Case Study 2: The Subscription Alternative

Can Platforms Profit Without Exploiting Attention?

Introduction: The Question the Critics Avoid

Critics of the attention economy — this book included — spend considerable energy describing the harms of advertising-supported platforms and the structural incentives that produce them. What they often underanalyze is the specific alternative. It is easy to say that platforms should serve users rather than advertisers. It is harder to describe exactly how a different model would work, what it would sacrifice commercially, and whether it is viable at the scale of the dominant platforms.

This case study examines three platforms — Substack, Patreon, and Wikipedia — that have built sustainable operations without advertising-based revenue. They represent three distinct variants of the subscription or donation alternative. Each has real commercial viability. Each has real limitations. Together, they illuminate both the genuine promise and the genuine constraints of building platforms that don't require exploiting user attention.

The evidence is clear enough to support a modest, qualified conclusion: ethical monetization is commercially viable, but not at unlimited scale, and not as a drop-in replacement for the specific social media use cases that the dominant platforms have colonized. Understanding why tells us something important about what change is actually possible.

Substack: The Direct Subscription Model

Origins and Mechanics

Substack launched in 2017 as a tool to make it easy for writers to charge readers directly. Its mechanics are simple: a writer creates a newsletter, sets a monthly or annual subscription price, and Substack takes 10% of revenue. Readers pay directly to writers. There is no advertiser in the transaction.

By 2024, Substack hosted tens of thousands of paid newsletters across every imaginable topic: political commentary, local journalism, science writing, literary fiction, sports analysis, financial research, personal essays. The platform reported approximately 35 million active subscribers collectively paying into this ecosystem. Several individual Substack writers — Glenn Greenwald, Heather Cox Richardson, Matt Taibbi at peak, later others — have built subscriber bases exceeding 100,000 paying readers, generating annual revenues in the range of $5–10 million per year from subscriptions alone.

Incentive Alignment

The Substack model creates a dramatically different incentive structure from advertising. A writer succeeds by retaining subscribers. Subscribers renew when they feel they are getting value from the writing — when it informs, challenges, entertains, or connects them to something they care about. Subscribers cancel when they feel they are not getting value, or when they feel manipulated, misled, or simply bored.

This incentive structure does not produce a frictionless alignment of writer and reader interests — no model does — but it produces significantly better alignment than advertising. An advertising-supported writer (a YouTuber, a blog monetized through display ads) earns more money by producing content that is watched for longer, that generates more reactions, that drives more clicks. A Substack writer earns more money by producing content subscribers are willing to keep paying for.

The distinction matters most at the margin where engagement and wellbeing diverge. Outrage-bait content performs well in advertising ecosystems; it performs less well in subscription ecosystems, because readers who feel manipulated by outrage-bait are more likely to cancel than readers who feel genuinely informed. The subscription model creates a direct financial feedback loop between reader satisfaction and writer revenue that advertising models simply do not have.

Limitations and Tensions

Substack's limitations are real and worth examining honestly.

Scale and discovery. The model works well for writers who already have audiences. Building an audience from scratch on Substack is difficult because the platform has limited discovery infrastructure — no recommendation algorithm, no viral distribution, no mechanism for surfacing writers to potential subscribers who don't already know to look for them. New writers must build audiences through other channels (often the advertising-supported platforms they are presumably trying to escape), then migrate them to Substack. The subscription model depends, to a significant degree, on the parasitic relationship with the very advertising ecosystem it is ostensibly an alternative to.

Economic concentration. Substack's revenue is highly concentrated among a small number of top writers. The platform's economics follow a power law: the top 10% of writers capture something close to 90% of revenue. The majority of Substack writers earn little or nothing from subscriptions. This is not unique to Substack — it is characteristic of any creator economy — but it means that the subscription model's promise of sustainable income is realized only by a small minority of practitioners.

Content quality and misinformation. The subscription model creates stronger incentives for accuracy than advertising, but it does not eliminate incentives for provocation. A subscriber who is emotionally invested in a particular political or ideological position will pay for writing that reinforces those positions, even when that writing is misleading. Substack has faced significant criticism for hosting writers who produce misinformation and conspiracy content — and has been slower to remove them than advertising-supported platforms, partly because the absence of advertiser brand-safety concerns removes one of the main commercial pressures to moderate content.

Patreon: The Creator Support Model

Origins and Mechanics

Patreon launched in 2013, three years after its founder Jack Conte began posting music videos online and found that advertising revenue was insufficient to support the work. The insight was specific: advertising revenue scales with views, but creating high-quality content requires sustained investment regardless of view count in any given period. A subscription model could provide stable income proportional to the size and loyalty of an audience rather than the volume of their consumption.

Patreon's model differs slightly from Substack's: creators offer tiered membership levels with different benefits at each tier (early access, behind-the-scenes content, direct access to the creator, physical merchandise), and patrons pay monthly for the tier they choose. Patreon takes between 5% and 12% of creator revenue depending on the service level.

By 2024, Patreon reported over 250,000 active creators and approximately 8 million paying patrons, with creators collectively earning over $2 billion annually through the platform. Like Substack, the distribution is highly unequal: top creators earn millions annually while the majority earn supplemental income.

The Wellbeing Alignment Problem (and Its Limits)

Patreon provides an instructive illustration of where subscription models genuinely differ from advertising models — and where they don't.

On the positive side: Patreon creators are insulated from the brand-safety pressure that pushes advertising-supported creators toward safe, high-engagement content formats. A Patreon-supported documentary filmmaker can produce a three-hour film on a difficult topic without worrying about advertiser flight from controversy. A Patreon-supported journalist can publish an 8,000-word investigation without optimizing it for shareability. The subscription revenue provides stable income that is not contingent on producing the specific content types that advertising algorithms reward.

On the less positive side: Patreon creators are not immune from their own versions of engagement optimization. The creator who communicates regularly with patrons, who posts frequently enough to justify the subscription, who provides the emotional intimacy and parasocial connection that high-tier patrons are paying for — this creator performs well on the platform. There is a different kind of engagement extraction happening here: not the anxious, infinite-scroll engagement of advertising platforms, but the emotional dependency and parasocial relationship that creators and patrons can develop. This is not the same problem as outrage-based engagement maximization, but it is worth noting that no business model is entirely free of incentives to keep customers psychologically invested.

What Patreon Solves and Doesn't Solve

Patreon solves the specific problem of content creators who produce high-quality work that is poorly compensated by advertising models. It provides those creators with a direct revenue source tied to audience loyalty rather than audience size. This is genuinely valuable and represents a meaningful improvement over advertising for the specific use case of individual content creator monetization.

What Patreon does not solve is the social media platform problem. Patreon is a creator-monetization tool, not a social platform. It provides no discovery infrastructure, no social graph, no mechanisms for the kinds of social sharing and community formation that define the dominant platforms' value proposition. The question of whether Facebook, Instagram, or TikTok could be replaced by something Patreon-like is not answered by Patreon's success as a creator tool.

Wikipedia: The Donation-Based Cooperative

The Model That Shouldn't Work

By any conventional analysis, Wikipedia should not work. Its content is produced entirely by unpaid volunteers. Its governance is managed through an elaborate democratic consensus process. It has no advertising, no subscription fees, no venture capital, and no profit motive. It competes in the information space with Google, which has billions of dollars in engineering resources, and yet Wikipedia consistently ranks among the most-visited and most-trusted information sources in the world.

Wikipedia is funded by user donations — approximately $150–160 million annually — and operates under the Wikimedia Foundation, a nonprofit. The foundation uses donations to fund the technical infrastructure: servers, bandwidth, a relatively small paid staff for legal, technical, and organizational functions. The content — 60 million articles in 333 languages — is created and maintained by a global community of approximately 280,000 active volunteer editors.

What Wikipedia Achieves

Wikipedia's most important achievement for the purposes of this discussion is negative: it has no engagement optimization algorithm. There is no mechanism within Wikipedia designed to maximize the time users spend on the site or to provoke emotional reactions that drive return visits. The goal of every Wikipedia article, as defined by the community's editorial standards, is to be accurate, comprehensive, and neutrally written. These are not engagement-maximizing properties — they are the opposite.

The result is a platform that serves users efficiently. Wikipedia's goal is to answer your question and let you go. The successful visit to Wikipedia is a short one: you find what you were looking for, you read it, you leave. By the metrics of the attention economy, this is a failure. By any reasonable measure of whether a platform is serving its users, it is a success.

Wikipedia also demonstrates that large-scale content platforms can be governed by community consensus rather than by corporate hierarchy. The editorial community has developed elaborate standards for sourcing, neutrality, and conflict resolution. These standards are imperfect and contested — but they exist, they are transparent, and they are enforced by the community rather than by an algorithmic optimization target that serves an advertiser.

Wikipedia's Structural Problems

Wikipedia's structural problems are worth examining because they reveal the genuine tradeoffs of the cooperative model.

Demographic skew. Wikipedia's editor community is approximately 85–90% male, disproportionately from North America and Europe, and disproportionately represents certain educational and socioeconomic backgrounds. This skew is reflected in the content: articles on topics important to underrepresented groups are systematically shorter, less well-sourced, and more often marked as needing improvement than articles on topics important to the platform's dominant editor demographics.

Governance slowness and hostility. Wikipedia's consensus-based governance process is famously slow and, for new contributors, often hostile. Articles are reverted by established editors who have built up institutional power within the community. The bureaucratic complexity of Wikipedia's policies creates significant barriers for new contributors, producing gradual demographic stagnation in the editor base and difficulties incorporating new knowledge efficiently.

The maintenance problem. Wikipedia depends on volunteer labor to maintain and update its content. When a topic attracts many volunteer editors — major public figures, popular cultural phenomena, breaking news events — the content is typically excellent. When a topic is important but attracts few volunteer editors — regional politics in developing countries, highly technical scientific fields, minority languages — the content quality degrades. The cooperative model produces uneven quality in ways that commercial models, with their ability to direct paid labor toward specific needs, do not.

What the Three Models Tell Us

Examining Substack, Patreon, and Wikipedia together, we can draw several conclusions about the commercial viability and limitations of the subscription alternative.

Ethical monetization is commercially viable, but at a different scale ceiling. All three models have demonstrated real commercial sustainability. Substack processes hundreds of millions of dollars in annual subscription revenue. Patreon has distributed over $2 billion to creators. Wikipedia sustains itself on $150 million in donations annually. None of these platforms has a revenue profile anywhere near Google's ($280 billion in 2023) or Meta's ($135 billion in 2023), and it would be dishonest to suggest the alternatives are equivalent at scale. But "commercially viable" does not require "as large as the current market leaders." It requires producing enough revenue to fund the operations and deliver the value the platform promises.

The social media use case is the hardest to replace. The subscription model works cleanest for individual creator monetization (Substack, Patreon) and reference content (Wikipedia). The specific use case of bilateral social interaction — sharing your life with your friends, discovering new social connections, building community around shared events — is the hardest to serve with a subscription model, because users have relatively little willingness to pay for the privilege of sharing their own lives. Social media's fundamental value proposition (stay connected with people you know) is currently bundled with advertising in ways that make it hard to unbundle.

Discovery and network effects remain the hardest problems. Every subscription alternative depends, at some point, on the advertising-supported ecosystem for audience discovery and user acquisition. Writers build Substack audiences by going viral on Twitter. Patreon creators build patron bases through YouTube or Instagram followings. Even Wikipedia depends on Google for the majority of its traffic. The advertising ecosystem's control of discovery infrastructure creates a dependency that subscription alternatives have not solved.

The models are complements, not replacements. The most accurate assessment is that subscription models, cooperative models, and contextual advertising represent better options for specific use cases — individual creator monetization, reference information, intent-based search — but do not yet represent viable replacements for the specific social networking functions that the dominant advertising-supported platforms provide. The question of whether those functions can be reorganized around different business models is one of the most important unanswered questions in the technology policy debate.

Conclusion: Viable, Valuable, and Incomplete

The evidence from Substack, Patreon, and Wikipedia supports a conclusion that is simultaneously encouraging and sobering. These models are real, they work, they provide genuine social value, and they do so with meaningfully better alignment between platform incentives and user interests than advertising-supported alternatives. They are not theoretical ideals; they are working systems that have served millions of users for years.

But they have not replaced the dominant platforms. They have not demonstrated the ability to scale to Facebook's two billion users or YouTube's five hundred hours of video uploaded per minute. They have not solved the social networking use case that currently anchors the dominant platforms' user lock-in. They have not broken the dependency on the advertising ecosystem's discovery infrastructure.

What they have done is establish existence proofs: it is possible to build sustainable, large-scale information platforms without exploiting user attention. The subscription and cooperative models describe a different way of organizing the relationship between platforms and users — one where the platform succeeds by serving users rather than by selling users. That the dominant platforms have not adopted these models tells us not that the models are unworkable, but that the structural conditions — competitive markets, capital expectations, talent allocation, regulatory vacuum — make adoption commercially dangerous even when executives might prefer to pursue it.

Changing those structural conditions is the task. The alternative models show us what we're trying to get to. Understanding the structural conditions that prevent the current market leaders from adopting them tells us what we need to change to get there.

Discussion Questions:

1. Substack has been criticized for hosting writers who produce misinformation, partly because the absence of advertiser brand-safety pressure removes one commercial incentive to moderate content. Does this represent a fundamental tradeoff between the subscription model and content quality, or is it a solvable problem within the subscription model?

2. Wikipedia's cooperative governance model produces systematic demographic skew in its editor community, which is reflected in content quality disparities. Is this a problem with the cooperative model specifically, or a problem with any governance structure that relies on voluntary participation? How might it be addressed?

3. If you were designing a new social networking platform intended to serve the "staying connected with friends" use case without advertising, what business model would you propose? What specific features would your platform need? What would users have to give up compared to the current dominant platforms?