The Self-consuming Commodity: Audiences,Users,and the Riddle of Digital Labor

Inheriting the Blindspot

Smythe (1977) set off the “Blindspot debate” by arguing that Marxist media criticism had overlooked the economic function of mass communication as a system, separately from its ideological function. He then proposed to shed light on this “blindspot” by showing that, insofar as the mass media audience was a commodity, the source of its value for capitalism was audience labor power, which consisted in learning to become consumers (Smythe 1977, 4). Jhally and Livant (1986) elaborated on Smythe’s insight by proposing that this learning process consisted in the labor of watching, measured in watching-time. They also astutely observed that media companies are able to extract a measure of rent by monopolizing access to audiences (Jhally and Livant 1986, 125).

In response to these arguments, Meehan (1984) objected that the audience commodity was in fact the product of the ratings industry. Consumer consciousness could not assume the form of a product to be exchanged, and watching-time had to be ascertained. Ratings and additional audience research and measurement techniques emerged to provide a suitable proxy, allowing TV advertising prices to be negotiated in terms of projected audience size and composition. In effect, it was the work of measurement that constituted the blindspot in question. With this, the terms of the debate were established. On one hand, Marxist authors have been unable to agree as to what constitutes the labor component of the audience commodity. On the other hand, a competing contingent has followed Meehan in arguing alternatively that the audience commodity is artificial or fictitious (Chen 2003), that the source of value is not audience labor but rent (Pasquinelli 2009), or that the labor generating it takes place elsewhere, such as the culture industries (Maxwell 1991) or the activity of “meaning making” (Nixon 2014).

This state of affairs has remained essentially unchanged, as the digital labor debate emphatically confirms. As Fuchs and Sevignani (2013, 237) helpfully summarize the stakes of the prevailing state of play, the central problem is whether and in what sense

the dominant capital accumulation model of contemporary corporate Internet platforms is based on the exploitation of users’ unpaid labor, who engage in the creation of content and the use of . . . sites for fun and in these activities create value that is at the heart of profit generation.

Disagreement chiefly concerns the precise nature of this process, and critics have offered varying and often incompatible descriptions of it, grounding differing conceptions of what, exactly, today constitutes the “blindspot” of media analysis. Authors who, like Mosco (2011), continue to insist that “labor remains the blind spot of communication and cultural studies,” invariably rely on a series of inferences that lead back to the assumption that advertising can and does “create demand” in some unspecified but functionally irreducible sense. ¹ This assumption, which remains largely unquestioned ² and central to other areas of cultural and media studies, places audience (or user) consciousness and activity at the center of communicative capitalism generally and the political economy of social media specifically.

This, in turn, results in another blindspot: the role of advertisers as the consumers of user “labor.” Among the tiny minority to have noticed this role, Bolaño and Vieira (2015, 58) have assigned it the designation of “third payer.” But here their analysis of this role stops, and they go on to replicate Meehan’s critique of Smythe by concluding that “The productive labor in SNSs is precisely the work done by employees, engineers, researchers, and much other kind of professionals that produce statistics, interfaces, and algorithms that make possible the constitution of the commodity audience” (Bolaño and Vieira 2018, 58). In other words, if it is the advertiser who pays, the source of value in the SNS economy is simply labor conducted elsewhere, on behalf of the advertiser. This is, to be sure, a crucial step forward, as it identifies an important driver of advertising expenditures that has little if anything to do with expanding sales revenues. But it raises as many questions as it settles, including the inevitable one of how this particular form of labor is financed. Surely, the incomes of employees, engineers, and researchers are entirely funded by revenues from the sale of the goods and services produced by advertisers. And this, in turn, would seem to justify the semantic slippage involved in referring to users as consumers in the first place.

From this vantage, little conceptual clarity is gained by focusing on the labor of those employed on behalf of the “third payer”; this new focus again returns us to the problem of user labor. Such a conclusion, however, would be misleading and unhelpful. Part of the reason can be glimpsed in the commonsense empirical objection to Fuchs’s approach offered by Arvidsson and Colleoni (2012, 138), who note that “in 2010 (according to its own figures) . . . each Facebook user was a ‘victim of exploitation of surplus value’ to the extent of $0.7 a year,” a paltry sum. The authors point this out to ground their claim that the source of SNS value cannot be user labor but derives instead from the price premiums assigned to successful brands by financial markets.

Yet as Andrejevic (2015) correctly observes, their analysis likewise fails to cohere, as it seems either to require a wholesale disarticulation of financial value from the productivity of the underlying assets (i.e., corporate profits), and thus leaves them unable to explain the genesis of financial value itself, or it requires them to posit this genesis in the affective involvement of users (e.g., brand equity), thereby returning them precisely to the model they are rejecting. As it is impossible to have it both ways, Andrejevic concludes that Fuchs is basically right, even adding that “The Wall Street Journal has noted . . . that ‘Facebook’s nearly one billion users have become the largest unpaid workforce in history’” (Laney 2012). It follows, in his view, that the source of value generated by SNS is user data, which has become a new “asset class” yet to be included in standard accounting practice and thus reflected in formal valuations of SNS.

At stake in this dispute is the precise source of “unaccounted” surplus value accruing to the platforms. Arvidsson and Colleoni find the market valuation of Facebook far in excess of the value generated by users (i.e., the revenues generated by the company through advertising and data sales), while Andrejevic finds that the “missing” value is simply not being counted yet. However, like the other participants in the digital labor debate, these authors overlook the role of the “third payer” in the production of the apparent surplus. Nor do they take account of the fact that Facebook, at least, does not sell data but rather offers advertisers free access to a highly refined data selection interface as a means of targeting ads, which are in turn sold via competitive auction that pits advertisers against each other within a marketplace monopolized by Facebook itself.

At present, these factors remain of marginal interest for critics of digital labor. ³ But they turn out to be indispensable for a full understanding of the political economy of social media. For the axiomatic premise organizing the debate is that these “third payers” are buying a commodity that, for them, has more or less determinate use value. The former “audience commodity,” reborn as user data and engagement, is on this view a means to an end: selling products and services, either directly or indirectly, via persuasion, branding, and affiliated strategies. ⁴ Yet signs of trouble with this assumption were already visible in the “Blindspot debate” from which it has been inherited. The return of this debate today reveals that, even accompanied by monopoly rents and financialization, the “audience commodity” has never been sufficient to account fully for the profits of the media industries, because it has never delivered the requisite use value to advertisers. ⁵ Those profits therefore require further explanation. And that explanation lies in the competitive, collectively self-defeating logic of advertising, which ensures that advertisers inevitably overpay for a commodity, the utility of which they thereby reduce.

All about the Data?

Like Fuchs and against Arvidsson and Colleoni’s “financialized branding” model, Andrejevic (2015, 9) argues that “it’s all about the data,” which is best understood as the product of uncompensated and concealed user labor:

while Facebook does list the value of intangibles in its financial reports, its data trove does not figure into this calculation. Data is not simply an input into the advertising and marketing process and the creation of the services that Facebook offers to users; it is also an asset with market value that can be bought and sold . . . There is little question, in other words, that the activity of users produces commodities with real market value—in other words, that this activity is labor in the technical sense . . .

It is important to notice that Andrejevic focuses on the market value of the data commodity. User activity qualifies as labor “in the technical sense” that it generates a commodity with exchange value. Marx, of course, famously insisted that the exchange value of a commodity had to be grounded in its use value. In his view, market prices facilitate exploitation precisely to the degree that they obscure the expropriation of use value—that is, a portion of the living labor necessary to produce commodities—by disguising it as exchange value. If the user-produced data commodity has exchange value, what is its underlying use value?

An answer consistent with Arvidsson and Colleoni’s approach is that the data commodity is “fictional,” virtual, or hyperreal, in the sense that its exchange value is generated more or less entirely within the financial sphere and has no meaningful connection to use value. Understandably, this is what Andrejevic is at pains to disprove. The alternative is to show that the consumers of data—chiefly the advertising and marketing industries—derive some use value from the product they purchase. Andrejevic does not offer such a demonstration, and it is quite clear that the utility of the “data trove” springs from the same promise of influence and control that underpins SNS ad sales. Virtually all of Facebook’s $28 billion in revenues for 2017 derive from ad sales, not bulk sales of data, which is almost exclusively provided free of charge to help advertisers target users (Ingram 2017; Tufekci 2018; Vaidhyanathan 2018). Meanwhile, Google’s holding company, Alphabet, is rapidly diversifying its business, so that ad sales, and thus user data, are becoming less important as a source of revenue. Fuchs and Sevignani makes the connection explicit, explaining that user SNS activity

creates a data commodity that is sold to advertising clients . . . [who] thereby obtain the possibility of presenting advertisements that are targeted to users’ interests and online behavior. Targeted advertising is at the heart of the capital accumulation model of many corporate social media platforms. (Fuchs and Sevignani 2013, 237; emphasis added)

In sum, the exchange value of platform user data has primarily derived, up until and including the present at least, from the use value it promises for advertising. It is this promise that the platforms purport to sell—which is not to say that it is what they deliver or what data consumers actually buy. In fact, a good deal of the difficulty in parsing the enigma of the audience/user commodity derives from the peculiar structure of the transaction in which this commodity seems to be exchanged. ⁶ The source of this peculiarity lies in the economic status and function of advertising itself.

Advertising Myths

Doubts about the efficacy and utility of advertising are as old as the profession itself. Indeed, John Wanamaker’s legendary quip that “Half the money I spend on advertising is wasted; trouble is, I don’t know which half” is practically the industry’s ironic motto, ritualistically invoked by lions as well as by most undergraduate textbooks in the field (e.g., Jones 2007; Moriarty, Mithcell, and Wells 2012). ⁷ In fact, uncertainty concerning the persuasive capacity of advertising has profoundly shaped the industry, professional culture, and norms of practice. As a result, this uncertainty has become a vital resource for the business and creative strategies of both marketing and advertising professionals. A substantial literature is devoted to the conflict between “creativity” and “effectiveness” in advertising practice, which results from the advertisers’ perfectly understandable—indeed, structurally mandatory—interest in realizing a specifiable return on its investment in the form of revenue growth, as well as the advertising profession’s steadfast resistance to such simplistic assessments of the value of its painstaking work (e.g., Kover, Goldberg, and James 1995; Lury and Warde 1997; Moeran 1996; Nixon 2003; Nyilasy and Reid 2009). In fact, despite the relentlessly intensifying pressure to demonstrate effectiveness of this sort, the industry has been remarkably successful at convincing clients to focus on a host of other supposed effects, from attention and awareness to feeling, understanding, and “connection” to the brands or products being marketed.

The industry-wide consensus around approaches to effectiveness assessment today strongly privileges self-referential measurements. For example, Moriarty, Mitchell, and Wells (2012, 105) exert considerable effort to show that, contrary to what advertisers may expect and demand, “focus on persuasion illustrates the problem that effects models don’t always cover the essential facets of marketing communication impact.” In their view, long standard in the industry, selling products is the smallest and least important of these impacts. Instead, they “believe communication effects include a wide range of consumer responses to a message—responses that may be just as important as sales” (Moriarty, Mitchell, and Wells 2012, 122). The authors insist on this even as they repeatedly cite John Philip Jones (2007, 4), who states unequivocally that “Psychological measures are very poor predictors of sales effectiveness. . . . These measures are most commonly used as substitutes for sales measures in the widespread belief that it is too difficult to measure reliably advertising’s contribution to sales.” This is because agencies are not interested in demonstrating such effectiveness but in reducing its importance in the minds of clients.

Nor should this be surprising. As Schudson (1984) pointed out around the time of the original blindspot debate, advertising has always been an “uneasy persuasion,” with companies most in need of effective campaigns repeatedly discovering the limits of the discipline. More recently, Cronin (2004) has argued convincingly that much of what practitioners as well as critics of advertising habitually assert about its power functions as a myth in Barthes’s sense of the term. Lury and Warde (1997) draw on a wealth of ethnographic evidence to show that the industry’s capacity to sell its expertise to clients relies centrally on its success in devising precisely the sorts of consumer research techniques that position the profession itself as the sole arbiter of efficacy. This ultimately leads them to call advertising “the modern witch doctor” and its measurement techniques “ultimately unverifiable myths” (Lury and Warde 1997, 96).

Similarly, McFall (2004; 2007) has mounted a thoroughgoing challenge to the prevailing view that advertising has, through new techniques and increased ubiquity, come to exercise considerable power to persuade consumers. Among the problems with this view is that it is compelled to position advertisers untenably both as absolute masters of meaning production and as lacking any substantive control over the meaning (or sign-value) of their products (Ang 1991). As an alternative, McFall (2007, 126) directs attention to “the generative significance of internal advertising processes . . . in shaping the reality they purport to measure.” What emerges in the literature examining these processes is “a glimpse of advertising as an inward-looking and self-referential culture” (McFall 2007, 136) whose products cannot be grasped in terms of their persuasive effects on consumers.

In a penetrating and revealing ethnographic study, Nyilasy and Reid (2009, 86) quote one creative professional’s formulation of the prevailing view that “By and large, advertising can’t make you buy something that you don’t want or don’t need.” Although the authors find that agency practitioners strongly believe that exposure to ads causes changes in human cognition, emotions, and behavior, the profession views these effects as fundamentally indirect and impossible to ascertain or reliably predict. This is important, because the prevailing “semiotic ideology” (Keane 2003) in the industry ensures that the only effects that can be measured are those divorced from sales but relatively easy to demonstrate to clients.

Indeed, independent empirical studies of advertising effectiveness consistently show, at best, mixed results. To wit, in a massive review of the literature, Sethuraman (2011) finds that from 1960 to 2008, advertising elasticity (i.e., the increase in sales or market share for a 1 percent increase in spending) has steadily declined. Similarly,

Based on over 260 estimates, the mean elasticity of sales or market share to advertising is 0.1 percent. . . . advertising is not the variable of choice for increasing sales. . . . Even if advertisers make a big increase or decrease in weight, sales do not increase or decrease by much. (Tellis 2009)

In an effort to make a more compelling case for effectiveness in an earlier literature review, Kim (1992) stipulated that ad agencies stand to benefit by convincing clients to focus on “long-term” goals and “branding” strategies that are largely assessed in self-referential terms. Such models allow agencies to replace the question of sales with measures of consumer perceptions, which are in turn defined in terms of the brand image itself. This pseudo-empirical circularity enables agencies to demonstrate effectiveness to clients in ways that defer indefinitely the question of sales, market share, pricing power, and so on. It also obviously intensifies the advertisers’ reliance on the profession, and now on SNS, as punctual campaigns are supplemented or replaced by proliferating forms of “relationship marketing” that require an interminable process of audience composition. Similarly, scholars of political communication have long understood that campaign advertising benefits political consultants far more than the candidates (Bennett and Iyengar 2008; Kalla and Broockman 2018; Mutz 1998; Zaller 1992). And for all the media attention devoted to the Cambridge Analytica scandal, the available research suggests that its refined targeting techniques predicated on Facebook data have produced negligible persuasive effects (Kalla and Broockman 2018; Martínez 2018; Nyhan 2018).

The progressive decline in what was always a dubious contribution to revenue growth attributable to advertising is undoubtedly linked to the well-known problem of clutter (Kent and Kellaris 2001; Rotfeld 2006). Indeed, Andrejevic (2013) himself has written brilliantly about this problem, which he elaborates into a powerful critical category, infoglut, and uses adroitly to explain a wide range of ostensibly unrelated phenomena comprising the cultural logic of communicative capitalism. The diagnostic scope and flexibility of the concept derives in good part from the fact that both clutter and infoglut refer to a familiar class of collective action problems, of which the paradigmatic instance is the arms race. The structure and dynamics of an arms race ensure that the participants waste resources while failing to gain an advantage over rivals. Put differently, an arms race guarantees that the use value being purchased—that is, the security provided by weapons systems—is never acquired. This is precisely the problem clutter represents: the more companies advertise, the less effective their advertising becomes (even on the self-referential measures of attention, recall, loyalty, etc.). Thus, while the advertising and marketing industry has spent decades convincing its clients to fund campaigns unlikely to improve their bottom line in proportion to expenditures, the spending itself makes such improvements even less likely.

Branding, and the accompanying shift in the definition of effectiveness, was devised in part as a solution to this problem. The techniques Andrejevic assembles under the rubric of infoglut (psychographics, neuromarketing, sentiment analysis, etc.) are more recent efforts to address it, this time not by reducing the clutter but by deploying it strategically. This is precisely what SNS do. First, by maintaining their “always on” experience, the platforms ensure that user attention and time on the site are increasingly fractured, so that the engagement sold to advertisers in the form of data is increasingly an artifact of feed engineering and assorted other techniques. And to this extent, the customers are being shortchanged, if not duped.

Second, insofar as this overload is a proxy for (no less than stimulus to) engagement, it is sharply at odds with the ostensible commercial aims ascribed to platforms. Advertisers want exposure and engagement, but they certainly do not want clutter, overload, or distraction. Thus, while Facebook and other platforms frame their sales pitch to clients in terms of data-driven targeting and the possibility of direct interaction with users, the techniques they employ to make this possible directly contravene their customers’ aims in paying for such access. Indeed, Facebook is far more cluttered than television (Nelson-Field, Riebe and Sharp 2013)—so much so that “The improvements in advertising recall . . . do not appear to be sufficient to justify the likely price premium that advertisers would have to pay to reduce clutter on Facebook.” And Facebook’s own research disputes the notion that engagement with a brand leads to increased sales (Smallwood 2016). Put simply, they actively generate the very glut that their customers pay to break through. ⁸ Clutter is their business model.

One way to understand the long-term consistency in ad spending as a fraction of revenues is to see it as an index of the arms race structure. On one hand, companies are loath to increase spending, in recognition of the limited and self-vitiating impact their campaigns have on sales, margins, and market share. ⁹ On the other hand, the major advertisers, which tend to comprise the same limited roster of large multinationals year after year, are reluctant to reduce or stop their advertising activities, because this would aid their competitors. In other words, they are effectively trapped in the circuit of advertising like drivers in rush hour traffic, expending resources on activities that, taken together with those of others, are collectively self-defeating in relation to the function supposed to justify them. It goes without saying that the chief beneficiary of this predicament is the media industry—and audiences/users. In addition to the content funded by advertising (at levels far in excess of optimality), the surplus revenues fund the infrastructure on which users come to rely. At the same time, the structure of the advertising arms race ensures that “the work of watching,” as well as “the work of being watched,” is actually minimized by means of distraction. Users are prevented from paying much attention to ads, just as TV audiences have always routinely used commercial breaks to perform other activities.

Indeed, Facebook in particular notoriously manipulates user feeds to manage the reach of its largest content creators, who find their posts presented to only a fraction of the audience they bring to the platform. On its face, this is perfectly sensible, inasmuch as users with the largest following would otherwise dominate the feeds of millions, sharply reducing everyone else’s reach in the process. Yet for this very reason, every audience sold to advertisers is a product not only of the filtering tools applied to user-generated data, but of the way feed algorithms continuously drive the sorts of interactions in which users partake—and thus the information represented by the resulting data. In effect, almost any group of users can be assembled in real time to play the part of an empirically verifiable ideal audience, since the feedback loop between constructing and mining user data renders this process self-referential (for a nuanced analysis of this circularity, see McStay 2011). Part of the problem is that SNS advertising response is subject to high levels of endogeneity, meaning that online ads “work” by insinuating themselves into an ongoing purchase process. Users click on ads served to them on the basis of their demonstrated interest in a product or category, allowing platforms to take credit for sales initiated by users via Google and Amazon searches (Blake, Tadelis and Nosko 2014; Marvin 2015), rendering SNS user data largely redundant.

Meanwhile, the sheer volume of data and variety of analytic tools made available to advertisers (e.g., Jackson 2018; Sterne 2010) is a common source of frustration for marketing professionals (Brenner 2016; Shaprio 2009), who tirelessly urge their clients to focus less on such metrics as impressions, reach, “likes,” or engagement and more on clicks and conversions. It is clearly in the platforms’ interest to oversupply advertisers with largely superfluous metrics, as this enhances the clients’ sense of control and the impression that campaigns are producing effects not captured by mere sales (e.g., McKay 2017). In so doing, the platforms are taking advantage of the advertising and marketing industries’ long-term self-serving project of expanding the very concept of effectiveness (LaPointe 2009). Hence, the countervailing tendency in the industry is to instruct clients that since SNS so dramatically expand and intensify access to potential customers, effectiveness should be measured even more expansively, so that “returns from social media investments will not always be measured in dollars, but also in customer behaviors (consumer investments) tied to particular social media applications” (Hoffman and Fodor 2010, 42; original emphasis). This division in the digital marketing industry on the question of what counts as effective is precisely “political” (LaPointe 2009), in the sense that it cannot be settled empirically because it concerns what empirical metrics can and should determine.

To be clear, this is not to say that advertising produces no useful effects for businesses, however assessed. Rather, it is to stress the pivotal role of structural divergence between what SNS appears to offer for sale, what advertisers seek to buy, and what the resulting exchange actually delivers. It is this divergence that best explains the disproportionate size of SNS profits in relation to the contribution of commodified user labor, by tracing them to their source in the transaction between buyers and sellers, a transaction in which the use and labor value comprising the ostensive commodity are partly destroyed. If what advertisers pay for is a consequent decline in the likelihood that their ads will reach, engage, or convince users to make purchases (Fahey 2017), then each additional dollar spent decreases its own marginal utility. Hence, the profits earned by Facebook and Google rise as a function of and in proportion to the coerced profligacy of their clients’ expenditures. In effect, the commodity actually purchased in this process is this perverse outcome. It follows that while users certainly engage in a form of labor within the chain of transactions comprising the SNS economy, strictly speaking the exploited parties are the companies paying for this commodity. From their point of view, this is sheer waste; from the SNS point of view, it is pure profit; and for users, this is the circuitous route by means of which “third payers” provide the platforms “for free.”

Nor has the conflict between the interests of the platforms and their clients escaped notice. Google’s metrics have been suspended by the Media Ratings Council (O’Reilly 2016), and P&G, the largest and most analytically sophisticated advertiser in the world, has been aggressively reducing its digital ad spending and vocally criticizing social media for wasting advertiser resources (Neff 2017; Seetharaman 2016; Vranica 2018a, 2018b). The conundrum is quite widespread, increasingly becoming evident in surveys of smaller advertisers, as well as SNS users. For example, 62 percent of small businesses advertising on Facebook find their ads are not seen by the audience targeted via the platform’s vaunted data tools (Sophy 2017). Even the largest, most heavily advertised brands enjoy shockingly small levels of user engagement (Sitta et al. 2018). Conversely, most users do not regard such advertising as “useful, meaningful, and leading to purchase intentions” (Baglione and Tucci 2018; Greenlight 2012). Nor do so-called “multi-touch” strategies—themselves devised to compensate for the shortcomings of paid advertising (Einstein 2016)—fare much better, as advertiser efforts to repeatedly engage users typically backfire, driving them away from the products or brands on offer (Baltas 2003; Brettel et al. 2015). Instead, engagement is typically initiated by users, rather than in response to marketing efforts (LaPointe 2012). Yet even effective engagement techniques beg the question of user labor, since unlike advertising, most such activity is not a source of revenue for SNS.

In a way, then, this development is a step backward for advertisers, as the shift to branding represented a departure from the entire question of sales revenue growth. Put simply, brand advertising is more or less explicitly a form of competitive consumption. Long misunderstood as a key aspect of the so-called “demand management industry” (e.g., Ewen 1976, 1982; Harvey 2006; Klein 1999; Maxwell 1991), it had less to do with stimulating demand than with occupying perceptual space to reduce the presence of competitors in it (Aaker 2011). It is in this sense that clutter has been deployed productively: firms could advertise to increase the scarcity of audience attention, crowding out competitors. By the same token, branding has served both symbolically to intimidate would-be competitors from entering competition (Bagwell 1990; Thomas 1999) and materially to increase the cost of advertising for them, raising the barrier to entry (Gable et al. 1995; Pehrsson 2009).

It is crucial to see that none of these motives or effects has anything to do with consumers themselves. Within the political economy of digital media, it makes essentially no difference whether or how the ads affect them. Functionally, if not always strategically, the primary targets of branding (as distinct from “direct marketing”) campaigns are rival companies. And indeed, the largest SNS advertisers—the ones effectively increasing prices for everyone else and each other—are almost exclusively the largest consumer brand portfolios, such as P&G and Unilever. In other words, the largest expenditures are made by companies with the least to gain from them, both at the margin and on net (Sitta et al. 2018). These brands already dominate their markets and enjoy massive structural advantages that competitors are unlikely to overcome, which is what enables them to fund expenditures far in excess of their economic use value. This surplus, produced by means of the purchase of platform data and user access, ensures that audience/user labor is functionally unproductive.

Selling the Blind Spot

In this regard, Arvidsson and Colleoni make a uniquely helpful contribution to the audience commodity debate. Insofar, as stock prices increasingly reflect the estimated value of brands as balance sheet assets, the inherent circularity of advertising measurements is a veritable boon. It now becomes possible to guarantee a positive return on investment on ad spending simply by demonstrating that target consumers have experienced and “understood” the ads. Up to a point, additional spending returns almost immediately in the form of higher brand valuation and share price. All this can now take place without adding a single sale or even increasing margins by boosting the brand premium (which remains constrained by price competition at the point of sale). In effect, companies can enhance their own book value by means of what are, from a strictly economic view, unproductive expenditures.

The largest advertisers undoubtedly understand this and tacitly cooperate with the platforms in coercing smaller companies into expenditures that contradict their interests and stretch their resources. By definition, market-leading brands have little room for additional growth, but they do tend to crowd out smaller competitors (Madden et al. 2011). Moreover, sales are typically expanded by means of effective (i.e., monopolistic) distribution strategies, and strategic acquisitions or partnerships, in which advertising plays at most a supporting role (Aaker 2011). That is, it is far more effective to get one’s products onto Walmart shelves than to advertise them relentlessly, but it is easier to do the former by assuring Walmart that the latter will help support demand. Conversely, massive long-term ad spending to promote Budweiser have proven no match for the rapidly rising popularity of microbrews with minuscule advertising budgets, and Anheuser-Busch responded by merging with InBev, not by intensifying its advertising efforts.

All of this helps explain why the largest ad buyers tend to be brands that stand to gain least from the spending. Simply put, expanding sales revenues is not their primary aim, and their ad campaigns are consequences rather than causes of market dominance. After all, “the vast majority of brands fail, and . . . [t]he stories that get told about brands in the professional and other literature are almost always success stories” (Moore 2003, 334). As in any status competition, the primary target audience consists of rivals, but more importantly, their expenditures serve, deliberately or not, to reduce the opportunities competitors have to address consumers. This necessarily means that the attention and response of actual users is, at best, of secondary importance. In effect, the visible extravagance of their spending itself functions as a market strategy.

For their part, SNS can only profit if they refuse or fail to provide their customers the service they appear to be selling. Their power resides not in the secrecy or obscurity of their algorithms and promotional “black ops,” but in the forced choice with which they confront their target market—advertisers and their consultants and agents. Beyond or by way of the virtual monopoly they have established, the rents they extract are not only a surplus over the marginal market value of their ostensible product but in fact can only be derived if the use value of this product is reduced by the very process of extracting the rent. In effect, the customers themselves are induced into facilitating their own exploitation. The more they compel each other to pay for access to users, the less this access is worth and the higher the surplus profits.

We thus arrive at the unexpected conclusion that SNS profits derive in critical part from “selling the blindspot,” or the structural, meticulously staged obstacle to influencing users. The ostensible audience commodity—in its new form as user data and interactivity—has, from the start, fueled a collectively self-defeating arms race that funds the digital media industries under what amount to false pretenses. It is not primarily by fulfilling the promise of delivering consumer data, attention, or spending to advertisers but rather by actively impeding its fulfillment that the digital economy functions. The labor involved is that of the consumers who pay for the marginal cost of advertising reflected in prices, as well as the workers in the advertising industry. Meanwhile, the value of access to users derives from its scarcity, which in turn results from the competition for consumer attention among advertisers. It is the advertisers, then, who are compelled, often against their interests or preferences, to generate this peculiar form of surplus value by means of ultimately profligate expenditures. Put simply, it is status competition between owners of industrial capital that drives the transfer of a portion of this capital to the media industries. In effect, this transfer is not fully explicable as a form of exchange but is better understood as an externality.