Human labor is inherently collective, in the sense that individuals rely on others’ labor in order to bring their own to fruition. This dependence calls for some kind of structure and organization to enable and support a division of labor within the collective. Organizing, however, incurs costs—of communication, coordination, control, and collaboration. These costs represent an interesting problem for organizations because they must expend a portion of their resources to manage the division of labor. The larger the number of people involved, the higher the cost of management.
Historically, societies differ in how they divide the labor and how they handle the costs. We have discussed how in feudal societies, the division of labor obligated the serf to work on the feudal land part of the time and his own land the rest of the time. The cost to the serf, i.e., working for the lord, was traded against the cost to the lord—providing land and protection. This arrangement was regulated through mechanisms of obligation and punishment (Burawoy 1979). Early capitalism, which freed labor from its bonds to the land, devised new mechanisms to ensure the production of value. Bringing workers under the same roof, it used spatial and temporal techniques, such as clocking in and timesheets, to guarantee the extraction of value from labor. The cost to the worker—partially unpaid labor—was traded against the cost to the capitalist—providing the means of subsistence. This arrangement provided a practical solution to the problem of control, although it incurred new costs (factories, warehouses, supervision, and so on).
The expansion of the scope of the operation of capitalist enterprises transformed them into multi-unit, hierarchical organizations, exacerbating the problem of control into a “crisis of control” (Beniger 1986). The evolution of the railroad system in the United States represents this phenomenon in a very clear manner. As railroad operators expanded their scope from 50 to 100 miles in the early nineteenth century to thousands of miles by the end of the century, running long stretches of railroad networks in different directions, there was an obvious need for coordination in terms of organizational and technological standards, rights of way, freight transportation, scheduling, and the like. Before the expansion, freight moving from Philadelphia to Chicago had to be unloaded and reloaded as many as nine times (Chandler 1980). To respond to this inefficiency, railroad companies created bureaucracies with two or three layers of middle managers running operating divisions, geographical offices, and central headquarters. Chandler (1980) names some of the advantages of this type of organization—economies of speed (cutting costs through standards and routines), flexibility in responding to customer needs, and a steady flow of cash, which reduced the cost of loans and credits. The disadvantage was that the layers of bureaucracy imposed huge expenditures for office buildings, accounting, training, and other overhead costs.
The transition in the later part of the twentieth century to decentralization and flexibility provided a scheme to adjust and reduce these costs. With the service sector and “knowledge organizations” coming to the forefront, and with business and management gurus advocating “liberated firms,” Total Quality Management, and project-based units of work (Peters 1992), the best solution to the problem of control was found to be “self-control”—that is, for people to control themselves. This “solution” speaks directly to the growing prominence of notions such as “workforce participation” and “intrinsic motivation,” which sought to highlight “the desire to do the work and the pleasure of doing it,” as opposed to systems of external rewards (Boltanski and Chiapello 2005, p. 80). The excitement in the United States and the rest of the world around Toyota and its “lean” model of organization largely derives from its successful implementation of notions based on participation and intrinsic motivation (Womack, Jones, and Roos 2007).
For our purposes, these developments find parallel incarnations in a type of heteromation we dub “organizing labor” because it draws on the human capacity to organize activity. Here we understand “organizing” in the broad sense of goal-directed social entities designing and implementing coordinated activity systems linked to the external environment (Daft 2011). We discuss two cases where organizing labor plays a prominent role in the extraction of value.
The implementation of self-control in organizations takes two forms: peer control and customer control. Peer control replaced the direct supervision of the old hierarchies with the evaluation of peers and team members, organized in projects, looking over each other’s performances. Customer control extends this model beyond the boundaries of organizations to the appraisals of customers, whose “satisfaction” has been portrayed as the holy grail of business practice for decades. The expansion of the internet and e-commerce has brought customer control into prominence, to the point where one can hardly interact with companies online or offline nowadays without being asked to provide feedback on one’s “experience.”1 The rhetoric of customer satisfaction and experience, while partly genuine, hides the more important benefits of customer reviews to business enterprises.
A survey of business literature on this topic illustrates this quite clearly. The list of the substantial benefits of customer reviews to businesses includes referrals, repeat business, loyalty, retention, customer acquisition, competitive advantage, reputation, sales opportunities, morale and motivation, and education (Letts 2013). Some consider reviews “a great SEO [Search Engine Optimization] boost” for business, arguing that even “negative reviews aren’t actually bad; they help establish the authenticity of your product” (Lentejas 2013). What is notable here is that the concern of these commentaries is how businesses stand to benefit from customer reviews, discounting the time and labor that customers put into them, despite the clear economic value of the reviews, as acknowledged by the businesses themselves. There is, needless to say, potential benefit to the customer in terms of the information gained from peer reviews, but such benefits pale in comparison to the profit-seeking benefits for enterprises.
We turn to an analysis of organizing labor in citizen science, an arena very different from commerce in its goals, norms, and outcomes. A relatively long tradition of “community archaeology,” for instance, has engaged citizen volunteers in various kinds of projects, some of which are run by amateurs (Simpson and Williams 2008). The Edgware Junior School in London, for example, which was the site of an air-raid shelter in World War II, became an archeological site in 2006. Prompted by the accidental discovery of a concrete structure in the school’s playing field, a team of excavators joined the local archeological society in a project with the pedagogical aim of teaching students about the history of the period, as well as introducing the principles and methods of excavation and archeology more broadly (Moshenska 2007).
Largely successful in its aims, this project is an example of an organization “with little or no professional involvement and guidance” (Simpson and Williams 2008, p. 74). As the authors point out, however, the general idea of “archaeology by the people for the people … is something of a naïve fantasy [because] in reality, community archaeology is censored and manipulated, and communication of information and access to the past is controlled through many different agencies” (ibid., p. 72). This situation begins to sound like the corporate governance of the modding community, where mods can be squelched instantly if a company does not like them.
The introduction of Web 2.0 technologies into archeology seems to have exacerbated these issues. Based on a survey of a broad range of archeological projects that leverage social media for public participation, Perry and Beale (2015, p. 154) found that
Discussions of power, labour, consumption, and capitalistic control that fuel much archaeological web-based engagement are often nonexistent or undertheorized by developers, users, and other implicated parties.
Drawing on the notion of heteromation, Perry and Beale (2015, p. 159) note that “via the social web, volunteers are regularly drawn into the project of resolving profound internal organisational problems: bureaucratic, personal and workflow oriented.” They criticize the “exploitative politics” behind citizen projects that do not provide actual learning experiences, “instead using amateurs and beginners as ‘workhorses’” (p. 158). One such project, for instance, makes the following recommendation: “Crowdsourcing application must be designed carefully to give the impression that participants volunteer and are not working for free” (p. 158; emphasis added by Perry and Beale). The “exploitative politics” behind such statements, according to these authors, is hidden behind the volunteers’ desire for personal fulfillment and socialization—an antidote to the professional labor bottleneck that has arisen in the era of austerity in Britain and elsewhere. In this light, Perry and Beale conclude that the gains to disciplinary and institutional centers are much more discernible than benefits to individual volunteers.
Other studies of citizen science have shown a smaller number of projects that enable fuller forms of participation, higher levels of engagement, and a balanced distribution of rewards compared to the greater number that disallow or limit contribution and recognition (Cornwall 2008). Qaurooni et al. (2016) studied a wide array of citizen science projects, discovering a similar bias toward what they call “Crowd Science” compared to the more collegial and participatory “Civic Science.” As examples of these types of citizen science, the authors discuss two map-making projects, Cropland Capture2 and Kite Mapping,3 both of which draw on citizen help for map construction. The projects differ on key dimensions such as citizen roles, organization, and technology. Cropland Capture recruits citizens to remedy scientists’ poor understanding of cropland locations across the globe, having them evaluate scientist-selected land areas through a simple yes-no question, using a gamified smartphone application and geotagged pictures. Kite Mapping, on the other hand, considers citizen participation as a means to contest official maps and challenge their beneficiaries. To that end, it allows people to pick the time and place of their interest, provides them with guidelines and support for buying, assembling, and flying a map-making kite using simple household items, and equips them with tailor-made software to stitch together their own map of an area. The result is then submitted to the public record and may be used for advocacy and activism, as well as research (ibid.).
The contributions of volunteers in a significant number of citizen science projects include activities such as organizing or hosting groups, classes, and workshops for topics such as climate change, environmental monitoring, volunteer training, brainstorming sessions, solving math problems, raising funds, and collecting debris. Organizers and participants of these events often report their findings to a lab or institution without having access to the bigger picture of how their findings are incorporated into scientific work and scholarly publications. Study after study has shown that there are significantly more citizen science projects that are like Cropland Capture than Kite Mapping (ibid.). The distribution of the projects illustrates a pattern that is observed in increasingly different sites and areas, as we see next.
Customer reviews and citizen science provide two examples of a larger trend in a constellation of activities referred to as “crowdsourcing,” “wikinomics,” or “human computation.” A common resource of these activities is the self-organizing capacity of human beings, which provides an alternative model to old hierarchies, eliminating or minimizing the cost of control and coordination (Benkler 2002). The trend reflects the potential of Web 2.0 technologies for facilitating coordinated activity at minimal cost. Shirky (2010) provides a large set of examples, including Wikipedia and its concept of collaborative content creation; the Linux operating system and its open-source approach to programming; the Voice of the Faithful campaign against child abuse in the Catholic church; the Coalition for an Airline Passenger Bill of Rights; the successful campaign of the student customers of the HSBC Bank to revert an overdraft penalty; Howard Dean’s supporters in the US presidential campaign of 2004, who did some of the first online political organizing; and the use of Twitter by political activists in Egypt.
The list is, indeed, compelling in giving us a glimpse of the power of what Shirky calls “social tools” in enabling new kinds of “organizing without organizations.” The power derives from the capacity these tools provide for coordinated action across time and space, at large scale and high speed, and with almost no cost. In spite of these advantages, however, the majority of participatory projects do not succeed, failing to accomplish their goals or sustain themselves. In fact, the performance of projects seems to reveal a pattern similar to what happens in citizen science. This pattern can often be expressed in the form of a power law distribution—that is, a distribution of projects with “a lot of failure, some modest success, and a few extremely popular [projects]” (Shirky 2010, p. 235). Of about 100,000 open source projects on SourceForge.net (a site that hosts many such projects), for instance, very few are downloaded millions of times, and the majority don’t get even a thousand downloads. Below the 75th percentile, projects have no downloads whatsoever (ibid., p. 244). Shirky explains this pattern in terms of the low cost of failure which allows projects to explore and experiment with different possibilities, essentially “for free.” Terms such as “success,” “failure,” and “free,” are relative descriptors, however, leading one to ask, Free for whom? and, Who bears the costs of failure or the benefits of success?
To answer these questions, we need to consider another power law distribution that is as commonly observed within these projects as it is among them. This distribution has to do with the division of labor among participants. Numerous studies have shown that the bulk of the work in participatory projects is done by a very small percentage of participants, with the majority contributing little. This is the case whether we consider Linux programmers, Wikipedia authors and editors, or people who tag photos on Flickr (Shirky 2010, p. 124). Shirky explains this “predictable imbalance” in terms of differences in motivation: “The number of people who are willing to start something is smaller, much smaller, than the number of people who are willing to contribute once someone else starts something” (ibid., p. 239). Motivational differences, according to this line of thinking, lead to a “spontaneous division of labor,” where “no effort is made to even out … contributions” (ibid., p. 125). Actually, however, the bigger point is that this spontaneous division of labor “wouldn’t be possible if there were concern for reducing inequality … [because] most large social experiments are engines for harnessing inequality rather than limiting it” (ibid.). Not only should this “natural” state of affairs in terms of motivational differences and inequality not be discouraged, it should be maintained and encouraged because it guarantees the emergence of a small number of “a few extremely popular” projects. Inequality is endorsed as natural and inevitable.
We find the sense of generality, spontaneity, and inevitability emanating from this logic problematic. First, to use individual motivation as the explanation of these phenomena inverts the causal order, putting the (social) horse of predicaments and possibilities behind the cart of individual psychology. It is true that people vary in terms of their motivations, but the generators and enablers of those motivations can be largely found in the socioeconomic circumstances of people’s upbringing and current life circumstances. Rather than attributing differences in participation level to differences in “intrinsic” motivations among individuals, therefore, it makes more sense to think of this in terms of their ability to maintain commitment to particular goals. If we understand committed behavior as the capacity to engage in “consistent lines of activities” (Becker 1960), then varying behaviors can be explained according to the outcomes they produce and the system of values that is applied in assessing those outcomes: “What kinds of things are conventionally wanted, what losses feared? What are the good things of life whose continued enjoyment can be staked on continuing to follow a consistent line of action?” (ibid., p. 39). Psychologists and behavioral economists have documented, for instance, the effect of poverty on mental stress, with deep influences on people’s willingness to set goals and on their commitment to work hard to accomplish them (Haushofer and Fehr 2014). Rather than motivation being an inherent attribute of individual psychologies, it is in large part the outcome of socioeconomic circumstances. For example, Körner, Reitzle, and Silbereisen (2012, p. 190) point out that, “Even in wealthy, stable countries with a strong safety net such as Germany, unemployment may lead to a downward spiral of psychological distress.”
Second, the “spontaneity” of the division of labor is relative to how group relationships are configured, resources are divided, and information is distributed. “Wikipedia, which looks like a reference work to the average viewer, is in fact a bureaucracy given over to arguing,” Shirky comments (2008, p. 278). Wikipedia has around a dozen administrative collections of pages, only one of which is for the actual articles (ibid., p. 279). What seems on the surface to be a spontaneous arrangement is actually a structure implemented to maintain control of the editing process and the integrity of the content. Forte, Larco, and Bruckman (2009) found that the Wikipedia community governed itself by articulating its norms as rules and strictly enforcing those rules (ibid.). This structure, embedded into the Wikipedia technology, along with resistance toward commercialization (and the financial backing of Jimmy Wales), have enabled Wikipedia to survive as an open system for many years. Wikipedia is not the result of “motivated individuals,” but of commitments that give rise to strong organizational and disciplinary structures.
While Wikipedia has thus far adhered to its founding principles of collaborative and open content creation, not all systems are founded on such principles.4 And this brings us to asymmetries in the distribution of rewards that is becoming increasingly prevalent in participatory projects. The asymmetries derive from a seeming paradox in the relationship between individuals and groups.
Describing the most highly connected people as the backbone of social networks, Shirky (ibid., p. 213–214) argues that these networks “are held together not by the bulk of people with hundreds of connections, but by the few people with tens of thousands.”5 A few pages later, he acknowledges: “Perhaps the most significant effect of our new tools, though, lies in the increased leverage they give the most connected people” (ibid., p. 225; emphasis added). These statements speak to a paradox: Is the powerful position of the highly connected people an inevitable outcome of differential motivations among people, or is it an artifact of the design of current networks and their “winner-take-all” incentive structure?
In the absence of empirical data based on a thorough exploration of both scenarios, we might not be able to settle this question. We do know, however, which alternative is more compatible with the spirit of capitalism. In particular, developments in the last few years show a clear trend toward the appropriation of participatory projects in favor of the interests of a select few at the expense of the large majority. The wealth accumulated in the last decade by technology companies such as Facebook, Google, Amazon, and others illustrates the “resolving” of the paradox toward the design of systems that have turned a very small percentage of people into the major beneficiaries of these projects. This trend provides an interesting example of how capitalist institutions have by and large managed to co-opt the outcomes and mechanisms of participatory projects in various activities. In so doing, capitalism has, yet again, turned a predicament into an opportunity for its own gain. Therein lies the true nature of the power law distribution: the long tail gets the short shrift.
A telling example comes from the case of the Canadian firm Goldcorp Inc. Around 2000, when gold markets were shrinking, Goldcorp’s 50-year-old mines on a property in Red Lake, Ontario, were depleted of cheaply extractable deposits, putting the company on the verge of bankruptcy. Having learned about Linux, the new CEO devised a plan for opening up the company’s databases to allow a participatory competition to identify promising locations for excavation. With a total purse of $575,000, an army of geologists, mathematicians, students, and other folk identified 110 targets on the 55,000-acre property, 50 percent of which had not been identified beforehand. The outcome of this initiative was turning a $100 million company into a $9 billion giant! Prize money of a few hundred thousand dollars, in other words, gave rise to many billions of dollars. Considered a gold standard, if you will, of participatory projects, the case of Goldcorp Inc. reveals the dynamics of relative rewards in power law distributions and how they hugely favor capital.
In this light, the increasing prominence of power law distributions, which has apparently come at the expense of bell curves and their normal distributions, should be understood as an outcome of expanding computer-mediated networks and the kinds of social arrangements they support. Power law distributions, in other words, serve the same analytic and normative function in networked environments that bell curves provided in hierarchically layered environments. They normalize a radically non-symmetric distribution of rewards, wealth, and prestige while bell curves normalized the “average” and the “normal.” In an ironic twist, however, while the “normal” distribution of bell curves hid individual idiosyncrasies, power law distributions do the opposite, accentuating individual differences and aggravating inequality. While old-fashioned statistics could “lie” by burying individual differences under the shadow of averages (Huff 1954), newfangled Bayesian statistics have the potential to bury a large percentage of the population under the logic of networks, rendering their contributions to the working of networks invisible. The old economists’ joke—that Bill Gates walks into a bar, and the average income of everyone becomes $1 million—loses its punch, giving way to a more current version: Mark Zuckerberg walks into a bar, and everyone likes everyone, but no one has enough money to pay for their drink. Then Eric Schmidt walks in and invites everyone to search on their Androids for moneymaking ideas using Google’s real-time analytics. Finally, Jeff Bezos walks in, and offers everyone a HIT!
The questions and problems introduced by organizing labor embedded in a capitalist economy highlight the ambivalent character of computing technology in a rather vivid manner, facing us with what can be described as the paradox of participation: Should we, or shouldn’t we, as individuals and communities, partake in these projects? Refraining from participation would deprive us of an effective mechanism of innovation, contribution, and community building. Taking part, on the other hand, not only might enhance social and economic inequality, it can expose us to unpredictable risks and consequences. While a sizable portion of the population seems to have resolved this dilemma in practice, favoring participation and contribution, the situation is more complex. The complexity is evident in a new initiative called Human Computation, which seeks to leverage the potential of computer-mediated participation by structuring and formalizing the division of labor between humans and machines in an explicit way, and arguing for its pervasive implementation throughout the economy and society. The Handbook of Human Computation (Michelucci 2013, p. 85) defines the field as one that considers “the design and analysis of multi-agent information processing systems in which humans participate as computational elements.” The idea of humans as “computational elements” provides a tellingly accurate depiction of what is involved in organizing labor and the mechanisms that make human labor invisible. While sounding innocuous, this conception of the place of human beings in the big scheme of things runs the risk, especially in a capitalist system, of undermining basic human rights and values. If industrial capitalism turned individuals into little cogs in huge manufacturing plants, giving rise to the predicaments of alienation and isolation, current capitalism turns people into “bits” of information, pieces of wetware, or simply abled bodies ultimately in the service of others, usually powerful others. The cycle of innovation and co-optation repeats itself, and generates new possibilities and predicaments.
What this implies for the future for the future of human individuals, communities, and societies is a question that we pick up in part III of the book.