The Great Unbundling: Creators, Platforms, and the Collapse of the Newsroom

The institution that used to capture the surplus

For most of the twentieth century, reaching a mass audience required owning physical infrastructure: a printing press, a broadcast license, a wire service subscription. Owning that infrastructure was expensive enough that only an organization could afford it, and the organization, not the individual reporter, became the unit that captured most of the money a story or a broadcast could generate. A journalist worked for a newsroom because the newsroom owned the means of reaching anyone at all, and that ownership was the actual asset being paid for.

Advertising followed the audience the newsroom assembled. A metro daily's classified section, a network's ratings share, a magazine's circulation count, each one functioned as an audience packaged for sale to an advertiser, and the institution that had built the audience kept most of what that sale produced once it paid the staff who reported, edited, and printed the thing readers actually wanted to read.

Platforms that host content without producing it break the link between owning distribution and employing the person who fills it. YouTube, Substack, and services like them supply hosting, discovery, and payment rails to anyone with an account, without employing the person whose video or newsletter runs through those pipes. The capital that once had to belong to an institution now belongs to the platform, and the content that fills it can be produced by one person working alone, with no newsroom in between.

The newsroom's ownership of distribution also carried an editorial function that is easy to overlook once distribution stops requiring an institution at all. An editor decided what ran, in what order, and under what standard of sourcing, and that decision doubled as a form of quality control the reader could rely on without checking it directly. A platform that hosts a creator's account performs no equivalent editorial check on what that account publishes; it decides only whether to rank, recommend, and pay for it. The unbundling separated the money from the newsroom and, with it, separated the reach decision from any editorial judgment at all.

What follows from that change is not a single story of decline. It is two stories running side by side: an institutional newsroom economy shrinking as its advertising base migrates, and an individually branded creator economy growing as some of the same advertising and subscription dollars land somewhere new. Both movements are now large enough, and disclosed in enough detail, to measure against each other directly rather than describe in the abstract.

Two economies, one advertising dollar

The institutional side of the story is documented in detail by Pew Research Center's analysis of Bureau of Labor Statistics data. U.S. newsroom employment fell 26 percent between 2008 and 2020, a loss of roughly 30,000 jobs, from about 114,000 newsroom employees to about 85,000. Underneath that headline figure sits a reallocation as much as a straight decline: newspapers, which held about 62 percent of all newsroom jobs in 2008, held fewer than 36 percent by 2020, while digital-native outlets' share rose from 6 percent to 21 percent over the same period. The newsroom did not simply empty out. Its jobs moved from print organizations toward newer, web-first ones, before the individual creator economy had entered the count at all.

The advertising dollar that used to fund those newsroom jobs collapsed on a similar timeline. U.S. newspaper advertising revenue fell from about $49.4 billion in 2005 to about $9.6 billion in 2020, a decline of roughly 81 percent. The pressure was not gradual or evenly spread across the calendar. Pew Research Center tracking found that roughly one in three large U.S. newspapers conducted layoffs in 2020 alone, the same pandemic year in which platform-based creator payouts were accelerating.

On the platform side, the same years produced very different numbers. YouTube says it has paid out more than $100 billion to creators, artists, and media companies since 2021, with chief executive Neal Mohan citing $70 billion of that total paid between 2021 and 2024 alone. Substack's growth over the same period was steeper still: paid subscriptions on the platform rose from about 2 million in 2023 to 4 million by November 2024 and past 5 million by March 2025, and total writer gross revenue across the platform rose from $300 million in 2023 to $370 million in 2024 to $450 million in 2025.

Substack the company, distinct from the writers earning on it, took a flat 10 percent commission on those subscriptions and reported annualized revenue of $45 million in July 2025, up from $37 million in 2024 and $30 million in 2023. More than 17,000 writers were earning money on the platform by that point, a floor low enough to include many earning very little, alongside the more than 50 who were earning over $1 million a year. The advertising and subscription dollars that used to flow into an institutional payroll are, by these figures, now flowing substantially toward individuals instead.

Concentration inside the unbundling

The creator economy's growth is real, but it is not evenly spread, and the unevenness echoes something the newsroom collapse also showed. The top ten Substack publications collectively earned more than $100 million annually by 2025, up from $25 million collectively in 2023, a fourfold increase in two years. That $100 million captured by ten publications is more than double the $45 million Substack the platform itself reported earning from its 10 percent commission on every paid subscription running through it. A small number of the platform's most successful writers are, in dollar terms, bigger businesses than the platform hosting them.

The more than 17,000 writers Substack counted as earning something by mid-2025 makes the concentration plain: most of that group earns far less than the fraction crossing $1 million a year, a pattern common to creator platforms generally and not unique to Substack. Unbundling the newsroom into individual accounts did not flatten the distribution of who gets paid. It relocated the concentration from an institution's masthead to a platform's leaderboard, and the leaderboard is steeper.

The institutional side concentrated its losses just as unevenly. Newsroom job cuts between 2008 and 2018 hit midcareer workers hardest: full-time newsroom employees aged 35 to 54 fell 42 percent, from about 52,900 to about 30,800. Younger and older workers, on Pew's tracking, absorbed less of the cut. The newsroom that shrank did not shrink evenly across its staff, and the digital-native share of the surviving jobs, rising from 6 percent to 21 percent, suggests some of that labor found a home in a newer kind of institution rather than in an individual creator account at all.

Read together, the two datasets argue against a clean story in either direction. The advertising dollar did not simply move from institutions to individuals; it moved from one kind of institution to a mix of newer institutions and individually branded accounts, and within that individual tier it landed disproportionately on a small number of creators, in much the way the institutional tier it left had disproportionately protected its newest and its most senior staff over its experienced middle.

The algorithm as editor

The economic reallocation carries a geopolitical one inside it. A nationally licensed broadcaster like the BBC operates under a charter granted by a government; a wire service like Reuters historically divided the world into territories its member papers would cover. Reach, in that older architecture, was a matter of law, license, and negotiated territory, decided in some part by national institutions carrying a public mandate or a commercial treaty behind them.

A creator's reach today is decided by neither. It is set by a platform's ranking and payout algorithm, the rules that determine which video gets recommended, which newsletter gets surfaced to new readers, and how much of the advertising or subscription pool a given account is paid. Those rules are proprietary, changeable without public notice, and set by a small number of companies headquartered in the United States. A creator's global audience, and the money it produces, both depend entirely on decisions an outside institution has never had to explain to anyone reading the result.

This does not mean influence has left institutions altogether. Newer digital-native newsrooms captured a growing share of a shrinking institutional pie, as the employment data above shows, and national broadcasters and wire agencies have not disappeared. What has changed is where the marginal unit of new cultural and political reach now sits. Increasingly it sits with an individual whose entire audience was assembled by a platform's ranking decision, rather than by a masthead, a license, or a territorial agreement with another publisher.

That relocation of reach is also a relocation of negotiating power. A government negotiating with a national broadcaster it charters, or a wire service negotiating a territory with another wire service, has an institutional counterpart to negotiate with. A platform's ranking algorithm has no comparable counterpart on the other side of the table, and the individual creator whose reach depends on it has even less standing to negotiate its rules than the institution the platform displaced.

What outlasts the institution

The specific mechanism driving this shift, individually branded video and newsletter accounts monetized directly by a platform, is new to this decade. The underlying question it raises is not. Every medium this series has covered settles, in its own way, the same question: who controls the infrastructure a mass audience uses, and what does that control let them decide about who gets paid and who gets heard. The newsroom answered that question for roughly a century and a half. Platform ranking and payout rules are answering it now, on a timeline measured in years rather than generations.

The chapter after this one is already visible in the same data. An AI system that assembles a shortlist of sources, businesses, or writers for a person asking a question performs a version of the same sorting a platform's recommendation algorithm performs for a creator's video or newsletter, deciding who gets surfaced and, by extension, who gets paid attention worth converting into money. The newsroom that shrank and the creator account that grew are both, at this point, subject to the same downstream test: whether the system doing that sorting can find them, read what they publish, and judge it worth naming.

That is a narrower claim than a prediction about where creator or newsroom revenue goes next, and it should be read as narrow. It is simply the observation that the pattern running through this series, control of a dominant medium determining who captures money and who reaches whom, has not stopped operating just because the medium changed from a printing press to a platform's payout algorithm to a machine assembling an answer. Each step changed who won. None of them changed the shape of the contest.

Neither the shrinking newsroom nor the growing creator account is, on its own, well positioned for that test. A newsroom built its authority on an editorial masthead and a beat reporter's standing, credentials a machine assembling an answer does not automatically read. A creator built an audience on a platform's recommendation feed, a signal that does not automatically transfer to a different system deciding whether to name a source. Both now face a version of the same unbundling that separated money from institutions in the first place: the entity that produces something worth reading is, once again, not necessarily the entity a new intermediary chooses to surface.

The limits of this reading

Some of what ties these figures together is inference rather than proof. The newsroom employment and advertising-revenue figures come from Pew Research Center's analysis of federal labor and industry data, a well-corroborated source for the institutional side of the story. The YouTube and Substack figures are platform-disclosed or tracked by a single analytics firm, Sacra, rather than independently audited, and should be read with that caveat. Both sets of figures are established in the sense that they are the best available public numbers, not in the sense that every underlying dollar has been externally verified by a third party.

The causal link between the two trends is weaker than the correlation. Newsroom advertising revenue collapsed for reasons well beyond platform payouts, including the broader migration of classified and display advertising to search and social platforms generally, a shift that predates and extends beyond YouTube and Substack specifically. That the creator economy grew while the newsroom shrank is well documented. That platform payouts specifically, rather than the wider advertising migration, caused the newsroom's decline is a plausible reading, not a demonstrated one, and is held here as contested.

The thread also cuts both ways, as every medium in this series has. The unbundling of the newsroom into individual accounts genuinely lowered the barrier to reaching an audience: anyone with a camera or a keyboard can now build a following that a printing press or a broadcast license once made impossible without institutional backing. The same unbundling concentrated a new kind of power in the platforms setting the ranking and payout rules, and within the creator tier itself, in the small number of accounts capturing most of the money the wider pool generates. Liberation and concentration, in this chapter as in the others, arrived together.