Not Enshittify
• Why Is Enshittification the Correct Frame Not Product Criticism
• Frequently Asked Questions
• Is enshittification avoidable in venture-backed SaaS?
• How do you evaluate a SaaS tool before it enshittifies?
• What is the difference between enshittification and bad product management?
• References

Enshittification gets framed as a failure of execution. The wrong leadership made the wrong calls. Investors pushed too hard. A good product got ruined by people who did not understand what made it good. This framing is comfortable because it implies the problem is fixable: install better leaders, find patient capital, hire product managers who actually use the tools they build. None of that is the problem. Enshittification is not a deviation from the SaaS model. It is the SaaS model running correctly.

Cory Doctorow named the pattern in 2023 and described its sequence with enough precision that it should have settled the debate about whether platform decay is structural or incidental. It has not settled the debate, because the people most invested in the model have the most to lose from the structural explanation. But the sequence is not ambiguous: platforms first serve users to build scale, then abuse users to serve business customers, then abuse business customers to claw back value for shareholders and creditors. Each stage follows from the economic logic of the one before it. The decay is the business model reaching maturity.

Venture-backed SaaS has one viable exit: growth to acquisition or IPO at a multiple that returns the fund. Neither of those exits rewards long-term user value. Both reward growth metrics, net revenue retention, and market share. The product decisions that maximize those metrics are not the same decisions that maximize user value. When the two align, the platform appears to be good. When they diverge, the incentive structure wins and the platform decays. The alignment is temporary. The divergence is permanent.

Lock-in is not a side effect of building a useful product. It is an explicit design objective. Every platform decision that makes data harder to export, every proprietary format that replaces an open standard, every workflow that only functions inside the platform's ecosystem is a lock-in investment. The engineering team may not frame it this way. The incentive structure does not require them to. It simply underfunds the features that reduce switching costs and overfunds the features that increase retention metrics and justify the next pricing tier.

Data portability is not a feature the engineering team forgot. It is a switching cost the strategy team intentionally maintains. This distinction matters because it means the solution is not a better engineering team. It means the solution requires a different ownership structure with different incentives, or it requires the user to stop depending on the platform's goodwill to access their own data.

The question for any SaaS platform is not whether enshittification will happen. It is where the platform currently sits in the sequence. A new platform with genuine momentum and a growing user base is in phase one: serving users well enough to generate the lock-in that makes phase two profitable. The product is good because goodness in phase one produces the scale that enables extraction in phase three. The goodness is instrumental. It is not the goal.

A platform with established market share, rising prices, and a product roadmap that consistently prioritizes revenue-generating features over user-requested improvements is in phase two or three. The enshittification is not imminent. It is already running. The businesses still paying monthly fees and building workflows inside that platform are in the extraction phase whether or not they have named it.

The sequence does not reverse without a structural change in ownership or incentives. A new CEO does not reverse it. A public apology about pricing does not reverse it. A product team reshuffle does not reverse it. The incentive structure that produced the decay remains in place. The decay continues at whatever rate the market will absorb.

If enshittification is structural rather than incidental, then evaluating tools on current product quality misses the point. The product that is good today is good because it is in phase one of its enshittification arc. The lock-in it is building while it is good is the mechanism that will extract value from you in phase three. Evaluating a tool without accounting for its ownership structure, its funding model, and the switching costs it is accumulating while you use it is evaluating a landlord based on the first month's rent.

The relevant questions are not: does this tool solve my problem well? They are:

• What does this tool know about my workflows and clients that I cannot export if I cancel?
• How much of my operation will exist exclusively inside this system in twelve months?
• What happens to years of accumulated data if the platform raises prices beyond what I will pay?
• Who currently owns this platform and what exit do they need?

Adobe's forced migration from perpetual licenses to Creative Cloud subscriptions in 2013 is instructive. The features were the same features designers had used for years. The ownership relationship changed: from a one-time license that ran on your machine indefinitely to a subscription that ceased to function if you stopped paying. Adobe built enough lock-in through file format dependency and workflow integration that the migration, despite significant user resistance, succeeded. The product did not improve at the rate of the price increase. The switching cost was high enough that most users paid anyway.

The SaaS model cannot produce platforms that do not enshittify because the model's exit requirements are incompatible with sustained user value. This is not a criticism of individual founders or product teams. It is a description of what venture capital requires and what the public markets reward. A founder who wants to build a tool that stays good for a decade and charges a fair price for what it delivers is not building a venture-backed SaaS company. They are building something else: a lifestyle business, a cooperative, a nonprofit, a self-funded product, none of which fit the model and all of which have different incentive structures.

Yanis Varoufakis frames this as the replacement of market capitalism with cloud feudalism: a system where the returns flow not from producing things that compete in markets but from controlling the infrastructure through which all production flows and collecting rent from everyone who depends on it. The enshittification sequence is the mechanism by which a competitive product becomes a feudal infrastructure. Phase one produces the product. Phase two builds the dependency. Phase three extracts the rent. The model is complete.

Platform cooperatives offer one structural alternative: governance structures that distribute ownership among users rather than concentrating it in investors, removing the exit pressure that drives enshittification. Self-hosted software offers another: infrastructure you own cannot extract rent from you regardless of who owns the codebase. Neither alternative is as convenient as a well-funded SaaS product in phase one. That inconvenience is precisely what the SaaS model is selling you in exchange for the lock-in it is building while you enjoy it.

The correct frame for evaluating any platform is not: how good is this product today? The correct frame is: what is the ownership structure, what does the exit model require, and what will this platform look like when the interests of its investors and the interests of its users fully diverge? That divergence is not a risk. It is a scheduled event. The only variable is timing.

Recognizing enshittification as structural rather than incidental changes what you build, what you adopt, and what you are willing to become dependent on. It does not mean refusing all software. It means being precise about which infrastructure relationships you are willing to let accumulate switching costs and which ones you insist on keeping portable. The distinction between those two categories is the difference between owning your operational stack and farming someone else's platform while it is still in phase one.

The platforms that are good right now will not stay good. The sequence is documented, the incentives are in place, and the timeline is the only unknown. The businesses building irreplaceable workflows inside those platforms today are the extraction targets of phase three. That is not a prediction. It is a description of how the model works.

No. Venture-backed SaaS requires an exit that returns investor capital, either through acquisition or IPO. Neither exit rewards long-term user value. The incentive structure guarantees that product quality will be sacrificed for extraction metrics once lock-in is achieved. The timeline varies. The outcome does not.

Evaluate the ownership structure, not the current product quality. Ask: who owns this platform, what exit do they need, and what will they do when user value and investor return diverge? Also ask what data this tool accumulates that you cannot export. what interoperability standards tell you about a tool before you adopt it.

Bad product management produces a declining product by accident. Enshittification produces a declining product by structural necessity. The distinction matters because bad product management can be fixed with better leaders. Enshittification cannot be fixed without changing the ownership model and incentive structure that make extraction more profitable than quality.

Doctorow, Cory. Pluralistic: Tiktok's enshittification. January 2023.

Doctorow, Cory. Pluralistic. pluralistic.net.

Varoufakis, Yanis. Technofeudalism: What Killed Capitalism. Bodley Head, 2023.

Varoufakis, Yanis. yanisvaroufakis.eu.

Hern, Alex. "Adobe's switch to subscription model for Creative Suite sparks protest." The Guardian. May 2013.