Right now, it is impossible to imagine capitalism functioning without the century-and-a-half-old industry of advertising. It is not just a part of the so-called superstructure but deeply nested, close to capitalism’s beating heart. Despite tech companies’ frequent claims to innovation and disruption — and their scrambling attempts to find different revenue streams — many still depend on it or find themselves having to adopt it as a business model. 

To understand what is at stake for tech companies, it is key to bear that in mind. Elon Musk’s recent disputes with Twitter about bots and whatnot don’t so much reflect some novel concern specific to how social media work but an age-old concern about advertising: When he asked in a tweet, “So how do advertisers know what they’re getting for their money?” he was rehashing the famous complaint, often attributed to department store magnate John Wanamaker, that half of any advertising budget is wasted, but the problem is knowing which half. 

Companies primarily use content to produce segmented groups of people

The underlying concern in that question seems to be how, or even whether, advertising performance can be measured. But a larger question is to think of what media companies like Twitter make and sell — that is, to think of them not as information circulators but as commodity producers. In a 1977 article, “Communications: Blindspot of Western Marxism” — which eventually developed into “The Audience Commodity and Its Work” (1981) — communications scholar Dallas Smythe offered the idea of the “audience commodity” as a corrective to media studies that focused on content analysis rather than media as means of production. From Smythe’s perspective, the content media companies transmit is epiphenomenal to their primary business, which is to manufacture audiences (of human viewers) to deliver to advertisers. What matters to understanding the media and advertising industries is not so much whether ads “work” or have some other particular social effect in themselves (the focus of much 20th century critical theory about advertising) but how those audiences are made and sold: What are the means for packaging, measuring, and valuing them for exchange? 

Smythe’s argument draws on the theory of “monopoly capitalism,” which extends Marx’s ideas to 20th century developments. If the pre-1870s form of capitalism was driven primarily by competition among smaller producers, decades of economic development eventually led to consolidation and massive multinational firms, which, in Smythe’s view, reproduced their dominance not by meeting a population’s existing or given needs but by manufacturing demand for goods that the monopolies themselves were optimized to produce. This helps explain why so much of waking life has been colonized by advertising; it is integral to the sustained supremacy of the companies that have already come out on top. Advertising not only provides an arena in which the dominant manufacturers of essentially interchangeable products can compete as brands, it also helps establish the idea that people’s needs can be properly fulfilled only through those commodities. 

But for advertisers to manufacture demand, they need a very specific material as an input: audiences. That is where the for-profit media steps in. In the words of Richard Serra and Carlota Fay Schoolman’s short video about this process (produced a few years before Smythe’s intervention): “Television delivers people” — as do newspapers, and radio stations, and websites, and so on. That is, media companies don’t make content as their primary product; they use content to produce segmented groups of people that cohere for a definite stretch of time in a form that can be analyzed, sliced, and offered to a variety of buyers — but especially the dominant monopolistic firms that control much of the economy. Hence, the audience commodity, a term that marks the contested site where media companies, advertisers, and the ratings companies that intercede between them struggle over how attention can be standardized, measured, and sold. 


The “audience commodity” allows us to frame more useful questions about how contemporary digital media and tech companies operate, especially those whose primary revenue comes from ad sales. Research inspired by the concept — for example, Eileen Meehan’s work on gendering the audience commodity — has proved to be among some of the most insightful because it situates media companies’ decision-making within capitalist production’s many contradictions and reveals how the logics at play in ad markets have a primary role in determining which content is produced at scale. 

But in post-internet media criticism, the audience-commodity approach has been less adopted and discussed than approaches that focus on “digital labor,” “surveillance capitalism,” or “techno-feudalism.” The “digital labor” approach, which developed out of Autonomist Marxism, to its credit made an early attempt to make sense of the internet from a political economic perspective. Theorists like Tiziana Terranova associated networked computing with a “tendency” in post-industrial capitalism that collapsed the distinction between production and consumption and intermixed leisure and labor at a fundamental level of the economy. From this vantage, tech users were seen more as producers than the product. The unwaged or “free labor” that users of digital tech had begun to provide, generating data and content for sites, became structural as tech companies found ways to capture this value. From this perspective, the free gift of knowledge and community that early internet users happily produced became the source of its own eventual negation in the form of giant tech monopolies and the platformized web.

Workers as much as consumers must be produced

Shoshana Zuboff’s concept of “surveillance capitalism” takes up the idea of digital capture in a different way, focusing not on the labor of internet users but the passive data collection and social manipulation of those massive platform companies that emerged. She argues that “authentic” human interaction is being alienated from its original owner, the user, and processed as raw material for a variety of underhanded things designed to predict and modify behavior. In this view, users are being exploited, but Zuboff’s critique treats this as an anomalous process, the nefarious doings of a rogue sector of capitalism rather than internal to capitalism’s logic in general. Google isn’t treated as an advertiser or even particularly interested in audiences as such; it is instead a machine for conducting an invasive mode of surveillance and dispossessing users of “their” data, conceived as their rightful property rather than something that techno-social conditions structure and produce. Users’ data and behavior become products that are taken from them, but users themselves are not seen as commoditized.

The techno-feudalism thesis — or digital rentierism, as it’s also known — evolved in part out of a dissatisfaction with the digital-labor thesis. Some, like Jakob Rigi and Robert Prey, claimed that the Marxist labor theory of value still holds in its classical form for tech companies, and free labor plays no part in their emergence. Others, like Jathan Sadowski and the scholars in this collection on global platform capitalism, insist that the massive revenues and valuations of tech companies can be explained only by their special status as digital marketplace owners, capable of exploiting a misclassified workforce, capturing data to sell to third parties, and extracting a toll from each side of every transaction they host. An extreme version of this line of argument — which Evgeny Morozov, in a recent essay, attributes to Yanis Varoufakis and Jodi Dean — posits that capitalism is in the process of giving way to a neo-feudal order that no longer depends on exploiting nominally free wage workers. Instead, techno-feudal “landlords” exploit humanity at a new scale without relying on the myth of an open labor market. People are not the product, but are serfs.

While each of these theories has compelling arguments in its favor, I come back to Smythe’s work because of his focus on media companies as capitalist enterprises in monopoly markets, and how this conditions their business decisions. Critical political economy asks that we return again and again to the question of what is being made and what is being sold, so a clear political economy of the internet must focus on what media companies can sell as specific commodities to the sort of customers that will allow them to compete. For many tech and media companies, as they scale, market pressures within the prevailing capitalist logic compels them to produce audiences for advertisers over and above whatever other business model they may have once sought to pursue.


In her account of Smythe’s theory, Meehan notes that “media industries were neither dream factories nor consciousness industries: they were hunter-gatherers of the audience.” This holds true of contemporary tech companies, even the ones not yet directly dependent on advertising. Netflix, for example, doesn’t currently air ads, but its stock valuation depends on subscriber growth, making it a “hunter gatherer” of subscribers at a global scale. As Roman Lobato has shown in Netflix Nations: The Geography of Digital Distribution, this requirement directs investment choices when it comes to content. Yet the subscriber-growth model is starting to show its limits. As this Bloomberg article details, Netflix plans to launch an ad-supported service by the end of 2022. 

The video-game industry, too, is becoming tangled up with advertising. Games sold as a stand-alone product or as a service are being out-earned by free-to-play games that shovel ads — including ones for their own in-game content — into players’ faces, guiding them toward battle passes, loot boxes, and avatar cosmetics. These games live and die on their user base and invest heavily in ad campaigns across media formats. In this way the data collected by platforms and games is then recycled to build their audiences, creating a series of nested commodity products. 

A recent deepening of this tendency is the merger agreement signed between Unity Technologies (a suite of production-oriented tools for digital games, animation, and virtual worlds) and ironSource (an app advertising and monetization company). Apps made in Unity are now pre-built into the production of audiences. 

Gig platforms also exist within a slightly altered framework of audience commodities: If platformization as a process is about lock-in effects and winner-take-all economies of scale, this can work only by way of manufacturing different “audiences” to comprise the sides of their multi-sided market. Workers as much as consumers must be produced, and dividing them against each other continues to be lucrative. Companies like Uber are constantly searching for a new “audience” of drivers to sign up for the service, sign leases, and provide rides. This search requires the kind of psychographic analysis and targeting that companies like the ratings provider Nielsen still do for media companies and is subject to political biases that cut across a variety of social inequalities. 

As Meehan showed, the way ratings companies articulate the audience commodity demographically allows advertisers and media companies to construct audiences on political as well as economic bases. The jostle between media, advertisers, and ratings companies yielded a system of cultural production organized around the search for “bona fide consumers”: those with the “income, access, and desire to loyally purchase brand names and to habitually make impulse purchases.” Typically, these were held to be men, ages 18 to 34. But as Meehan points out, women in the 1970s and ’80s were often the main buyers of most household items (even if they weren’t necessary the main earners), so it would have been rational to develop TV programming to make audience commodities of them. Yet chauvinism — a political choice — oriented the system toward men, who also mainly ran the TV industry. A highly profitable system of production still found room to sacrifice some profit to sexism. 

Something similar persists in how audiences are constructed today. The audience commodity approach allows us to see both the profit-maximizing and political-hierarchy-reproducing sides of this, showing how the media-advertising complex and its circulation of capital contributes to perpetuating social inequalities — between geographic regions, between genders, and between races. Despite the claims of digital platforms that their data collection is accurate, we should keep in mind that there’s strong interest in misrepresenting this data, which likely does not reflect the truth on the ground at all. After all, a capitalist is interested in their profits and position, not reality.

Smythe’s focus on the production of audiences offers a way forward for making sense of the confusing layers of how money circulates in the digital economy. It not only returns to the simple question of what is being sold and to who; it also highlights the most important dynamic of contemporary markets, that they are simultaneously monopolistic and competitive at the highest level. It also shows that the 20th century logics of capitalist media companies still play a considerable role in how the digital economy is structured. It helps explain “why everything is an ad network,” as more and more companies use technology to build advertising into their business model, creating a series of nested ad networks all trying to find their bona fide consumer.