Platform Enshittification: Doctorow's Framework for Why Things Get Worse

Cory Doctorow coined the term “enshittification” in 2023 to describe a pattern so consistent across digital platforms that it functions as a law of platform dynamics rather than a series of coincidences. The pattern is this: platforms start by being good to their users, then they abuse their users to be good to their business customers, and finally they abuse their business customers to claw back value for themselves. The trajectory is predictable. The destination is a degraded product that serves no one well except the platform’s shareholders — and even that advantage is temporary. For anyone building a livelihood, an audience, or an identity on rented platform infrastructure, this framework is not academic. It is a forecast of what your landlord will do next.

Doctorow’s thesis complements Zuboff’s surveillance capitalism framework in a specific and useful way. Zuboff explains why platforms extract behavioral surplus — because a behavioral futures market exists and is profitable. Doctorow explains how that extraction escalates over time — through a predictable lifecycle that degrades the platform for everyone who depends on it. Together, these two frameworks give the sovereign individual both the diagnosis and the prognosis.

Why This Matters for Sovereignty

The sovereignty case against platform dependence is not about privacy alone. It is about structural reliability. If you build your audience on a platform, publish your content through a platform, or conduct your commerce on a platform, you are subject to that platform’s lifecycle. The platform does not owe you stability. It owes its shareholders returns. When those interests diverge — and Doctorow’s framework predicts they will — the platform will optimize against you.

This is not cynicism. It is documented behavior, repeated across every major platform of the past fifteen years. The sovereign response is not to avoid platforms entirely — that would be impractical and, for many purposes, counterproductive. The sovereign response is to build on infrastructure you control before the platform you are renting degrades beneath you. Thoreau did not refuse to visit Concord. He made sure he had a cabin to return to.

How It Works: The Three Phases

Phase One: Good to Users

In Doctorow’s model, platforms begin by subsidizing the user experience. They offer genuine value — better search results, free content hosting, efficient marketplaces, frictionless communication — because they need to attract the critical mass of users that makes the platform valuable to business customers. During this phase, the platform’s interests and the user’s interests are aligned. The product works well because it has to.

Amazon in its early years exemplified this phase. The company offered lower prices than brick-and-mortar retail, extensive product selection, fast shipping, and a review system that genuinely helped consumers find quality products. The experience was good because Amazon needed users to attract sellers, and it needed sellers to attract more users. The platform was subsidizing both sides of the market to build the network effects that would later become its leverage.

Google Search in its first decade was another clear example. The algorithm prioritized the most relevant results because relevant results attracted users, and users attracted advertisers. The product’s quality was the growth engine. There was no tension between what was good for Google and what was good for the person searching.

Phase Two: Good to Business Customers at Users’ Expense

Once the platform has achieved sufficient scale and network lock-in, the interests begin to diverge. The platform starts extracting value from users and redirecting it to business customers — the advertisers, sellers, and commercial partners who pay the platform directly. The user experience degrades, but not enough to overcome the switching costs that keep users on the platform.

Google Search’s quality decline between 2020 and 2025 is documented in multiple independent analyses. Search results pages became dominated by ads — in some queries, the user must scroll past four or more sponsored results before reaching organic listings. The “People Also Ask” boxes, knowledge panels, and AI-generated summaries consume screen real estate that once went to direct links. Several academic studies and journalistic investigations between 2023 and 2025 measured declining result quality, with increased SEO spam and decreased diversity of sources in top results .

Facebook’s treatment of business pages tells the same story with precise numbers. In 2012, a business page’s organic reach — the percentage of its followers who would see a given post without paid promotion — was approximately 16 percent. By 2024, that number had fallen below 2 percent for most pages . Businesses that had spent years building audiences on Facebook discovered that access to those audiences now required payment. The platform had attracted businesses with free audience-building tools, then monetized the audience the businesses had built. The value transferred from the business to the platform.

Amazon Marketplace followed the same arc. The platform that once connected buyers with the best products at the best prices became a platform where sponsored product placements dominate search results, where seller fees approach 50 percent of the sale price when advertising, fulfillment, and referral fees are combined , and where the review system that once served consumers has been degraded by incentivized reviews and algorithmic manipulation. The marketplace still functions, but it functions primarily for Amazon’s benefit.

Phase Three: Good to Shareholders at Everyone’s Expense

In the final phase, the platform extracts value even from its business customers to maximize returns for shareholders. Advertising prices increase. API access that was once free becomes paid. Features that attracted business customers are degraded or removed. The platform has achieved sufficient dominance that neither users nor business customers can easily leave, and it exploits that lock-in to extract maximum value.

Doctorow argues that this phase is not sustainable indefinitely — eventually, the degradation becomes severe enough that users and business customers begin to leave, or regulators intervene, or a competitor offers a sufficiently better alternative. But the timeline can be long, and the damage to anyone who built on the platform can be severe before the correction occurs.

The mechanism that enables this lifecycle is what Doctorow calls “switching costs.” Your social graph lives on Facebook. Your review history lives on Amazon. Your content archive lives on YouTube. Your professional network lives on LinkedIn. These assets are hostages — not because the platform explicitly prevents you from leaving, but because leaving means abandoning the accumulated value you built on the platform’s terms. The higher the switching costs, the more degradation users will tolerate, and the further down the enshittification curve the platform can travel before facing consequences.

The Zuboff-Doctorow Synthesis

Zuboff and Doctorow are describing the same system from different angles, and the synthesis is more useful than either framework alone. Zuboff identifies the economic engine: behavioral surplus extraction feeds a prediction market. Doctorow identifies the lifecycle: platforms degrade predictably as they optimize extraction over time. Together, the frameworks explain both the mechanism and the trajectory.

The surveillance capitalism framework tells you why your data is being collected and sold. The enshittification framework tells you why the platform that seemed like a good deal five years ago is getting worse — and will continue to get worse. For the sovereign individual, the practical implication is clear: any platform whose business model depends on behavioral surplus extraction will follow the enshittification curve. The only question is where on the curve it currently sits.

This is not a reason to avoid platforms entirely. Platforms provide genuine value, especially in their early phases. The sovereign response is to use platforms instrumentally — extracting value while the value is good — while simultaneously building on infrastructure you control. Your own domain. Your own email list. Your own content archive. Your own payment processing. These are the assets that do not degrade when a platform optimizes against your interests.

The Proportional Response

The proportional response to enshittification is architectural, not emotional. You do not need to rage-quit platforms. You need to build redundancy.

Start with the asset inventory. What do you have on platforms that you could not easily replicate elsewhere? Your content. Your audience. Your reputation. Your transaction history. For each of these, ask: if this platform degraded significantly tomorrow — raised its prices, reduced your reach, changed its terms — what would you lose? The answer tells you where to invest in sovereignty.

For content creators and business owners, the highest-priority action is establishing an owned platform — a domain you control, with an email list you own, publishing on infrastructure that cannot be unilaterally altered by a third party. This does not mean abandoning platform distribution. It means ensuring that platform distribution feeds owned infrastructure rather than replacing it. Post on social media, but drive traffic to your own site. Build an audience on YouTube, but convert viewers to email subscribers. Use the marketplace, but build direct customer relationships in parallel.

For individuals who are primarily consumers rather than creators, the proportional response is simpler but still meaningful. Diversify your digital life so that no single platform’s degradation is catastrophic. Use multiple communication channels. Maintain local copies of content you value. Resist the convenience of single-platform lock-in where the cost of diversification is low.

The core principle is this: enshittification is predictable. It is not a betrayal — it is a business model reaching its logical conclusion. The sovereign individual does not express surprise when a platform follows the curve. They have already built the cabin.

What to Watch For

As of early 2026, several platforms are at different points on the enshittification curve, and monitoring their trajectory helps calibrate your response.

Google Search continues to degrade for organic content discovery, with AI-generated answers increasingly replacing links to source material. For anyone whose traffic depends on Google referrals, this is a Phase Two to Phase Three transition worth watching closely. The platform is extracting value from content creators (who provided the training data for AI answers) and redirecting it to the platform’s own products.

Social media platforms are consolidating around short-form video, which intensifies the engagement-optimization dynamics while making content less portable. A text post can be republished anywhere. A video optimized for TikTok’s algorithm is a platform-specific asset with limited value elsewhere.

Amazon’s increasing dominance of e-commerce search means that for many product categories, the enshittification of Amazon’s marketplace is not escapable by simply choosing a different marketplace — there is no equivalent alternative with comparable scale. This is the lock-in condition that enables Phase Three extraction.

The pattern holds. The timeline varies. The sovereign response is the same: build before you have to, because by the time the degradation is obvious, the switching costs are highest.

This article is part of the Surveillance Capitalism & The Proportional Response series at SovereignCML.

Related reading: The Business Model Is the Problem, The Attention Economy: How Your Focus Became a Commodity, The Five Things That Actually Matter