Abstract
What mechanisms has Amazon deployed in its effort to control the labor of its warehouse employees? This question holds both practical and theoretical interest, given Amazon's prominent position in the economy and the wider importance of the logistics sector for consumer capitalism. This paper, part of a broader mixed-methods study of Amazon's workplace regime, uses a small national sample of interviews with Amazon warehouse workers (N = 46) to identify the mechanisms of labor control the company invokes. In keeping with accounts propounded by activists and journalists, we find evidence of highly coercive labor controls, chiefly in the form of what we call techno-economic despotism (which applies algorithmic technology to a precariously employed workforce). Yet many workers also experience forms of labor control that rely not on coercion but on the generation of consent. We identify three such mechanisms of hegemonic labor control - normative, relational, and governmental – that Amazon uses to foster workers’ consent. The efficacy of Amazon's workplace regime stems largely from its ability to deploy a multiplicity of labor controls that resonate with different groups holding distinct positions in the labor process. Given shifts in the social and economic conditions that bear on the company's regime, cracks have begun to appear in Amazon's armor, potentially reducing the traction its labor control mechanisms have gained with segments of its employees.
What mechanisms has Amazon deployed in its effort to control the labor of its warehouse employees? The question holds particular importance at a time when the public has grown increasingly aware of our dependence on logistics systems and when a resurgence of labor conflict has seemed to occur throughout many parts of the US retail sector. Yet in contrast to firms such as Walmart, whose operations have been the object of much scholarly attention (see Appelbaum and Lichtenstein, 2006; Lichtenstein, 2009; Petrovic and Hamilton, 2006; Reich and Bearman, 2018), Amazon's labor practices have attracted relatively few social scientific studies (see Allison et al., 2018; Alimahomed-Wilson et al., 2020; Delfanti, 2021). In this vacuum, journalistic accounts have been helpful in showing how the company's obsessive focus on customer satisfaction has led it to impose extraordinarily intense demands on its blue- and white-collar employees alike (e.g., Kantor and Streitfeld, 2015; Kantor et al., 2021a, 2021b). Yet such accounts have tended to focus on the physical dangers and harsh working conditions that Amazon workers confront, with only occasional allusion to the managerial strategies and labor control mechanisms that underlie such practices. Especially in the case of so prominent and powerful a firm as Amazon, whose operations have enormous impact on employment relations more generally, we need careful and sustained inquiries into the labor control mechanisms on which it relies.
In this paper we provide an initial effort to fill this social scientific gap. Part of a larger mixed-method study of Amazon's labor practices, the paper uses interview data gathered from a small national sample of the company's warehouse workers (N = 46) to shed light on the mechanisms management uses to ensure the compliance of its hourly employees. Our analysis has bearing not only on Amazon but also on the nature of work in the broader logistics sector (Bonacich and Wilson 2008), in which Amazon holds a dominant position. Our analysis holds theoretical value in that it seeks to address long standing debates within labor process analysis, as outlined further below (Burawoy, 1979, 1985; Edwards, 1986; Mears, 2015; Wood, 2020, 2021a). Its practical import stems from the fact that Amazon represents a major site in the contemporary struggle for workers’ rights, as witnessed in recent union formation efforts and legislative measures aimed at regulating the firm's working conditions. Moreover, given the ubiquitous role that Amazon has come to play in everyday life, empirical research on its labor practices can help inform public debate about the future of work while sensitizing consumers to the working conditions that lie on the other side of the digital portal that is so effortlessly used (Hill, 2020).
The paper unfolds as follows. We begin with a brief presentation of labor process theory and the perspectives toward labor control it has advanced. Following a discussion of our research methods, we then discuss the case at hand, describing the nature of Amazon's operations. Next, we present the findings of our qualitative analysis, which make several interrelated points. First, we find partial support for critical accounts which portray Amazon as deploying a harsh and coercive workplace regime. Here, we focus on one prominent and coercive mechanism of labor control that we call techno-economic despotism, in which management polices worker productivity using a sophisticated algorithmic system that places workers at risk of termination if they fail to “make rate.” This mechanism owes much of its effectiveness to the company's ability to recruit workers who have been chronically exposed to precarious forms of employment and have few alternative sources of work–conditions that serve as force multipliers, intensifying the power of algorithmic control.
Second, our evidence suggests that Amazon's regime does not rely on coercive mechanisms alone; rather, it has developed a multi-faceted system that also seeks to elicit workers’ consent to managerial authority. We identify at least three hegemonic mechanisms that supplement the company's coercive controls and that resonate with groups at various positions in the labor process. These include normative control, in which the firm selects and rewards workers who invest in their identity as “diligent workers,” often assigning them to positions that confer a “trusted worker” status on them; relational control, which fosters a sense of obligation among workers who are aware of the low esteem in which their labor is held and are thankful to the firm for hiring them, as if the job resembled a gift; and, following Foucault (2008), governmental control, which articulates a system of rules and resources that interpellate the workforce and imbue workers’ positions with a semblance of individual choice. At Amazon, a multi-faceted array of labor control mechanisms exists, with distinct mechanisms variably applied to workers holding different positions in the firm's operations enabling management to marginalize worker resistance and maintain a hegemonic bloc of employees willing to comply with company norms. Stil, there are signs that management's armor has begun to exhibit certain cracks. Changes in labor market conditions, cultural shifts affecting the tenacity of the “work ethic,” and instances of regulatory intervention are likely to affect the potency of the labor control mechanisms we identify here.
The Question of Managerial Control
Labor process theory has been at the center of much theoretical discussion in the sociology of work. Adherents of this view have often used historical and ethnographic methods to refocus conventional analysis of the management/labor relation. A key insight stemmed from this approach's inversion of the scholarly tendency (especially evident in the Human Relations tradition) to view restriction of output as irrational (Roethlisberger and Dickson, 1939; Roy, 1952; Vallas, 2012). Rather than asking “why do workers resist management's authority?”, as generations of industrial sociologists had asked, labor process scholars have flipped the script, asking why workers so often comply with management's directives (Burawoy, 1979, 1985; Mears, 2015). In this view, the purchase and sale of labor power is a most peculiar transaction. Although capitalists gain formal ownership over the labor power they buy, they are confronted with the fact that workers retain possession of this commodity and typically resist efforts at its exploitation. This infuses an element of uncertainty into the wage labor bargain and prompts capital to develop class-based systems of control and surveillance that involve far more than the exercise of unity and coordination. The question labor process theory pursues is how management seeks to accomplish this goal, thereby ensuring capital's ability to extract the full value of the labor power it has bought.
There the unity of labor process theory ends, however, as scholars have engaged in a long-running and often fractious debate over how best to address the question of labor control across different periods and sectors of capitalist societies. While early theorists emphasized the role played by coercive mechanisms such as Taylorism and the correlative process of de-skilling, later formulations stressed the structural conditions that enable capital to elicit the consent of workers to managerial authority. Burawoy (1979) provided the most explicit instance of this shift, arguing that the rise of labor unions and welfare state provisions forced capital to rely more heavily on hegemonic rather than despotic mechanisms of labor control. Edwards’ (1979) formulation was more complex, in that it traced a broad historical arc that led from simple control (based on personal supervision) to technical control (reliance on machine pacing, as with the assembly line) to bureaucratic control (which used internal labor markets and formal rules and policies to elicit worker commitment to the firm). 2 This emphasis on consent was equally evident in Kunda's study of a Cambridge software firm, which unearthed a potent system of “normative” control, in which management developed an ensemble of rituals, narratives and beliefs that fostered “an escalating commitment to the corporation and its definitions of reality, coupled with a systematic and persistent attack on the boundaries of their privacy” (1992: 224). Subsequent studies explored the ways in which firms managed to shape worker identity, engulfing them in a subtle form of control (e.g., “liberation management”) whose seductive powers stem from their egalitarian appearance (Fleming, 2014; Ross, 2004).
More recent efforts to advance labor process theory have moved beyond this emphasis on worker consent, arguing that shifts in the wider political economy (e.g., capital mobility and reduced state intervention) have led capital to reduce its reliance on hegemonic controls, instead resorting to other, more coercive or despotic ones in their place (Burawoy, 1985; Wood 2020; Crowley et al., 2020; Hatton, 2020; for an important exception see Sallaz, 2015). Burawoy himself made this argument when he introduced the concept of hegemonic despotism, in which capital has turned the tables on labor and forced workers to make collective concessions to the firm. More recent efforts to trace this despotic turn have emphasized either precarity or else technology. In the first of these accounts, capital's dismantling of the Fordist work arrangement has enabled firms to externalize any obligations to their workers, in effect using their structural power in the labor market as a potent lever of labor control (Hatton, 2011; Kalleberg, 2018; Lorey, 2015; Pugh, 2015). In a second version, theorists have built on longstanding views regarding the disempowering effects of digital technologies (e.g., Noble, 1984; Thomas, 1994), arguing that the spread of algorithmic management has enabled capital to tighten its grip on the labor process. This trend is explicitly manifest in platform work, where quantitative metrics, GPS technologies and surveillance systems lead toward “algorithmic despotism” (Griesbach et al., 2019; cf. Glavin et al., 2021). But as such technologies are adopted even by firms in the conventional economy, similar effects may spill over beyond gig work (Wood, 2021b), not least because firms such as Amazon operate in both domains. Though some case studies report exceptions to this disempowering trend (Shestakofsky, 2017; Wu, 2020), research using national data (Kristal, 2013, 2019, 2020) suggests that digital technology increases the flow of information to strategically important managerial groups, enabling them to expand their structural power over labor in various ways. This point is often reaffirmed by studies of call centers, where digital technology is central to the labor process (see Taylor and Bain, 1999 and below).
In his recent studies of the retail sector, Wood (2020, 2021a) has creatively combined these two versions of the despotism thesis, stressing the importance of both digital technology and precarity in shaping post-Fordist workplace regimes. Constructing a comparative study of two prominent retail chains in the US and the UK, Wood finds that management increasingly uses algorithmic systems to predict variations in consumer demand, enabling management to adjust staffing levels on short notice, despite the hardships such methods impose on workers’ lives. Such just-in-time scheduling practices endow managers with a highly coercive lever of labor control (“flexible discipline”), since they can cut the hours and earnings of workers they deem insufficiently compliant. Yet supplementing this coercive mechanism is a second element that relies on more hegemonic means in which managers selectively accommodate workers’ needs. Citing anthropological theory regarding the gift relation, Wood argues that this second mechanism (functioning as “schedule gifts”) encourages workers to misrecognize the nature of managerial power. Workers who receive such schedule gifts (including Wood himself during the course of his research) feel obliged to reciprocate, showing their thanks by conforming to management's expectations (see also Mears, 2015). The result of these mechanisms is what Wood dubs “flexible despotism” (cf. Chun, 2001), a regime that enables management to achieve Buraroy's couplet: the securing and obscuring of surplus value.
Although studies in the labor process vein hold great value, scholars have repeatedly wrestled with a number of issues that have bedeviled existing research, such as the precise relation between labor control and worker resistance (Ackroyd and Thompson, 1999; Knights and McCabe, 2000), the conflicting assumptions made by materialist and post-structuralist formulations (Thompson and O’Doherty, 2011; Willmott, 1990) and the bearing of ascriptive inequalities on labor control (Cockburn, 1983; West, 1990).
While acknowledging the importance of these and other issues, we wish to highlight a different yet related set of concerns: the assumption that managerial control functions via the application of a single well-integrated regime that envelops the entirety of a firm's or industry's workforce (Edwards, 1986: 2–3; 41; see also Storey, 1985). This view is especially evident in the work of Burawoy (1979, 1985), and more recently in Sallaz (2015); it has also been explicitly defended in the above-cited studies by Wood (2020, 2021a). In response, we argue that such approaches tend to elide the complex and often contradictory nature of managerial control systems (Vallas, 2003a). Even the most powerful and influential of firms is likely to rely on a form of bricolage as its managers set about controlling workers who perform distinct functions and exhibit different orientations toward the firm. Only by recognizing the internal complexities and differentiated arrangements that characterize workplace regimes can we begin to identify the sources of worker disempowerment. Only then can we anticipate possible shifts in the efficacy of management's labor control strategies, better informing efforts to challenge the subjugated position in which workers have been held.
The Company and its Warehouses
Founded as an on-line book seller in 1994, Amazon now enjoys a $1.7 trillion market capitalization (Kenney et al., 2021). It has achieved a massive presence in myriad economic domains, combining them in ways that multiply its market advantages. The firm controls a massive share of the world's internet infrastructure through its Amazon Web Services. It developed Amazon Mechanical Turk in 2005, one of the first crowdworking platforms (Berg, 2016). It now offers Prime subscription services that include video streaming, producing much of its content in-house. These achievements are in addition to the control Amazon enjoys over the retail sector. An e-commerce giant, Amazon can unilaterally set the terms for third-party sellers in the 18 countries where its Amazon marketplace services operate. Its $15 hourly wage is not unlike Ford's famous five-dollar day (Meyer, 1991) in that the wage itself is part of the socio-political mechanisms that define its especially potent production regime, as we shall see.
Perversely, Amazon has greatly benefitted from COVID-19, which has swollen the company's revenues, profits, and user base since early 2020. Responding to the public's reliance on e-commerce, the company leveraged its strengths as a logistics powerhouse, increasing its profits by an extraordinary 50 per cent, from $5.2 billion in Q2 2020 to $7.8 billion in Q2 2021 (Amazon.com, 2021). Anticipating further growth, the company hired more than 500,000 new employees during the second half of 2020. Its status as an employer of the second largest private workforce in the United States (behind only Walmart) is deceptive, since the company also employs an outsourced delivery workforce. This includes Amazon Flex, a digital labor platform staffed by on-demand contractors, as well as the nominally independent “Distribution Service Partner” firms that Amazon conjured into existence in 2016 (Alimahomed-Wilson, 2020).
The firm's labor practices have garnered critical coverage at the hands of journalists and public commentators, who have drawn attention to its extraordinarily poor safety and health record, its extreme intensification of labor, its high turnover rate, and its anti-union animus (Green and Alcantara, 2021; Gurley, 2020; Human Impact Partners, 2021; Kantor, Wiese, and Ashford, 2021a; Roosevelt, 2021). In 2018, when journalists reported that a substantial share of Amazon employees relied on food stamps and Medicaid, the company sought to address its growing legitimacy problem by raising its hourly wage from $11 to $15 an hour. Several efforts to unionize Amazon's warehouse employees have recently emerged (with one major success at JFK8 in Staten Island NY), publicizing the firm's Achilles heel –the overbearing intensity of its labor process. Public criticism has been especially important in that it has paved the way for direct government intervention, an example of which is California AB 701, which bans injurious production quotas by warehouse companies (Hussain, 2021 and Roosevelt, 2021).
Yet despite the attention Amazon has garnered, social scientific research on its labor practices has been remarkably scarce (exceptions are Alimahomed-Wilson et al., 2020; Delfanti, 2021). We therefore ask: How has Amazon sought to institutionalize its control over the workers it employs? What mechanisms of labor control has it deployed in its effort to limit worker resistance to its authority? And what implications does Amazon's regime hold for workers’ ability to organize on their own behalf?
Methods
Our findings are based on data collected over a fourteen-month period beginning in June 2020. Part of a larger mixed-method project on Amazon's logistics workforce, we conducted a national survey of warehouse workers and in-depth semi-structured interviews with both warehouse workers (our concern here) and Amazon Flex drivers. Utilizing the sampling method developed by Schneider and Harknett (2019; see also Griesbach et al., 2019), we recruited respondents using targeted ads on Facebook. Using images of Amazon warehouses, our ads were aimed at workers whose profiles indicated that they worked for Amazon and resided in or near zip codes adjacent to Amazon warehouses. The survey instrument asked respondents to indicate their interest in being interviewed, allowing us to follow up. We also used ad campaigns to recruit interviewees directly, running ads for Facebook groups in both English and Spanish and conducting interviews in both languages. Interviews were conducted via telephone and lasted roughly an hour; workers were offered compensation of $50 to $75 depending on when they were recruited. The final data included 46 interviews with warehouse workers, 25 additional interviews with Flex drivers, and 741 valid survey responses with Amazon warehouse workers nationwide.
Topics in our interviews covered a range of themes, including the nature of workers’ tasks, levels and forms of supervision, and the discipline they encountered, as well as their employment histories, exposure to financial insecurity, health and well-being, and the fairness of the company's labor practices. We also asked workers to share their perspectives on the episodic labor protests that have taken place at Amazon warehouses and on unionization. Interviews were transcribed and coded using nVivo 12. The coding scheme began with themes emphasized by labor process theory. This was augmented through an inductive effort to identify themes that were important to workers beyond those which theory had stressed. As the coding proceeded, the authors repeatedly met to modify and refine existing codes on the basis of emergent patterns.
Demographic data on our sample are presented in Table 1. Although our sample is highly diverse by ethnicity, age, and gender, its composition does call for caution in at least two respects. First, our sample under-represents workers of color. Though the company's blue-collar workforce is predominantly Black or Latinx (Long, 2021), workers in these groups comprised only a third of (15 of 46) of our sample. This bias was also true of our survey and may reflect greater distrust of researchers among this portion of Amazon's workforce. If so, our data are likely to be skewed toward more favorable accounts of Amazon (though our data suggest that the bias here is relatively small). 3 The second limitation stems from our use of individual-level data, which means that we cannot capture phenomena at the meso- or organizational level of analysis. We have done our best to use existing accounts of Amazon's operations to fill in the blanks. We note, finally, that the cross-sectional nature of our data makes it difficult to capture the unanticipated turns that labor struggles often exhibit (Fantasia, 1988). In our conclusion we comment on recent efforts at union formation among Amazon employees. In the following text all names of individuals who are quoted are pseudonyms.
Characteristics of the Sample.
Findings
Amazon's logistics system is comprised of a complex array of airport hubs, fulfillment centers, sortation warehouses, and delivery stations that are scattered across the country. Many are situated in exurban areas, on the perimeters of large urban markets (for example, there are 50 facilities surrounding Philadelphia alone; Blumgart and Hetrick, 2021). Fulfillment centers are the major nodes in the company's distribution network. These are massive, multi-floor establishments equipped with loading docks and conveyer belt systems that rival public transit systems in their complexity. Fulfillment centers can employ as many as 5,000 workers in a variety of jobs. At receiving docks, workers labor in small crews, unloading tractor trailers one after another, breaking down pallets of boxes and loading them onto conveyer belts, scanning each package and sending it to stowers, who then store each item in random locations, digitally marking each location for optimal retrieval at a later point. Pickers fill customer orders, either by walking to their storage points or (in more advanced fulfillment centers) retrieving them from robot-driven pods and placing them in plastic totes. Finally, packers box items at the outbound end, where machines label them and send packages to sortation centers and delivery stations for their final delivery to customers.
Amazon's workers take their places in an elaborate hierarchy. At the bottom are workers in Tier One positions, the largest category of hourly employees. These workers are differentiated in several respects. Most workers in fulfillment centers are hired on a permanent basis and hold “blue badge” status, making them eligible for the company's benefits “from day one” (a company-favored phrase that workers often used). Temporary or seasonal workers hold “white badge” status and have few if any rights or benefits. A further distinction runs between workers who perform direct and indirect functions. The former category, which includes stowers, pickers, sorters, and packers, have the most immediate impact on the flow of packages through Amazon's system. Direct function workers typically use digital scanners in their daily tasks and their expected volume of work reaches astonishing levels. Workers who perform indirect functions in quality control, training, or “non-con” (jobs handling non-conforming or oversized packages) also labor intensely, but encounter few of the pressures faced by workers in direct functions, especially in fulfillment centers serving large urban areas. A further division stems from the chain of command, for Tier One workers are overseen by Process Assistants (PA's) ranked at Tier 3. Above PAs are assistant managers, who oversee specific departments and in turn answer to Area Managers, who oversee entire buildings. The company fills supervisory and managerial positions partly from within (with provisions for internal labor markets) and partly from the wider labor market. Workers with military backgrounds seem to be favored by the company for positions at various levels in the hierarchy.
Journalistic accounts of Amazon's labor process have emphasized the unrelenting physical demands and bodily risks that workers encounter in their working lives, and our research confirms this point. One recurrent theme in our interviews was that of sheer exhaustion. Apart from short breaks amounting to one hour per 10 or 12 hour shift, full-time workers must remain on their feet for the entirety of each working day, walking, lifting, reaching, bending, or kneeling in physically awkward positions. When we asked Rick, from Pennsylvania, how the job affects workers’ bodies, he said: Oh, it breaks them down. It just breaks them down. […] You're not allowed to sit down at any time, other than break. You're constantly on your feet. You're walking on a hard concrete surface. You're constantly lifting. There's times when I went home from there and wonder if you're going to be able to wake up the next morning to go do it again. [And] I’m in actually really good shape… (58, white)
Exhaustion and the risks of injury were compounded by the pandemic in many respects. Given the public's increased reliance on e-commerce, many workers felt pressured to work at a far greater pace than before. (As one worker put it, “it's Prime Day every day now.” 4 ) Workers also confronted the risk of infection (at least five had themselves contracted COVID) despite the protective measures the company took. Indeed, one prominent theme involved resentment and confusion at the company's handling of information concerning internal outbreaks of the virus. The company began sending texts to workers when verified cases of COVID-19 occurred but these were often delayed by weeks and omitted key information such as where and when the cases had occurred, leaving workers alarmed or confused. Though the company temporarily increased wages and modestly expanded break times during the pandemic, many workers expressed resentment that more had not been done, viewing the company's efforts as little more than PR or efforts to forestall litigation.
Given the company's relentless demands on workers’ bodies, the risk of injury, and the company's fraught handling of the pandemic, one might expect workers to engage in fervent resistance against the company's practices, particularly in light of the company's rising profits and the seemingly unlimited wealth enjoyed by its founder, whose fortune was a frequent flash point in our interviews. And indeed, resistance has periodically surfaced, as with the COVID job actions at the outset of the pandemic (Dzieza, 2020) and the unionization efforts mentioned previously. Not surprisingly, our interviews unearthed many expressions of discontent and contempt for the company's treatment of its workforce. Instances of overt resistance to management's practices did emerge as well. One worker ripped down instruction sheets her manager had posted throughout her department, challenging the company's effort to implement a speed up. At team meetings, workers reportedly contested specific management rules or decisions (prompting some managers to end this ritual entirely, citing social distancing). We heard one instance in which a worker intentionally angled packages in such a way as to bring the line down, earning a bit of rest as a result.
Yet such instances of overt defiance were extremely rare. Pressured to keep working under demanding and often unsafe conditions, workers have for the most part been unable to mobilize in their own defense. The question then is why. What mechanisms has Amazon's workplace regime deployed and how do they limit workers’ capacity to resist the exploitation of their labor?
One obvious argument is that Amazon's wage and benefit package is sufficient to ensure worker compliance. While this is surely part of the story, many workers viewed their compensation levels as insufficient, especially in view of the extraordinary work effort the company demanded and the enormous wealth it enjoyed. We view the wage and benefit package as the material underpinning of the production regime that Amazon has developed, the broad elements of which are spelled out below, moving from its despotic to its hegemonic parts.
A second factor involves the high levels of turnover at Amazon's establishments, which make it difficult for workers to cultivate or share counter-hegemonic views. It is true that workers exhibited relatively thin ties to one another, owing to the constant churning of the workforce (and to the noise and frenetic pace of their jobs as well). One worker commented that he no longer bothered to learn his co-workers’ names since so many of them would be gone in a week or two (“I call them ‘Hey you, person.’”). COVID-related social distancing (which impeded the use of break rooms) compounded this fact. In our view however the importance of turnover is that it accentuates the individualizing tendencies established by the array of labor controls the company has established.
Techno-Economic Despotism
Labor process theory has increasingly emphasized capital's growing use of coercive labor controls but with few exceptions, it has treated two drivers of this shift –technology and precarity— separately. In line with Wood (2020, 2021a, 2021b), our data suggest the need to reconsider that assumption. Indeed, our interviews speak to the confluence of technology and precarity, yielding what we term techno-economic despotism, in which powerful algorithmic systems gain traction on workers precisely because of their vulnerability in the labor market, leaving them with no apparent alternative but to comply with management's production demands.
Although most of Amazon's workers hold standard work arrangements (full-time, permanent jobs with benefits), they are no strangers to economic vulnerability. As in the warehouse industry more generally (Gonos and Martino, 2011; Gutelius, 2016), the company recruits populations with chronic histories of low wage, insecure and sometimes abusive forms of employment. Our interviews included older workers who had exhausted their retirement funds and had struggled to find any work; former gig workers who had faced wildly unpredictable earnings; former workers in fast food and retail stores who suffered from a lack of benefits; former stay-at-home moms whose “resumes look like crud” (as one put it) because of their lack of paid employment; immigrants seeking refuge from factories that offered even harsher treatment than Amazon's; and workers with college and even Master's degrees in theater, journalism, sociology and other liberal arts fields who could not find work using their degrees. Reflecting the pandemic, the sample also included many workers whose jobs had evaporated, leaving them no alternative but to work during the height of the pandemic. Few of our subjects felt at all financially secure; only six of the 46 had enough savings to cope with a $1,000 emergency should the need arise. 5
A few examples will illustrate the point. Rick, the Pennsylvania man quoted above, once owned his own small business but had fallen on difficult times and knew that age was no longer his friend. “I'm old…You know, you're kind of, you're kind of like, I have no choice. You either work to survive or you, or you die” (58, white). Kathy, a Southern California woman who had left the labor force to care for her ailing mother, resorted to Amazon, which “was the only place I could get a job because I hadn’t worked in so many years” (59, white). Chaz, a full-time sorter in New Jersey who drove for Uber previously, observed that “It used to be basically the only time I ever paid my bills was when they were threatening to shut stuff off” (46, white). Workers made frequent reference to the ease with which they could be replaced and the gnawing fear that they might fail to “make rate,” putting their jobs at risk. Bob, working in Miami, described feeling too insecure in his job to use the sick days he is formally allotted. “I’m in no position to lose my job right now…I’m afraid that if I take sick days or if my job performance falls, I'm going to be let go” (20s, Latino).
Crucially, when workers describe their fear of being fired, they almost always reference the company's sophisticated system of digital or algorithmic controls, which management uses to police each worker's productivity. This is most pronounced among workers performing direct functions, who use scanners to process each item they must handle. These input devices enable management not only to track the precise location of items in the system but also to generate a detailed, real-time record of individual workers’ performance levels, which are automatically compared against the expected production rate and against other workers in the department. The system automatically flags workers with deficient performance records, who are then vulnerable to “coaching,” as pro-company workers called it, followed by formal warnings, and eventually termination. 6
Management uses two distinct measures of individual performance. The first –Units per Hour or UPH—measures the speed at which workers operate. Each worker's rate and percentile ranking are prominently displayed on monitors and hard copy print outs in many production areas, the better to foster a sense of accountability, discipline, and competition among workers. So powerful is the monitoring system that at certain junctures the scanners even prompt workers to pause and to engage in brief meditation or stretching exercises that don’t count against the worker's performance metric. UPH functions as an overbearing reality for many workers, who must monitor their own numbers at their work stations to ensure they can “make rate” by the end of their shifts.
The second performance metric, Time off Task (TOT), records the number of minutes during each shift when workers have shown no measurable productivity. The system normally allots workers a five-minute period during which inactivity is not flagged. Beyond this, the digital clock starts ticking, measuring the number of minutes each worker has failed to engage in productive activity. TOT levels amounting to an hour or more can be cause for summary termination, but smaller levels are also cause for discipline. Supervisors speak in almost admiring terms of legendary cases in which workers racked up unimaginably high TOT numbers, e.g., by hiding in hallways or bathrooms for hours at a time or even sleeping in their personal cars. Such tales aside, the mundane reality is that the TOT metric heightens workers’ sense that they are digitally tethered to their jobs and must constantly keep working.
Especially when applied to an economically vulnerable workforce, the company's algorithmic system for policing productivity imposes a powerful mechanism of labor control. While the classic assembly line worker could at times devise ways of carving out spaces of partial autonomy (by working ahead, doubling up, or infusing work with elements of collective play; see Hamper, 1991 and Milkman, 1997), Amazon workers lack such possibilities, not least because they can be held individually accountable to a greater degree than was true of assembly line workers in the past.
7
This is why workers like Stephanie can quietly express their indignation at the coercion they encounter on the job. A mother of three children, she said of the job “it's a slave trade…or a sweatshop actually” (40 s, white). Or why Sara, the 42 year old South Carolinan, can say that Amazon is ridiculously strict…It's as if you are held at such a level of perfection that if you make one minor mistake … [pauses]. You're human, you're not a robot. You can instantly be written up and then you can't do projects, you can't apply for other jobs, you get frustrated. It feels like you're constantly under pressure to be perfect all the time… Every day, every moment, every decision you make you worry that you're going to lose your job.
There is a resemblance between certain direct function jobs at Amazon and call center work (see Callaghan and Thompson, 2001). The latter sector uses Automatic Call Distribution systems, for example, routing jobs to workers in much the same way that orders are sent to pickers or stowers in Amazon's warehouses. Performance levels are carefully monitored in both settings, as the literature makes clear (Taylor and Bain, 1999). Putting aside debates over the rigors of call center work, a key difference is that it involves direct interaction with customers, rendering such work more complex and unpredictable than the handling of inert packages (Korczynski, 2009). That fact, coupled with the economic precarity that low-wage labor markets have imposed on Amazon's workforce, works to heighten the power that techno-economic despotism typically wields.
Hegemonic Mechanisms of Labor Control
Despite the power of techno-economic despotism, our data indicate that it would be a serious mistake to assume that coercion is the only labor control mechanism on which management relies. Evidence to this effect was apparent in a number of ways. First, a substantial number of the workers in our sample –especially those performing indirect functions— expressed little or no fear of termination or stress over making rate. Second, even workers who had experienced serious injury were often willing to overlook their exposure to bodily risks, concluding that the job provided a viable and advantageous bond. Indeed, it was not uncommon to hear workers challenge the notoriety the company had acquired for its labor practices, suggesting that their own experience had been far more favorable. While techno-economic despotism was a real and abiding source of labor control, it shared organizational space with other mechanisms that act to foster consent to managerial authority.
We identified at least three analytically distinct but empirically overlapping mechanisms –normative, relational, and governmental controls—that seemed to secure worker compliance by eliciting their consent. The first two rely partly on conditions rooted outside of the firm as such –a source of embeddedness (Wood, 2021a)-- articulating with them in ways that legitimate managerial domination. With normative control the firm can capitalize on workers who adhere to the work ethic, reinforcing it via structural arrangements that provide workers with important symbolic rewards, thus encouraging them to invest (as some do) in what we have termed a “diligent worker” identity. Similarly, relational controls hinge their efficacy on the dismal state of the low-wage labor market, with workers internalizing the poor value of their labor power and exhibiting a sense of gratitude and obligation toward the company for its hiring of them. The third mechanism functioned more independently, operating as a formal governance apparatus that interpellated workers as individual subjects or “industrial citizens” (Althusser, 1971; Burawoy, 1979: 109–113). Despite these differences, all three mechanisms are similar in that they help the company engage workers, leading them to adopt behaviors and orientations that align with management's goals.
There are of course practical reasons for workers to invest in diligence, since doing so can provide a shield against company discipline. Indeed, workers who felt that they were secure in their jobs commonly explained that they had earned favorable reputations as trusted and competent workers and were therefore unlikely to be let go. Yet workers’ investment in diligence was more than purely instrumental. Consider the case of Seline, a former realtor working in Florida who held an indirect job. When we asked whether the time drags for her at work, she indicated that it does not. For me, I oftentimes get all the work done that's immediately in front of me, and then I seek out further tasks, based on the talents that I've acquired in the last almost six years being there. I jump in and help out wherever I can… If there is a little bit of problem-solving that somebody needs help with, I ask associates, “Hey, do you need a barcode printed?” Or “You look like you need something. How can I help you out?"…I pay attention to my surroundings, so I know a lot of what needs to be done in that building. (44, white)
Seline uses the company's preferred lexicon (“associates”) and takes pride in her ability to play the “problem solver” role, wearing it as a badge of honor.
Like Seline, Nathan is proud of his work record. A young veteran of the war in Afghanistan, he makes it his business to exceed the company's production rate. He also takes pride in being assigned the more difficult tasks in his building. Actually my daily rate is about 130% over [the company norm]. I'm a pretty driven person at work, so they know me as that. And I've been doing it for three years constantly at that…Usually, what they have me do, and I kind of volunteer for, is I'll take the [tasks] with the harder product, because I know the building so well. And they know I can make that rate. Where they have some of the people that it's kind of harder for, they'll typically send them a little bit easier stuff. (30, white)
“I like my job. And for me, it is competitive. Do you understand me? I try to be on top. It’s my daily goal to always be on top. To be number 1, or 2. To always be there. The more I put [out], the more I get excited about it, and I love my job more [Laughs]” (47, Latina).
Magda watches her co-workers during the working day and checks her own UPH constantly, making sure she remains at or near the top. Estefania (40, Latina), working as a packer in Arizona, exhibited much the same outlook as Magda. When she transferred from one warehouse to another, she made it her business to learn who the top performers were in her new department, the better to compete with them and stay at the top. 8
Normative control represents the company's effort to establish job structures and rewards that encourage workers to invest in the diligent worker identity. By rewarding workers with job assignments that confer a modicum of recognition, the company manages to elicit active consent, encouraging a stratum of workers to align their identities with the company's goals.
Workers exhibiting this view commonly defended the company against workers who criticized management, whom they viewed as ungrateful or immature. Marie, who had worked in both New York and a Midwestern facility, epitomized this view. Well, I know what people are saying. [They say] “Oh, well, Jeff Bezos doesn't care about us.” Well, I don't know what they expect… I was able to get a job even though I had a gap in my employment history. I was treated fairly on a company level and they've gone out of their way to do things [for me]. When I had a crisis, they helped me. When I had to relocate, they did a hardship transfer. (50, Black)
Kathy, the California woman quoted earlier, was saddled with massive credit card debt and behind in her property taxes. She was thankful that the company had provided her with a job when no one else would, even though she admitted she had suffered serious injuries on the job. “I think we're treated fairly. We're very fortunate to have jobs… I think people are just complaining because they don't like the way things are, because they're not doing their jobs and they're lazy”. Meanwhile, Jessie, a middle-aged woman in Dallas, appreciated the company's willingness to vary her shifts, allowing her to care for her disabled mother. She frames her work effort using a logic of reciprocity. “Whatever Amazon expects from me, I expect from Amazon. If they expect for me to do my job, I expect for them to do their job with me, accommodating me. You give what you get in return, so Amazon is a great company” (55, Black).
Amazon has a well-deserved reputation for its harsh labor practices, but our interviews suggested that this is not the only face it shows toward its workers. Since much of its workforce has encountered high levels of hardship in a turbulent labor market, providing even a modicum of accommodation for some workers suffices to win their allegiance in ways that provide a bulwark against labor unrest.
The importance of the company's mobility provisions was somewhat surprising to us, given the corporate trend toward the dismantling of internal labor markets (Cappelli, 2001; Sallaz, 2015). Our interview schedule made no mention of mobility, but workers often discussed it, attaching much importance to their hopes of transferring to less onerous jobs or facilities or of being promoted to a Tier 3 position as a PA. Such workers tended to fasten on individual mobility and achievement as features of the company that they particularly liked. August, a young worker from Louisiana had long sought a promotion and had finally succeeded. “I have this feeling that I could move up in a higher position and keep making my way up this company” (33, white). Marie, mentioned above, worked at JFK8 in Staten Island when a COVID strike occurred. She told us that “Amazon gives everyone a chance to prove themselves…You can work hard, you can take on more responsibility, and prove that you can move up”. Chaz, from New Jersey, explained that “figuring out promotions and how to get ahead” requires being “a self-learner, a self-starter, you got to be motivated to do it”. All three of these workers expressed their frustration with the company's critics, defending the company despite all the bad press. The importance of promotions was also apparent in the many complaints we heard from workers who had been passed over for promotions and who lamented the company's willingness to hire supervisors from outside Amazon itself. Such complaints only served to underscore the value that workers attached to the company's provisions for mobility, which functioned as something of a safety valve, reducing conflicts and tensions by supporting workers’ aspirations for improvements in their individual work situations.
Even more significant were the company's provisions regarding working time, which seemed to serve multiple functions for workers and the company alike. It is important to note that logistics companies often confront unpredictable fluctuations in consumer demand which, coupled with high fixed capital costs, generate even greater pressures for temporal flexibility than Wood (2020) found in his study of two retail chains. Simply externalizing the risk of overstaffing often invites industrial conflict, as shown in a recent study of warehouses (Dörflinger et al., 2021). Rather than adopt the punitive approach that Wood (2020) uncovered in retail settings (which he dubbed “flexible discipline”), Amazon has creatively adopted provisions that define variability in workers’ schedules as a reward –and one that workers highly value since it enables them to escape the relentless demands of their jobs.
Each three months the company provides workers with an allotment of 20 hours of unpaid working time (UPT) which they can “spend” to gain a measure of freedom over their working schedules. Workers can use their UPT allotment to shield themselves from discipline if they arrive late or cannot attend work. They can also use their UPT balance when management issues a call for Voluntary Time Off (VTO), at which point management allows a certain number of workers to end their shifts early. Doing so reduces the workers’ wages, but VTO opportunities are so highly valued by workers that management can distribute them via a lottery (as if they were prizes to covet) or else bestow them on workers who have shown the best UPH rates for that shift. In this way management has framed a profit driven imperative –its need for temporal flexibility—as a benefit that workers can choose to invoke. Sara, the South Carolina worker, speaks for many when she says: [When] they throw out VTO, or voluntary time off, people will seize, like they lose themselves, they're like, “Oh my God, I got to go!” People just, they want out. You don't get paid to take VTO, you're shortchanging your paycheck, it's just so miserable people are just like, “I want to go!” They're happy when they get it and they run out the door.
Thus being eligible for transfers, promotions, and flexibility regarding one's schedule were highly valued parts of workers’ everyday working lives. Workers could hope to pursue individual solutions to their subjugated positions and find some sources of relief (however fleetingly) from the difficulties their jobs impose on them. Yet crucially, these job rewards were only available to workers who had maintained their status as “workers in good standing” –i.e., records that were free of any performance issues, warnings, or disciplinary infractions. Being written up thus meant that you were frozen in a liminal position without organizational rights, a stigmatized condition that workers –much like the prison workers in Hatton (2020)— eagerly sought to avoid.
Discussion
Scholars adhering to labor process theory have often embarked on a search for the predominant type of workplace regime that characterizes a given period in the development of capitalism. Such formulations, rooted in theories of social reproduction, have often tended to view workplace regimes as smoothly articulated systems whose elements fit together in mutually reinforcing ways, thereby accounting for the subjugated position in which workers are employed. With some exceptions (e.g., Friedman, 1977), they also tend to assume that a firm's workplace regime envelops the entirety of the firm's workforce in much the same way, enabling capital to both secure and obscure the existence of surplus value. Our analysis leads us to question these assumptions. While our findings partly confirm media reports that emphasize the company's highly coercive labor practices, these accounts also tend to oversimplify the realities of the company's labor practices. The apparent product of organizational bricolage, Amazon's workplace regime seems to lack the coherence that labor process theory expects. It exhibits an array of labor control mechanisms that mix hegemonic and despotic elements and differentially applies them to different groups and categories of workers, in accordance with managers’ needs.
Given workers’ exposure to physical exhaustion, safety and health hazards, the company's fraught handling of COVID-19, and the enormous profits it has reaped during the pandemic, it is reasonable to expect workers to challenge the many demands the company has imposed on them, even without recourse to union representation. To be sure, workers expressed much distrust and contempt for the company overall, but they engaged in few overt challenges to managerial rule. The question we have posed is precisely why.
The answer lies in a complex array of labor control mechanisms that rely on both coercion and consent. The most coercive mechanism is that of techno-economic despotism, an application of algorithmic technology whose efficacy is greatest among workers with few economic alternatives. This mechanism was most apparent among workers performing direct functions, whose scanners are tools in a double sense: Workers use them to process the items they handle, and management uses them and the data they generate to police the productivity of each worker, maintaining detailed real-time productivity data capturing both the speed and the continuity of the worker's performance. By quantifying and algorithmically monitoring compliance, ranking workers against one another and making these rankings broadly visible, management imposes the constant threat of discipline and termination on workers who fail to make rate. Understandably, workers without savings and few alternative sources of employment were made to feel chronically anxious about their positions, concluding that they had little alternative but to comply with the company's demands.
Yet this was not the whole story. Alongside despotic mechanisms we found forms of labor control that rested not on pressure, intimidation, or fear but on affirmation, accommodation, and interpellation. In the case of normative control, the company found it possible to rely on the wider culture's emphasis on the virtues of paid employment (Weeks, 2011). Leveraging this cultural emphasis, the company selected workers who invest in the “diligent worker” identity for movement into indirect functions that enabled them to enjoy a “trusted worker” status. Access to such symbolic rewards enabled management to cultivate many workers’ allegiances. The point is that job structures reflect not only technical concerns but also political imperatives –a point that was articulated early in the labor process debate (Marglin, 1974; Noble, 1984) but has been ignored of late.
Relational control emerged among workers whose blemished histories in the labor market left them with an enduring sense of gratitude to the company for hiring them despite their poor records. Here, hiring was construed less as a purely economic transaction than as a relational tie akin to a gift, implying the need for the worker to reciprocate with loyalty in one form or another. This is why workers in the thrall of relational control often viewed labor protesters as showing a lack of gratitude to the company for its willingness to hire them even when no one else would. This finding aligns with Wood’s (2020) emphasis on “schedule gifts” at two major retail chains (see also Mears, 2015).
Perhaps the most all-encompassing form of hegemonic control is the one we have dubbed governmental control. This mechanism is composed of organizational provisions that help Amazon to interpellate its workforce, defining workers as individual subjects who, should they satisfy the company's demands, enjoy important moments of individual agency. The firm's practices with respect to transfers and promotions were an important part of governmental control, fostering hope on the part of many workers that they might improve their job situations through individual mobility. Several of the workers in our sample had in fact been promoted to Tier 3 positions and a great many more were aware of such possibilities. More significant were the company's provisions regarding working time. By providing workers with a fund of UPT –essentially, a form of currency that can be ‘spent’ when management authorizes VTO, or voluntary time off— the company found it possible to adjust its staffing levels, achieving temporal flexibility not by imposing it on a potentially resistant workforce but instead imbuing it as a source of freedom, however temporary, that workers obviously prized.
Perhaps the single most important theme that emerges from this analysis concerns the need for labor process theorists (and for public critics of Amazon as well) to conceptualize workplace regimes in ways that acknowledge the complex and multi-faceted nature of the labor controls that workplace regimes can wield, enabling managers to apply sharply differentiated mechanisms to different groups and categories of employees. Indeed, Amazon's ability to subjugate its workforce stems precisely from the wide array of labor control mechanisms on which its managers can rely, equipping them with an elaborate repertoire of techniques that can be used to enforce their will. For example, if managers identify a given set of workers as insufficiently cooperative and thus poor candidates for “trusted worker” status, they can easily confine them to positions involving the most coercive form of labor control. Arguably, languishing in such positions implicitly encourages the attrition for which the firm is well known. By contrast, workers who seem more cooperative can be handled differently, using more hegemonic means that in effect include them in the manager's cadre of diligent or loyal employees. Such varied mechanisms of labor control equip managers with a diversified toolkit with which to enforce their rule. Since workers experience the wage labor relation in decidedly different ways, however, they naturally adopt different orientations toward the firm and, consequently, toward one another.
These differences emerged quite clearly in our interviews. Workers who seemed engaged by normative or relational controls often depicted their co-workers as lazy, inexperienced, or as refusing to pull their own weight (sometimes literally). Interestingly, workers subjected to coercive controls made fewer derisive representations regarding pro-company employees; instead, they viewed company-compliant workers as acting out of fear or insecurity rather than any moral deficiencies. The point is that the hydra-headed nature of Amazon's workplace regime operates differently across the workforce and sets in motion dynamics that almost inevitably segment or fracture the coalitions that workers would need to build if they were to challenge managerial imperatives. Divided against one another, isolated and individualized by the very design of their jobs and by divergent experiences of management's controls, workers are thrown back on their own devices. They cooperate with one another as their jobs require (even, our data suggest, across racial boundaries) but they fail to challenge the intensity of the demands the company places on their working lives –and their bodies.
Our analysis should in no way be read as indulgent toward Amazon's labor practices. On the contrary, we contend that the company's labor controls are all the more pernicious because they elide the uniformly abusive images that many activists have advanced. Though such images are partly true and may help to gain public support for workers, they also distort the realities that working for Amazon entails. To be sure, our interviews have unearthed a wealth of instances in which workers have suffered chronic injuries, some quite serious. They have identified stress-inducing surveillance and control systems that take a significant toll on worker well-being. A large proportion of workers feel treated as a disposable resource or as machines. Such workers harbor few illusions about the company's interests, the arbitrary and exploitative nature of which is rarely obscured. Yet such conditions exist side by side with starkly different mechanisms of labor control that instead elicit workers’ consent. This presents additional obstacles for workers to generate the trust, unity, and sense of shared fate that would be needed to challenge Amazon's regime. Indeed, Amazon's ability to elicit worker consent among a sizeable proportion of its workforce allows it to build a strategically important hegemonic bloc on which it can rely, thereby marginalizing and containing resistance from workers who encounter the most coercive controls.
Conclusion
Recent events have heightened public awareness of the central role played by workers engaged in our nation's logistics sector, the arterial systems that furnish the goods on which our daily lives depend. Yet few social scientific studies have targeted the core of logistics labor to date –the labor of warehouse workers, the assembly line workers of consumer capitalism. Fewer still have explored the hidden abode of Amazon, America's largest and most powerful retailer whose success rests in large part on its logistics prowess. The question this paper has posed is how Amazon has succeeded in limiting the capacity of its workforce to resist the extraordinary demands management places on them –demands that often place their bodies at great risk.
Drawing on interviews with a national sample of Amazon's warehouse workers, we find a more complicated pattern than labor process theorists typically expect. Rather than unearthing a well-articulated set of labor control mechanisms that function both to secure surplus value while obscuring its existence (Burawoy, 1979, 1985; Wood, 2020), we detect a multi-faceted array of mechanisms that seem only loosely coupled with one another. Indeed, their disparate nature suggests that they reflect a pattern of managerial bricolage rather than any systemically organized ensemble. Much as journalistic accounts have stressed, Amazon's workplace regime does indeed rely on harsh and coercive labor control mechanisms, fusing algorithmic technologies with the structural power that management enjoys over workers with histories of labor market precarity. Complementing these mechanisms, however, are other sources of labor control that are more subtle and sophisticated than naked coercion alone. Indeed, the efficacy of Amazon's regime seems to rest precisely on the broad and differentiated array of labor control mechanisms it has at its disposal and that can be variably applied to different groups and categories of workers as managers prefer.
In theory, the variegated nature of Amazon's labor control mechanisms might well introduce tensions and contradictions into its handling of its workforce. We have found little evidence of this, however, largely because the company's control mechanisms seem to be applied to workers performing different functions (e.g., those handling “direct” functions who are tethered to their scanners versus those performing supplemental or “indirect” functions, who rarely encounter techno-economic despotism). Indeed, the varied tool-kit that managers can wield seems to allow them to foster the kind of informal coalitions they need to maintain a compliant workforce. The result tends to informally fracture or even segment the company's workforce, as the company's workplace regime engenders different orientations among workers toward their jobs, management, and the firm. So powerful and multi-faceted an array of labor controls has thus far succeeded in subjugating the company's workers, bending them (and their bodies) to the meet the firm's preferences, which in effect view labor as a mere instrument toward customer satisfaction.
It is conceivable that research focused on specific Amazon establishments might find its labor control mechanisms to be unevenly distributed across the various facilities (e.g., fulfillment centers versus sortation centers or other smaller establishments), a point that would be informative in itself. Such research might also find a somewhat different set of control mechanisms. Regardless, the contribution of our typology is not one of mere taxonomy –as if the effort to identify labor control mechanisms resembled botany— but rather one that opens up questions relating to the imperatives the labor process imposes on workers under contemporary capitalism. For example, recent research has suggested that the search for temporal flexibility (to vary staffing levels to accommodate the on-demand economy) is pervasive in many industries (Schneider and Harknett, 2019), resulting in the imposition of coercive controls that what Wood has termed “flexible despotism.” Yet our discussion of governmental control–a type that seems to open up choices for workers, especially involving their work schedules—suggests that companies like Amazon can achieve temporal flexibility in perniciously creative ways, not only avoiding the conflicts that flexible discipline would provoke but even reducing conflict through the provision of Voluntary Time Off. Apparently the drive for temporal flexibility can assume a wide variety of forms.
A second point emerges from our discussion of both normative and relational controls. The former enables the company to harness the work ethic organizationally, via the adoption of job categories that select and reward workers who have invested in the diligent worker identity; the latter allows the company to capitalize on the low value employers has assigned to many workers’ labor power, leading them to feel obliged to Amazon for giving them a chance. As noted above, both these mechanisms operate in league with external conditions, and their efficacy may change as those conditions evolve. This point aligns with Wood's recent (2021a) call for labor process scholars to scrutinize the embeddedness of the labor process in its wider structural and cultural context, in effect transcending the longstanding tendency among many scholars to view the labor process as a relatively autonomous domain. This point seems particularly important now, as labor market conditions have begun to lend workers greater structural power than they have enjoyed in recent decades and when many workers are beginning to rethink the meaning of work – developments that may well reduce the efficacy of Amazon's labor controls. Surely, the efficacy of techno-economic despotism depends on the company's ability to access endless supplies of workers who have long histories of precarious employment. As the labor market improves, the company may no longer be able to fervently apply its Units Per Hour and Time off Task metrics. Arguably, these economic and cultural shifts in Amazon's environment help explain why cracks have begun to appear in the company's armor, as per organizing campaigns that have begun to succeed (Weise and Scheiber, 2022).
The typology we have introduced has at least one further implication. As the importance of digital technology for labor control has begun to attract increased attention, questions have emerged as to how to understand the simultaneous effects of competing variables. As noted earlier, with important exceptions, the tendency in much of the literature has been to view technology and precarity as orthogonal to one another; this is why the recent study of alienation among platform workers (Glavin et al., 2021) explored the effects of digital technologies while controlling for financial hardship. Yet our analysis of techno-economic despotism suggests a different view, in which the effects of technology interact with varying levels of labor market uncertainty. Indeed, in a subsequent paper on platform work and mental health, Glavin and Schieman (2022) find precisely such an interaction effect, in which the health effects of platform work depend on the labor market conditions to which workers are exposed.
One last question to be addressed concerns the broader relevance of our analysis. We began this study with an effort to reach into the hidden abode of logistics workers, whose importance for consumer capitalism has grown more palpable in recent years. Our choice of Amazon as a research site was guided by the company's extraordinary capacity to leverage its logistics knowledge and expertise (a form of power/knowledge that has equipped it to overtake Walmart, its predecessor). Yet as the logistics sector has grown in size and influence, and as the on-demand economy has expanded, we suspect that the labor controls that Amazon has devised may find their equivalents in industries either upstream from its warehouses (into manufacturing) or downstream (not only in retail but also in other services, such as health care). Future research is needed to explore this possibility.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research,authorship, and/or publication of this article: This research received generous support from Northeastern University's Office of the Provost.
