6. The Discontents of Web 2.0: Centralization, Surveillance Capitalism, and the Data Question

The previous chapter described how the "read–write" web of the 2000s turned the passive reader into a participant, and how a handful of companies learned to convert that participation into extraordinary value (see Chapter 5, Web 2.0: The Read–Write Era and the Platform Economy). This chapter examines the bill that came due. It sets out, rigorously and even-handedly, the structural critique of the platform web: the concentration of power in a few intermediaries, the business model that the scholar Shoshana Zuboff named surveillance capitalism, the unresolved question of who owns personal data, the politics of lock-in and deplatforming, the economics of the attention economy and algorithmic curation, the rents extracted from the creators and developers who build atop the platforms, and the fragility that comes from concentrating the world's communication in a few systems. These grievances matter to this book because, taken together, they constitute the motivating problem that Web3 claims to solve. This chapter is therefore the hinge between the history (Part II) and the proposal (Parts III–V): it does not yet evaluate whether Web3's answer works—that is the burden of later chapters—but it establishes, fairly, what the question is.

From an Architecture of Participation to a Concentration of Power

The defining promise of Web 2.0 was participation. In his canonical 2005 essay, the publisher Tim O'Reilly described the new web as software that "gets better the more people use it," harnessing collective intelligence through what he called an "architecture of participation" (O'Reilly, 2005). The same essay, however, contained the seed of the later critique. Among O'Reilly's design patterns was the maxim "Data is the Next Intel Inside": the claim that control over unique, hard-to-recreate data—a city's worth of map corrections, a billion social connections, the world's clickstream—would become the decisive competitive advantage of the era (O'Reilly, 2005). Participation produced data; data produced advantage; advantage produced concentration.

The mechanism of that concentration is the network effect.

A network effect is the property by which a service becomes more valuable to each user as more people use it. The classic analogy is the telephone: a single telephone is useless; a telephone is valuable in proportion to how many other people you can call. A social network with all your friends on it is, for that reason, very hard to leave for an empty rival.

In plain terms: the more people who join a platform, the better it works and the harder it is for any competitor to lure them away—so the biggest platform tends to keep growing until it is nearly the only one.

Network effects are not a conspiracy; they are an economic gravity. But their consequence is that markets shaped by them tend toward a small number of dominant winners. The media scholar Robert W. Gehl gave this outcome a precise name: distributed centralization. Web 2.0, he argued, presents an architecture that appears decentralized and participatory—millions of users posting freely from millions of devices—while in fact "funneling" the value those users produce into a small number of central portals that own the servers, the data, and the terms (Gehl, 2011). The participatory surface and the centralized substrate are two sides of one design.

In plain terms: the web feels like a public square that anyone can speak in, but the square itself is privately owned. The crowd does the talking; the landlord keeps the rent.

This tension—between a participatory experience and a centralized structure of ownership and control—is the root from which every other discontent in this chapter grows.

Surveillance Capitalism: The Logic of Behavioral Surplus

The most influential scholarly account of what the dominant platforms actually do with participation is Shoshana Zuboff's The Age of Surveillance Capitalism (2019). Zuboff's central claim is that a new economic logic emerged, first and most fully at Google, that "unilaterally claims human experience as free raw material for translation into behavioral data" (Zuboff, 2019). The phrase repays careful unpacking, because it is the analytical core of the entire critique.

Surveillance capitalism, in Zuboff's definition, is an economic system that treats the record of human behavior—what you click, search, watch, say, and where you go—as free raw material to be harvested, processed, and sold. The analogy Zuboff draws is to the early industrialists who treated nature as a free input: just as a logging company once felled a forest it did not plant, the surveillance capitalist harvests behavior it did not create and was never asked to pay for.

The crucial concept is behavioral surplus. A platform needs some data about you to provide its service—your search query so it can return results, your address so it can deliver a map. Zuboff's argument is that surveillance capitalism captures data far beyond what the service requires, and claims that excess—the surplus—as proprietary property (Zuboff, 2019).

In plain terms: when you ask a map app for directions, it needs your destination to do its job. Surveillance capitalism is the practice of also keeping a record of everywhere you went, when, how often, and who you met there—data the app did not need to give you directions, but which is enormously valuable to sell.

That surplus is then fed into what Zuboff calls "machine intelligence" and fabricated into prediction products—forecasts of what you will do next, what you will buy, how you will vote, whether you will churn—which are sold to business customers in what she terms behavioral futures markets (Zuboff, 2019). The advertiser does not, on this account, buy your attention so much as buy a prediction about your future behavior. Zuboff's most provocative extension is that the logic does not stop at prediction: the most reliable way to predict behavior is to nudge it, so the system drifts from forecasting what people will do toward intervening to make them do it.

Zuboff is a critic, and the more aggressive parts of her thesis—particularly the claim that platforms systematically engineer behavior—are contested by other scholars and by the firms themselves; this chapter returns to that contest below. But the descriptive core, that the dominant platforms' revenue derives from the large-scale collection and monetization of behavioral data, is not seriously disputed by the companies, whose own financial filings show that advertising is the overwhelming majority of their revenue.

The critical-media scholar Christian Fuchs reached a structurally similar conclusion from a different tradition, arguing that Web 2.0 platforms function as engines that "store, process, analyze, and sell" vast volumes of personal and behavioral data (Fuchs, 2011). Fuchs emphasized a feature specific to the read–write web: the user is a prosumer, simultaneously a producer and consumer of content.

A prosumer is a user who produces the very content they consume. On a social network, the users write the posts, take the photos, and supply the comments that are the product other users come to see. The platform supplies the venue; the users supply the goods—unpaid.

For Fuchs, this is what makes platform surveillance economically distinctive: the users generate the content, the social connections, and the behavioral data, while the platform captures the resulting value (Fuchs, 2011). Web 2.0, in this reading, did not merely enable surveillance as a side effect; surveillance became the business model.

The Data Question: Who Owns Your Data?

Surveillance capitalism makes urgent a question that Web 1.0 never had to ask: who owns your data? The question is deceptively simple and legally tangled. When a user uploads a photograph, writes a post, builds a list of contacts, or simply moves through a platform leaving a trail of clicks, who holds the rights to the resulting record—the user who generated it, or the company whose servers store it?

In practice, under the standard terms of service of the platform era, the answer has overwhelmingly favored the platform. Users typically retain nominal ownership of content they create while granting the platform a broad, often irrevocable license to use it; the behavioral data—the clickstream, the social graph, the location history—is generally treated as the platform's own asset, the very "Intel Inside" that O'Reilly predicted would be the source of advantage (O'Reilly, 2005). The data that is arguably most intimate—the record of one's behavior—is the data over which the individual has the least control.

In plain terms: you may "own" the photo you posted, but the platform owns the far more valuable record of who looked at it, for how long, and what they did next. You generated that record simply by existing on the service, and you cannot take it with you.

This asymmetry has provoked a counter-movement aimed explicitly at returning data ownership to individuals—and, importantly, one that predates and is distinct from blockchain-based Web3. The most prominent effort comes from Tim Berners-Lee, the inventor of the Web, through the Solid project (a contraction of "Social Linked Data"), developed at MIT. Solid proposes that each person store their data in a personal data store called a pod, which the individual controls, granting and revoking applications' access at will (Wikipedia: Solid, n.d.).

A pod, in the Solid model, is a personal online data vault—like a safe-deposit box for your information that you, not a company, hold the key to. Applications must ask your permission to read or write to it, and you can revoke that permission at any time, the way you might cancel a houseguest's spare key.

In 2018 Berners-Lee took a sabbatical from MIT to found a commercial venture, Inrupt, to develop Solid, framing the goal as restoring "true data ownership" to users (TechCrunch, 2018). Significantly for this book's central task of disambiguation, Berners-Lee has deliberately distinguished his decentralization effort from blockchain "Web3," resisting that label for Solid (Fast Company, 2022). The data question, in other words, has generated more than one proposed answer; Web3 is one candidate among several, and Chapter 7 takes up the resulting terminological confusion in detail.

The Attention Economy and Algorithmic Curation

If behavioral data is the raw material of surveillance capitalism, attention is its scarce resource. Because the dominant platforms are funded by advertising, their revenue is a function of how much of users' attention they can capture and hold. This produces the dynamics of what is commonly called the attention economy.

The attention economy is the competition among media and platforms for a fixed and finite human resource: hours of attention in a day. Because attention cannot be manufactured—there are only twenty-four hours, and a person can attend to only so much at once—platforms funded by advertising are structurally driven to maximize the time and engagement they extract from each user.

The instrument of that extraction is algorithmic curation: the automated ranking and selection of what each user sees.

Algorithmic curation is the use of software to decide, for each individual user, which posts, videos, or recommendations to show and in what order. Rather than presenting content chronologically or by explicit choice, the platform's ranking system predicts what will keep a given user engaged and surfaces that. The analogy is a newspaper editor who rewrites the front page differently for every single reader, optimizing not for importance but for how long that reader will keep turning pages.

The concern raised by critics is that when the optimization target is engagement, the content that wins is not necessarily the content that is true, healthy, or socially valuable, but the content that is most arresting—often that which provokes strong emotion. This is the point at which the attention economy connects back to Zuboff's thesis: the same behavioral data that fuels prediction products also fuels the ranking systems, and the line between predicting what will hold attention and engineering content to hold it is precisely the line Zuboff argues surveillance capitalism crosses (Zuboff, 2019). It is worth marking that the strongest version of this claim—that algorithmic feeds systematically degrade public discourse or mental health—remains an active and unsettled empirical debate among researchers, not a settled finding; the critique is real and serious, but its magnitude is contested.

Lock-in, Deplatforming, and the Censorship Debate

Concentration of power has consequences beyond economics. When a small number of platforms become the de facto venues for commerce, speech, and social life, two further problems emerge: users become locked in, and platforms acquire the power to exclude.

Platform lock-in is the condition in which leaving a service is so costly—in lost data, lost social connections, or lost functionality—that users stay even when dissatisfied.

Lock-in describes a situation where switching away from a product is so painful that you remain despite preferring an alternative. The analogy is a phone number you have had for twenty years: even if a rival carrier is cheaper, the cost of telling everyone a new number keeps you put. On social platforms the equivalent cost is your entire social graph and history, which the platform does not let you transfer.

Lock-in is the network effect experienced from the inside: the same gravity that concentrates users also traps them.

The sharper edge of platform power is deplatforming—the removal of a user, group, or application from a service, up to and including removal from the underlying infrastructure that makes a service reachable at all. The episodes that crystallized this concern were concentrated in a remarkable few years. In August 2019, the infrastructure company Cloudflare terminated service to the forum 8chan after its CEO described the site as a "cesspool of hate" linked to multiple mass shootings (TechCrunch, 2019). In January 2021, in the aftermath of the U.S. Capitol riot, Twitter permanently suspended the account of a sitting U.S. president, citing "the risk of further incitement of violence" (Twitter Inc., 2021); within days, Amazon Web Services suspended the social network Parler from its hosting platform, taking the entire application offline because it could not quickly find another host (TechCrunch, 2021).

These cases illustrate that exclusion can occur at multiple layers. A platform can remove a user; an infrastructure provider can remove an entire platform. The deeper the layer at which exclusion happens, the fewer the alternatives and the more total the effect—Parler did not lose a user, it lost the ability to exist online.

Even-handedness is essential here, because deplatforming sits at the center of a genuine and unresolved values conflict, not a simple wrong. The same power that critics call censorship is what others call content moderation, and platforms exercise it for reasons many find compelling: to remove incitement to violence, child sexual abuse material, harassment, and coordinated disinformation. A web with no gatekeeping is not obviously better than one with too much. The critique that motivates Web3 is therefore not that moderation is illegitimate, but that the power to decide—unilaterally, opaquely, and with global reach—rests with a handful of private companies accountable to no electorate, whose decisions can be neither appealed nor, in many cases, even explained. Reasonable people disagree about whether these companies have used that power well or badly in specific cases; what is harder to dispute is the structural fact that the power is concentrated and largely unchecked.

In plain terms: when a few private companies control the main venues for public speech, every one of their moderation decisions becomes a high-stakes act of governance—but they are not governments, and there is usually no court of appeal. Critics object less to any single decision than to the concentration of the power itself.

Rent Extraction from Creators and Developers

A distinct strand of the critique concerns not surveillance but economics: the claim that dominant platforms extract excessive rents from the very creators and developers who supply their value. In economics, a rent is income earned from controlling a scarce resource—here, access to an audience or a market—rather than from producing something new.

The clearest example is the app-store commission. To reach users on the two dominant mobile operating systems, a developer must distribute through the platform's app store and, for many transactions, accept its standard commission of up to 30 percent. The dispute reached a landmark courtroom in Epic Games v. Apple, where the trial court found that Apple's maintenance of its 30 percent commission was "not a product of competition" but of its market power, even as the court accepted Apple's arguments that centralized app review provides security and privacy benefits (Epic Games v. Apple, 2021; Wikipedia: Epic Games v. Apple, n.d.). The case captures the structural complaint precisely: the platform sits as an unavoidable tollgate between creators and their customers, and sets the toll.

In plain terms: to sell to the people on a platform, you must pay the platform a large cut—not because it built your product, but because it controls the only road to your customers. Critics call this a toll on a road no one else is allowed to build.

A second form of extraction is the API shutdown. An API (application programming interface) is the doorway through which one piece of software is permitted to talk to another; platforms expose APIs to let outside developers build on top of them—third-party clients, tools, and analytics.

An API is a set of rules and access points that lets external software interact with a platform's data and features—like a service window at a building through which approved visitors can make requests. Whoever owns the building controls the window: they can widen it, narrow it, charge admission, or board it up.

Because the platform controls that window, developers who build businesses on it are exposed to sudden, unilateral changes. In 2023, Reddit—whose API had been free since 2008—announced pricing that one prominent third-party client, Apollo, calculated would cost it roughly twenty million dollars per year, forcing its shutdown and triggering a protest in which thousands of communities went dark (Wikipedia: Reddit API controversy, n.d.; NPR, 2023). Twitter restructured and sharply restricted its own developer API in the same period. Each case made the same point: a developer who builds atop a platform's API builds on rented land, and the landlord can change the terms—or evict—at will.

The grievance extends to ordinary creators. Musicians, writers, and video-makers who depend on a platform for their audience are subject to opaque, shifting payout formulas, demonetization decisions, and ranking changes over which they have no control and no recourse. The structural pattern is consistent across app stores, APIs, and creator payouts: those who generate the value sit downstream of an intermediary that captures a large share of it and dictates the terms unilaterally.

Single Points of Failure and Control

A final discontent is technical but its implications are social: extreme concentration creates single points of failure.

A single point of failure is a component whose breakdown brings down an entire system—like a bridge that is the only route to an island, so that if the bridge closes, the island is cut off. When critical services run on a few centralized systems, the failure of one can disable much of the web at once.

The vivid illustration came on 4 October 2021, when a configuration error during routine maintenance caused Facebook, Instagram, WhatsApp, Messenger, and Oculus to vanish from the internet simultaneously for roughly six hours—a service relied upon by billions of people worldwide (Wikipedia: 2021 Facebook outage, n.d.; Meta, 2021). In several regions of the world where WhatsApp is the primary channel for personal and business communication, ordinary life was disrupted by the failure of a single company's network. The same concentration that produces market power produces systemic fragility: when one company is the channel, one company's outage is everyone's outage.

The point generalizes beyond outages to control. The same centralized chokepoints that can fail by accident can also be used deliberately—to surveil, to block, to censor at the behest of a government, or to exclude an entire region. Critics argue that an architecture with single points of control is, by construction, an architecture vulnerable to capture, whether by error, by commercial incentive, or by state power. It is precisely this property—centralized chokepoints that are simultaneously single points of failure and single points of control—that Web3's proponents target most directly, and the proposed remedy of decentralization is examined in Part III.

Where the Critique Is Contested, and What the Platforms Delivered

Intellectual honesty requires that the critique just assembled be tested against its strongest counterarguments, because a one-sided account would misrepresent the field and weaken the book's later evaluation of Web3.

First, the platforms delivered, and continue to deliver, enormous and genuine value—much of it at zero monetary price to users. Search engines made the sum of recorded human knowledge instantly retrievable; mapping services replaced the paper atlas with real-time global navigation; social and messaging platforms reconnected diasporas and let small creators reach global audiences without a publisher's permission; video platforms democratized instruction and entertainment alike. The "free" services funded by advertising lowered barriers to information and communication for billions of people, including many who could not have afforded paid alternatives. Any assessment that treats Web 2.0 as purely extractive ignores the immense consumer surplus it created.

Second, the surveillance-capitalism thesis, while influential, is contested in its strong form. Critics of Zuboff argue that she overstates the novelty of behavioral data collection (commercial surveillance long predates Google), risks technological determinism, and underplays the agency of users and the role of regulation. The claim that platforms can reliably engineer behavior—as opposed to merely predicting and nudging it at the margins—is empirically uncertain; advertising has always claimed more influence than it can always demonstrate. These are reasons to hold the strongest versions of the thesis provisionally, even while accepting its descriptive core.

Third, the data-ownership framing itself is questioned. Some economists and legal scholars argue that "who owns your data" is the wrong question, because much valuable data is relational—a social graph belongs jointly to everyone in it—and because individual data, sold or controlled one person at a time, is nearly worthless; its value is aggregate. On this view, propertizing personal data might entrench inequality rather than remedy it, and the better remedies are regulatory (rights of access, deletion, and portability, as in data-protection law) rather than proprietary. This matters for later chapters, because Web3's pitch leans heavily on individual ownership as the solution.

Fourth, concentration has prompted real responses within the existing system: antitrust enforcement, comprehensive privacy regulation, mandated data portability, and the deplatforming debates themselves are evidence of a society contesting platform power through law and politics rather than through new architecture. Whether those tools are adequate is itself contested—but their existence complicates any claim that the platform web is an unchallengeable monolith that only a technological revolution can dislodge.

In plain terms: the platforms genuinely improved billions of lives, the harshest version of the critique is not fully proven, and there are non-technical ways to address platform power. None of this refutes the grievances—but it means the grievances are a serious problem to be weighed, not a settled verdict that automatically validates whatever proposes to replace them.

Synthesis: The Motivating Problem That Web3 Claims to Solve

This chapter has assembled, from verified scholarly and journalistic sources, the structural critique of the platform web. Its elements cohere into a single diagnosis. The participatory web of the 2000s, driven by network effects, concentrated power into a few intermediaries (Gehl, 2011). Those intermediaries built a business model—surveillance capitalism—that harvests behavioral data as free raw material and sells predictions about human behavior (Zuboff, 2019; Fuchs, 2011), leaving unresolved the question of who owns the resulting record of one's life (TechCrunch, 2018). The attention economy and algorithmic curation orient the system toward engagement rather than welfare (O'Reilly, 2005; Zuboff, 2019). Lock-in traps users; deplatforming gives private companies quasi-governmental power over speech (Twitter Inc., 2021; TechCrunch, 2021); rent extraction subordinates creators and developers to unaccountable tollgates (Epic Games v. Apple, 2021; NPR, 2023); and the resulting concentration produces systems that are at once single points of failure and single points of control (Meta, 2021). The chapter has also marked, honestly, where each of these claims is contested and what the platforms delivered in return.

The significance of this diagnosis for the book's central question is that it is the question to which Web3 is offered as the answer. Web3's proponents take precisely this list of grievances—centralization, surveillance, the data question, lock-in, deplatforming, rent extraction, single points of control—and propose to dissolve them by removing the trusted intermediary altogether, re-architecting the web so that users rather than platforms control identity, data, and value. That ambition is not incidental to Web3; it is its founding rationale. As the next chapter shows, when Gavin Wood coined the modern meaning of "Web3" in 2014, he framed it explicitly as a response to a world in which "entrusting our information to organisations in general is a fundamentally broken model" (Wood, 2014). The discontents catalogued here are the broken model he meant.

Whether Web3's proposed cure is sound, partial, or illusory—whether decentralization can be achieved and sustained, or whether the same network effects and human preferences for convenience simply recreate the old concentration in new form—is the question that the remainder of this book must answer. This chapter has done its narrower job: to establish, fairly and on the evidence, the problem that gives the Web3 project its purpose. We now turn, in Chapter 7, to the intellectual history of the term itself.

References cited in this chapter

Epic Games, Inc. v. Apple Inc., No. 4:20-cv-05640-YGR (N.D. Cal. Sept. 10, 2021). United States District Court for the Northern District of California. (See also Wikipedia entry below.)

Fast Company. (2022). Tim Berners-Lee is building the web's 'third layer.' Don't call it Web3. https://www.fastcompany.com/90807852/tim-berners-lee-inrupt-solid-pods

Fuchs, C. (2011). New Media, Web 2.0 and Surveillance. Sociology Compass, 5(2), 134–147. https://compass.onlinelibrary.wiley.com/doi/abs/10.1111/j.1751-9020.2010.00354.x

Gehl, R. W. (2011). Distributed Centralization: Web 2.0 as a Portal into Users' Lives. Lateral, 1. https://csalateral.org/issue/1/distributed-centralization-web-2-0-portal-users-gehl/

Meta (Engineering at Meta) / Janardhan, S. (2021, October 5). More details about the October 4 outage. https://engineering.fb.com/2021/10/05/networking-traffic/outage-details/

NPR / Allyn, B. (2023, June 12). Thousands of Reddit communities 'go dark' in protest of new developer fees. https://www.npr.org/2023/06/12/1181376050/reddit-communities-go-dark-protest-new-api-developer-fees

O'Reilly, T. (2005, September 30). What Is Web 2.0: Design Patterns and Business Models for the Next Generation of Software. O'Reilly Media. https://www.oreilly.com/pub/a/web2/archive/what-is-web-20.html

TechCrunch / Lomas, N. (2018, October 9). Tim Berners-Lee is on a mission to decentralize the web. https://techcrunch.com/2018/10/09/tim-berners-lee-is-on-a-mission-to-decentralize-the-web/

TechCrunch / Lunden, I. (2019, August 4). Cloudflare will stop service to 8chan, which CEO Matthew Prince describes as a 'cesspool of hate'. https://techcrunch.com/2019/08/04/cloudflare-will-stop-service-to-8chan-which-ceo-matthew-prince-describes-as-a-cesspool-of-hate/

TechCrunch / Hatmaker, T. (2021, January 11). Parler is officially offline after AWS suspension. https://techcrunch.com/2021/01/11/parler-is-officially-offline-after-aws-suspension/

Twitter Inc. (2021, January 8). Permanent suspension of @realDonaldTrump. Twitter Blog. https://blog.twitter.com/en_us/topics/company/2020/suspension

Wikipedia contributors. (n.d.). 2021 Facebook outage. Wikipedia. https://en.wikipedia.org/wiki/2021_Facebook_outage

Wikipedia contributors. (n.d.). Epic Games v. Apple. Wikipedia. https://en.wikipedia.org/wiki/Epic_Games_v._Apple

Wikipedia contributors. (n.d.). Reddit API controversy. Wikipedia. https://en.wikipedia.org/wiki/Reddit_API_controversy

Wikipedia contributors. (n.d.). Solid (web decentralization project). Wikipedia. https://en.wikipedia.org/wiki/Solid_(web_decentralization_project)

Wood, G. (2014, April 17). ÐApps: What Web 3.0 Looks Like. Insights into a Modern World. https://gavwood.com/dappsweb3.html

Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. PublicAffairs.