Reproducing disposability: Unsettled labor strategies in the construction of e-commerce markets

Continuous revolution

Since the logistics revolution began in the 1970s, the sector has not only continued to rationalize in pursuit of cost reduction but also restructured in order to produce new services and commodities (Allen, 1997; Cowen, 2011; Hall, 2004). Retail capital’s use of logistics to construct e-commerce markets is the latest iteration of this ongoing experiment. Implicitly drawing on Marx’s insight in Capital, volume 2, that transport is itself a commodity, e-commerce has slowly expanded from selling books to selling groceries and other necessities of everyday life (Cowen, 2014: 100–102; Marx, 1978: 225–229). This restructuring of “last-mile” logistics intensifies the commodification of transport as delivery becomes a consumer good and highlights the centrality of labor in these projects of market construction. Big box retail externalizes the physical work of shopping to consumers, who are required to transport themselves to retail stores, walk around the store, and collect the items they want to purchase. By contrast, e-commerce internalizes this physical labor and industrializes its processes in warehouse and delivery networks.

Despite the differences between e-commerce’s and big box retail’s organization of labor, warehousing has remained central to the logistics and market strategies of both models. Walmart’s distribution centers became famous for their cross-docking and continuous movement of commodities, while Amazon “fulfillment centers” have become notable for their innovative inventory storage and hyper-Taylorist methods of controlling warehouse workers (Bonacich and Hardie, 2006; Cidell, 2015: 20–22; McClelland, 2012). While distribution centers move pallets and cases through their warehouses relatively quickly, workers in fulfillment centers have to open up those cases in order to store and pick items at the individual rather than the case level. It is because of these changes in the requirements of warehouse work that warehousing has become a critical site of restructuring operational and labor processes for emerging e-commerce markets. These new requirements drastically increase warehousing’s demand for low-wage labor and the expectations of quality in their work, which is based primarily on accuracy when picking, packing, and sorting customers’ orders. Interrogating this restructuring of warehouse work for e-commerce markets, this article asks two main questions: How is warehouse labor being re-produced and mobilized to make these markets possible? And how are logistics workers being re-positioned and re-valued within production networks as logistics increasingly takes a lead rather than secondary role in capital’s accumulation strategies?

To answer these questions, this article interrogates the production of warehouse labor for e-commerce markets, and the tensions they produce for the sector’s widespread use of labor market intermediaries in the form of both temp agencies and third-party logistics companies (3PLs) (De Lara, 2013; Gutelius, 2015; Rowe, 2012). ¹ My research demonstrates that one strategy of warehouse managers in response to the difficulties from e-commerce has been to reduce their reliance on labor market intermediaries and directly hire more warehouse workers. Even though this shift in labor strategy brings some warehouse workers closer to the archetype of “standard” employment – often used as a benchmark for working conditions (Kalleberg, 2000) – workers’ disposability is reproduced in the process. This article identifies two mechanisms for reproducing workers’ disposability even when warehouses hire more of them directly; first, warehouse management mobilizes discourses of “unskilled” labor to degrade the work that is core to e-commerce warehousing. Second, management mobilizes conceptions of retention that focus on controlling workers’ movement in and out of warehousing jobs at a systems rather than individual level. This logistical practice of retention makes individuals highly replaceable even as management must make sure the workforce does not turnover too fast or at the wrong time.

To elaborate this argument, I first outline how e-commerce affects warehousing and creates demand for more low-wage labor and higher quality work. Drawing from research on the restructuring of last-mile logistics in central and northern New Jersey, I analyze two attempts by warehouse managers’ to address emerging problems with their existing labor strategies. In both examples, management adjusts to the new world of e-commerce in part by reducing their reliance on labor market intermediaries. In the first warehouse, managers mixed qualitative and quantitative forms of flexibility in order to eliminate their previous dependence on temp agencies and achieve the increased efficiency of workers demanded by e-commerce warehousing. These hybrid forms of flexibility served as an alternative to the re-valuation of core workers involved in flexible specialization and provided a means for management to develop a more skilled workforce while maintaining their disposability by shortening or lengthening the working day mid-shift. To understand this process, I build on feminist research in global production to consider how socially constructed ideas of skill influence the production of flexible labor.

In a second warehouse, management experimented with increasing the amount of employees they directly hired in an effort to improve their worker retention. At this warehouse, worker retention was low in large part because of the way that labor market intermediaries were organized in their facility. I argue that the idea of retention managers mobilized in this experiment is different from Fordist understandings. Management in this example were concerned with retention to the extent that they wanted to control the entire workforce’s short-term movement in and out of warehousing jobs. These managers were not concerned long-term investments in individual worker training or employment. This analysis builds on existing arguments that logistics’ rationalities equate workers with commodities and machines (Cowen, 2014: chapter 3; Kanngieser, 2013). Like commodities, workers’ movement in and out of warehousing jobs is being managed at a systems rather than individual level, elucidating the terms under which capital values and retains labor (Harvey, 1998; Wright, 2006).

Given the analytical weight that scholarly work has placed on the subcontracting relationship in degrading workplace standards, one might expect that workplace conditions would improve with lead firms employing more workers directly (Tsing, 2009; Wills et al., 2009: 6). However, my research demonstrates that without effective worker empowerment, workers in the supply chain remain disposable even if they move closer to “standard” forms of employment. Within logistics research, this insight develops our understanding of flexible labor’s production within an increasingly crucial sector of the US economy. For labor geography, the reproduction of disposability without labor intermediaries is important because it raises questions about whether subcontracting relationships are as ubiquitous as assumed, and hence whether subcontracting can be adequately understood as the “paradigmatic form of employment across the world” (Wills et al., 2009: 6). From an organizing perspective, a detailed analysis of subcontracting is significant because policy and activist work on warehousing has often used a commodity chain approach that prioritizes mapping and organizing strategic relationships between contractors and lead firms (Cho et al., 2012; Quan, 2008; Rowe, 2012). This analysis has been important in the organizing strategies of groups like Warehouse Workers United, which has won legal battles that gained joint employment status for temp workers at a subcontracted Walmart warehouse in Riverside California (De Lara et al., 2016; Miles and Jamieson, 2012). My paper suggests that within the logistics industry, these subcontracting relationships are coming under new pressures from the construction of e-commerce markets. It is important to understand this dynamic landscape of subcontracting and warehousing, as their limits may provide new leverage points for labor organizing (Doussard, 2013: 42).

Restructuring the last mile

E-commerce’s attempts to re-structure the space between retailers and consumers represent a significant shift from Walmart’s paradigmatic logistics and market strategies (Bonacich and Hardie, 2006; Lichtenstein, 2009: 2; Tsing, 2009). Schematically, Walmart’s story is that it built out its own centralized distribution network rather than using wholesalers to purchase goods from producers (Vance, 1971). Walmart internalized this back end of its supply chain so that it could buy directly from suppliers and leveraged its size to reduce costs and force producers to submit to the requirements of its “lean” distribution network (Abernathy et al., 1999; Gereffi, 1993). These transformations of the retailer-supplier relationship ushered in a new business model known colloquially as big box retail, which has since been adopted by many major retailers (Graff, 2006).

While the Walmart retail model exploded throughout the 1990s and early 2000s, the last decade has seen big box retailer’s growth rates decline and level out at around a 2% annual rate. In contrast, e-commerce retail has been booming and in many cases growing at an annual rate of 30% and up. ² Initially, the lean fantasy of e-commerce retail was to get completely out of logistics and distribution. Retailers could in theory use the internet to sell goods without having to maintain any inventory. They could sell items on a website, then purchase them from suppliers, and finally have them shipped directly from suppliers warehouse’s to a consumer’s doorstep. This vision would have retailers becoming branders, designers, and marketers in the vein of Nike rather than deal in the physical world of commodity distribution (Collins, 2003). However, as the dot-com bubble boomed and busted in the early 2000s, the dream of outsourcing commodity movement died with it (Brandt, 2012). Logistics again became a key strategic resource in a new “retail revolution.” Only this time, retail’s horizon for logistical restructuring became the socio-spatial relations between consumers and retailers, rather than the socio-spatial relations between retailers and suppliers.

Industry boosters frame the restructuring of “last-mile” logistics as a passive response to changing consumer demand (Ganash and Engleson, 2015; Oracle, 2014). In contrast, one can view this as a response to the stagnation of big box retail’s growth. Qualitatively restructuring the way consumers shop offers retail capital new avenues of growth and accumulation by enabling them to sell delivery as much as the item delivered. ³ Because e-commerce warehouses ship products directly to consumers rather than retail stores, this restructuring has required the industry to re-organize its warehousing and its biggest cost: labor. When goods move directly to consumers, cases must be opened and individual items must be picked according to each customer’s order. During an interview, one warehouse manager estimated that this process of picking individual items is 15 times less efficient than picking cases, meaning that 15 times more labor is needed to handle the same volume of product. ⁴ At the same time that warehousing becomes more labor intensive to fulfill the needs of e-commerce, the expectations of “quality” in this work are also increasing. When goods are shipped to consumers rather than retail stores, mistakes become much more costly because they create both angry customers and return shipping costs. Therefore, “quality” in most warehouse work has to do with being able to undertake rote and repetitive labor without making a mistake, such as picking a wrong item or putting it in the wrong location. The demands of e-commerce markets therefore intensify both the labor requirements of this work along with its quality expectations.

These demands have been central to the restructuring last-mile logistics, and as I will demonstrate, they have produced tensions within the sectors reliance on labor market intermediaries to produce flexible labor (De Lara, 2011; Gonos and Martino, 2011; Gutelius, 2015). The restructuring of last-mile logistics has therefore unsettled warehousing’s flexible labor strategies and prompted some managers to directly hire more workers. In order to analyze this labor strategy of reducing warehouse’s dependence on labor market intermediaries, the next section reviews relevant literature on the production of flexible and disposable labor. In particular, I concentrate on feminist work investigating skill in the workplace and on conceptions of worker retention through the lens of logistics research.

Producing flexibility

To undertake my analysis of the reproduction of disposability when warehouse managers increase their number of directly hired workers, it is necessary to review academic discussions of skill, retention, and the workplace. I will begin by reviewing analyses of work and skill in Geography. Research on post-Fordist work has long maintained that the distinction between qualitative and quantitative flexibility is crucial to understanding the production of flexible labor (Standing, 2014: 52–69; Terry, 2011: 661–662). Within this distinction, qualitative or functional flexibility is often associated with a re-skilling of labor that contrasts the Fordist standardization and degradation of work (Braverman, 1998). In general, this re-skilling involves cross-training workers on multiple machines and/or tasks in a workplace, which theoretically makes workers more valuable to their employers and more amenable to specialized production (Gertler, 1988). In contrast, quantitative flexibility is associated with unskilled manual labor. Employers value this kind of labor because it enhances their ability to change how many workers they put to work on a day-to-day basis. In terms of skill, management tends to treat these workers as “warm bodies” that can be called in one day and sent away the next, rather than people with a desirable set of characteristics (Peck and Theodore, 2001: 478).

In contrast to this analysis of skill in the workplace, feminist research on global production has long argued that the notions of skill implicit within these conversations are not static or technical in nature. Ideas about skill are sites of contention and struggle that perform and reproduce particular hierarchies of social difference (Collins, 2003: 171). For example, Marion Werner traces the Dominican apparel industry’s attempts to upgrade its garment-manufacturing sector after losing the protection of US trade quotas. Initially, a resource for national industrial development, the feminized symbol of “unskilled” garment assembly work became a sign of stagnation in elite circles (Werner, 2015: 54). Within this context, Werner demonstrates how attempts at industrial upgrading had to re-signifying the labor process from feminized and unskilled assembly work, which involved repeating the same task all day, to masculinized and skilled sewing labor, which involved working on multiple aspects of the production process (Werner, 2015: 65–73). This case demonstrates that industrial upgrading towards more skilled or “value-added” operations is not a neutral or technical process, but is a struggle over the meaning of work and the inclusion, exclusion, and valuation of socially differentiated workers.

Similarly, Jane Collins details how notions of skill within the global garment industry are ingrained with ideologies of gender and place. Collins does this by demonstrating and analyzing arguments from Liz Claiborne’s management that sewing skills had been lost by US women and retained by women in places like Mexico. Collins argues that this naturalization of skill to place devalues the labor of worker’s in Liz Claiborne’s factories and delegitimizes their claims to skill, even if that skill was not much more than speed. Skill within Jane Collins work is therefore a site of struggle between management and workers, with workers claiming they have skills to legitimize claims for better wages and working conditions, while management attempts to undermine these claims in order to lower labor costs. Through this use of gender ideologies to devalue labor, along with hyper-Taylorist control of the labor process, Liz Claiborne was able to construct a massive outsourced global production network (Collins, 2003: 60). Building on these analyses of the contested nature of skill, this article argues that restructuring warehousing for e-commerce depends on devaluing warehouse work in order to meet increased labor and quality demands in a cost-effective manner.

In addition these analyses of skill, interrogating the idea of retention is important for analyzing how warehouse managers can continue to produce disposable workers even when they express concern with retaining them. Within Fordist work regimes, the central institution for securing the retention of predominately white male workers was the internal labor market (Vosko, 2000: 14–45). Michael Burawoy explains internal labor markets as follows:

In the context of a wider external labor market, the problems for the internal labor market are to ensure that individuals do not seek places outside the firm, on the one hand, and that, in filling places within the firm, priority is given to employees. (Burawoy, 1982: 97)

Burawoy famously updated this analysis of workplace regulation, arguing that a hegemonic and market-based despotism followed the Fordist era of consent through workplace citizenship (Burawoy, 2015). To maintain worker participation in capitalist social relations, practices of coercion replaced those of consent as the threat of job loss replaced the guarantee of workplace rights. A flurry of employment classifications such as agency, part-time, and contract work proliferated within companies themselves, shattering the singular border between internal and external labor markets (Kalleberg, 2000; Mezzadra and Neilson, 2013). Beth Gutelius (2015) makes this point in her research on the hierarchical organization of temp warehouse workers as temp, perma-temp, and super-temp. As Gutelius argues, employer methods of hierarchically segmenting the workplace are aimed at “reproducing relations of power and subordination on an uneven and unstable landscape of restructuring” (Gutelius, 2015: 60). But does this hierarchical production of flexible and disposable labor within workplaces suggest that employers have abandoned a valuation of workers that addresses concerns with retention? Or, have the rationalities of retention themselves changed form, similar to Burawoy’s diagnosis of the changing form of competition between internal and external labor market?

While logistics scholars have not contributed to conversations of workplace regulation through the lens of worker retention, they have been keen to highlight how practices of supply chain management conflate workers and commodities (Cowen, 2014; Kanngieser, 2013; LeCavalier, 2008). The “hyper-coordination” of distribution networks as a whole, constitutes workers and packages as differences of degree rather than differences in kind. These claims by logistics scholars implicitly assert that logistical technologies increasingly make labor more disposable and expendable (Kanngieser, 2013: 606; Neilson, 2012: 329). However, the specific terms of disposability remain largely un-interrogated and it remains unclear the degree to which distribution networks still depend on retaining workers in order to profitably function. I draw on these insights from logistics research to explore how warehouse managers still depend on retaining workers, even while these workers remain individually disposable. I argue that the conceptualization of retention elaborated by warehouse managers has shifted to conform with the logics of supply chain management (Cowen, 2014: chapters 1–2). In place of the Fordist concern with institutionally creating borders between internal and external labor markets, managers interviewed for this research were concerned with controlling the frictions that arise from moving workers in and out of their warehousing jobs. These managers approached controlling the flow of workers through their warehouses at a systems rather than individual level, similar to how they manage the flow of commodities through their doors.

The next sections interrogate two examples of warehouse management reducing their dependence on labor market intermediaries. These examples draw from research on the experimental restructuring of last-mile logistics in central and northern New Jersey, conducted over a 10-month period in 2015 and 2016. The research comprised 69 semi-structured interviews with managers and workers, and participant observation of warehouse and delivery work, which entailed both shadowing warehouse workers and working as a driver and in a warehouse. My sampling of warehouses was oriented towards e-commerce operations and their logistics services, although I also did a small number of interviews in other retail warehouses for the purpose of comparison. The first example draws on research where I was given permission by warehouse management to spend a week in the warehouse and shadow workers who were willing to participate in my research. In this context, despite my self-identification as a geographer, I was often read by management as a supply chain student aspiring to become a warehouse manager myself. As the week wore on, the general manager disclosed that part of his motivation for letting me shadow was to demonstrate to his supervisors that there was a future in supply chain management. On the other hand, workers often read me as a future company owner who was visiting a warehouse to understand how everything works. This fact became painfully clear when a worker asked me to keep them in mind for employment while we walked together to the break room. The second example draws on interviews with warehouse and temp agency managers recruited through snowball sampling. These interviews had a much more formal feel to them. While some saw me as an up-and-coming supply chain graduate student, others perceived me as an unremarkable university student working on an obligatory class project. I now move on to the details of the first warehouse that reduced its dependence on labor market intermediaries.

Discourses of skill

Midway through July 2015, I spent a week of participant observation at the facility of a small e-commerce apparel retailer in central New Jersey. Each day I shadowed a different worker doing a different job in this warehouse. On the third day, I was shadowing and learning about the picking process, the process that has been the most affected by the growth of e-commerce. As explained above, workers in e-commerce warehouses must pick individual items to be shipped to consumers rather than cases of items to be shipped to stores. This shift to picking individual items affects not only the amount of labor that is required, but also the kind of work that dominates a facility and the gendered expectations of who will undertake it. Picking cases often requires the use of forklifts to move pallets or pick cases from pallet racks that can be 40 feet off the ground. In contrast, picking individual items usually involves someone walking around the warehouse floor with a pushcart, getting items stored only up to 6′ off the ground. Jobs that involve forklift driving and machine operation are predominately masculinized in warehouse work. In contrast, women are segmented into lower paying light manual labor, such as the picking processes that are core to e-commerce operations. Explaining this segmentation of warehouse labor, managers often rely on generic tropes that women have smaller hands or do not like heavy machinery (Wright, 2006). One agency manager reiterated patriarchal understandings of the household when he rationalized intentionally sending women to lower paying manual jobs by claiming that women are secondary rather than primary household incomes. ⁵

At the e-commerce warehouse I was at on this day, the general manager estimated that 70% of the workers were women. By contrast, when I visited warehouses that required all or most workers to be able to operate a forklift, women often only represented 10% or less of the workforce. The general manager at this facility explained that previous mangers had used a lot of temp labor in the warehouse. When he came on about five years prior, he had switched from this reliance on temp agencies to directly bringing on all of his workers and training them across all of the warehouse tasks. He explained this switch in hiring from temp to direct by claiming that temp labor is highly problematic for two main reasons. First, he degraded temp workers themselves as “low quality” individuals who go to temp agencies only because of “something in their background or something in their character.” He continued by asserting that with temps there was a lot of drama in the warehouse, with employees stealing, getting in fights, and dealing drugs. The manager’s second argument was that temp agency markups become quite costly for warehouse operations. Agency markups can be up to 50% of wages, and this becomes especially costly when the turnover rate among temp workers is high, resulting in many lost training hours. Sometimes with temp workers, he complained, you might only get a day of work out of someone and then they disappear. From his perspective, temp agencies did not provide cost-effective labor strategies for the companies’ e-commerce operations and he attributed his predecessor’s dependence on agencies to ignorance and “bad management.” Further, temp workers were blamed for performing their disposability on their own terms rather than on their employer’s terms. As our conversation wore on, the manger’s depiction of temp workers was counter posed to his conception of an ideal worker. The manager argued that because warehouse work is relatively mindless, the ideal worker for this job is someone who is able to zone out at work and be content doing the same thing over and over and over, all day long.

This warehouse’s shift from temp to direct hire was not a transition away from flexible labor strategies, but a shift from a total reliance on quantitative flexibility to a mix of quantitative and qualitative approaches. Qualitative flexibility was utilized in the sense that direct hires were cross-trained on multiple tasks (including picking, stowing, and receiving) and shifted around the warehouse depending on where they were needed that day. These direct-hire workers received wages that at the time were mid-scale for the industry’s low-wage standard, starting at $11 an hour. However, shift lengths were very short, usually about five or six hours a day. Shifts were also flexible, with management regularly increasing or decreasing the length of the shift partway through the day. On an average day, workers would show up at 7 a.m. and find out after their first break, around 9:30 a.m., how long their shift would be for that day. By running this warehouse on short shifts through most of the year, management was able to “flex up” its operations for peak season simply by extending the working day rather than bringing in a large number of temp workers. The limited hours that workers received most of the year were by far the largest source of discontent voiced by workers during my participant observation.

Most workers at this facility, who were now directly hired and cross-trained but given hours through highly flexible shifts, had been there for two to three years. This level of retention stands in stark contrast to the general manager’s hypothetical example of temp workers not coming in after the first day. Management did not openly admit that retaining experienced workers improved the quality, efficiency, and productivity of workers, or that this was part of their rationale for shifting to direct hire. However, participant observation with warehouse workers suggests that even in these rote and mundane jobs, the quality and efficiency of the work significantly benefitted from worker experience. For example, training and experience appeared to be more important to the picking process than management admitted. Picking is repetitive and numbing work, which led management to categorize it as “mindless.” However, while it takes a day or two to learn how to pick, the worker I shadowed while he picked explained that it can take months to really learn the most effective ways of navigating the racking locations where picking occurs.

To learn about the picking process in this warehouse, I was paired with a Puerto Rican man in his 50s who had come to this job about five years earlier, after being laid off from a supervisor position at a nearby window manufacturer. After greeting each other, my partner took me to the central dispatch station and began explaining how the picking process works. We walked up to two tables where one person was stationed all day; printing out, organizing, and preparing pick lists. Pick lists are narrow sheets of paper with stickers on them, one sticker for each item that needs to be picked. These stickers were ordered from top to bottom according to what a computer algorithmically determined to be the most efficient route to walk when picking this group of items. When we got to the dispatch station, my partner looked through a stack of available pick lists and took those for zones 10 and 12. He would repeat to me throughout the day that zones 10 and 12 were his favorite to pick in. He knew these zones well and picked faster in them than in other zones in the warehouse. At first this statement seemed odd since the picking process was theoretically the same throughout the entire warehouse, but throughout the day the reasons behind this preference became clearer.

After getting the lists for his favorite zones, my partner took a sticker off the bottom and placed it on a sheet at the central dispatch station. This sheet was used to track which worker had which list, so management could record their productivity and accuracy. He then collected two bins and a pick cart, and we headed off to zones 10 and 12 to begin picking. Arriving at zone 12, I was shown a single pick sticker that represented an item that needed to be retrieved. I was told that at the top-left corner of each sticker is a number that provides the location where an item is supposed to be stored. From left to right, this number gives the aisle number, the row number within that aisle, the shelf number, and the bin number where the item is located. It is with these numbers that pickers navigate the physical storage areas they are working in. Every storage position in the warehouse therefore has a unique number made up of these four variables. Individual items are scanned into these storage locations and it is based on this location information that warehouse management systems arrange the order of items on pick lists into the most efficient walking path for pickers.

Despite that pick lists are automatically arranged for efficiency, my partner explained that directly following those pick paths is a slow strategy. People only follow the assigned paths when they first start on the job. Going straight down this list is slow because it creates a lot of backtracking and unnecessary walking. This backtracking is created because the parameters of the warehouse management system require that every item on one side of an aisle be picked before moving on to other side of the aisle. In practice, this means that when a picker is picking one item, an item four stickers down the list might be immediately behind them on the other side of the aisle. However, the pick list will instruct them to walk down one side of the entire aisle first and then backtrack to their original location, to pick the item on the other side of the aisle. To avoid such backtracking, my shadowing partner showed me how he re-routed the algorithmically generated pick-paths he is given. Instead of going straight down the pick list, my partner first scanned the entire list to determine where in the picking zone he wanted to start and finish, getting a general idea of the locations he had to cover for this pick list. After determining where he would start, he focused on about five or six items at a time and calculated for those items what he thought would be the quickest picking order and route to take. Far from putting on headphones and zoning out as management had described their ideal workers, my partner was continually routing and re-routing the paths he took through the storage area in order to avoid as much backtracking and walking as possible. The ability to quickly re-route one’s pick path depends in many ways on a worker’s accumulated experience and knowledge of particular pick zones. This is one reason why my partner favored picking in zones 10 and 12. Whenever we were in areas other than 10 and 12, he would make sure I knew that he did not like picking there. Because of the level of focus required, my partner also explained to me that unlike other people in the warehouse, he could not listen to music while picking because it was too distracting for him as a musician. Our conversations suggested that he was far from unique or the only person that used these tactics. Throughout the day, my partner would often point out other pickers and say that they looked like they were moving very slowly but were actually very fast pickers. In contrast, he would point out other people moving very quickly and say that they actually pick quite slow. Given that each worker’s pick rates were publicly posted in the warehouse, these assessments were likely based on tracked data rather than personal opinion. These examples suggest that picking can be not only very physically demanding, but cognitively taxing as well. It is because of the cognitive aspects of this work that an individual worker’s efficiency and potential productivity appear to significantly depend not only on their experience at that facility but also on their experience in and knowledge of certain segments of the warehouse.

In sum, a new and self-proclaimed more experienced manager brought innovation to this e-commerce retailer’s warehouse by shifting from a dependence on temp agencies to cross-trained direct-hire employees that were made quantitatively flexible through daily changes in shift lengths. This strategy has been beneficial enough for management to maintain it for five years since its implementation. My research suggests that the benefits management realized from this switch can be attributed in part to two dynamics: increased efficiency in the labor process, gained from retaining experienced and cross-trained employees, combined with a continuing degradation of this more “skilled” labor as mindless. Through this degradation, the material benefits of the transition to direct hire have been primarily captured by ownership rather than distributed to workers in the form of either increased wages or stable hours. In the absence of strict rate-based incentives or punishments to squeeze these efficient employees, workers at the warehouse were able to capture some benefits by achieving their minimum requirements with minimal physical effort. But rather than a victory, this achievement seems more like a survival strategy for workers forced to work multiple jobs because of their limited hours. ⁶ Management’s adjustment to hire all workers directly was therefore oriented towards producing higher skilled labor at lower costs, through the continued degradation of this work and last minute changes to the working day. This insight is important in context of e-commerce market construction because the commoditization of delivery will depend on finding these kinds of labor strategies in order to meet warehousing’s increased quality and productivity demands in a cost-effective way.

Redefining retention

Another example of moving away from labor market intermediaries occurred in the warehouse of a big box store retailer who predominately sold furniture and home goods. This warehouse was located in central New Jersey and was about 800,000 sq. ft. Across the street, a 3PL operated a warehouse about the same size that exclusively handled their product. Inside the lead firm’s warehouse, there was a national and a local temp agency along with a 3PL that only unloaded inbound trucks. This kind of arrangement is referred to as “in-sourcing,” which means that a contractor runs part of a warehouse that otherwise is operated directly by the lead firm. While this company handled both online and in-store orders, the warehouse that I was at focused mostly on replenishing retail stores and even some light assembly of furniture. The company’s online orders predominately went through the warehouse across the street operated by a 3PL. Management estimated that women made up roughly 13% of the workforce in this warehouse, which stands in stark contrast to the 70% estimated by management in the first example. Despite the fact, that e-commerce was a relatively small part of this warehouse’s operations, my interviews made clear that they were feeling the effects of e-commerce’s increased demand for labor in the form of tightening labor markets and a heightened interest in worker retention.

In comparison to other warehouses where I conducted research, this warehouse had a high degree of outsourcing in its operations, both in terms of the number of intermediaries they utilized and in the kinds work they were temping out. A manager stated that they aimed to temp out 50% of their workers but that in reality it fluctuated between 40% and 50%. Additionally, they were temping out forklift operating jobs, which are more regulated and require more training than picking. On the one hand, it is plausible to think that since this warehouse functioned with a high degree of subcontracting, it was operating smoothly under these conditions. However, this warehouse was in the process of scaling back the degree to which it depended on agencies and third parties by increasing their percentage of direct hire workers and integrating permanent and temp workers on the shop floor. The issue of retention was a recurring theme in my conversation with management here, which they took to mean gaining greater control over the movement of workers in and out of their jobs rather than a concern with keeping individual employees. This manager elaborated that the reasoning behind decreasing their reliance on intermediaries was to deal with problems created by their organization of subcontractors and temp agencies in the warehouse. The first problem he elaborated had to do with the fixed term structure of temp worker’s contracts.

Author: So what was the reasoning behind the decision to go more direct hire?

The way this manager conceives of retention is reminiscent of logistical rationalities based on the hyper-coordination of all movement in the supply chain. As Kanngeiser states:

The management of bodies and commodities now encompasses the entire spectrum of movement, from the minute gestures of box packers and the pathways of cranes in the warehouse, to the rest breaks of freight drivers, the call content and duration of call centre workers, and the passage of commodities shipped around the globe. (Kanngieser, 2013: 596)

Such an understanding of retention as management control over worker movement in and out of warehousing jobs was evident in a second issue dealt with by this same manager. The second issue in this warehouse was the unauthorized movement of workers between subcontractors and temp agencies. In this facility, two temp agencies were tasked with finding already trained forklift drivers to put pallets away for storage and take them out for picking. These forklift positions were the majority of the jobs filled by the national agency that I interviewed. The lead firm was not concerned with how agency workers were trained on forklifts as long as they passed a 10-minute test onsite validating they knew how to drive one. Through this organization of recruitment, the lead firm externalized the cost of training forklift drivers and depended on there being enough trained drivers available in the labor market.

In contrast to the two temp agencies that had to find already trained forklift operators, the in-sourced 3PL that unloaded goods on a piece rate basis was contracted to train its workers for up to two weeks on forklift operation. The mismatch in the warehouse between contracting a 3PL to train forklift drivers and temp agencies to find already trained drivers became a cost problem for management. Once trained, workers would leave the contractor for better-paying jobs that the temp agencies offered in the same warehouse.

Warehouse Manager: So generally what we see is that (laughs) [the in-sourced 3PL] hires people with no experience on a forklift, they will train them for two weeks, they’ll work for two weeks and then quit and go work for [a temp agency in the same warehouse] because they pay them more money…. So that’s why we’re transitioning away from giving [the in-sourced 3PL] more stuff, right? So we just went through that pitfall right there. I went through a cycle of about six months where we pay the training but we get zero production out of it. So we’re actually eating the cost. So what we decided, I decided, about two months ago is I’m not doing it anymore. For six months I had eighty hours I had to pay them every week to train people that would eventually quit and go work for someone else. So then I said, “Forget that. Hire me ten humans please.”

This example offers a window on to the broader pressures facing the warehousing industry, which are creating conditions that may make directly hiring workers an attractive strategy. This is because e-commerce’s large demand for labor is tightening the entire labor market for warehouse work. My interviewee explained that in recent years Amazon had opened up warehouses 30 miles to the north and south of this facility and that it immediately became harder and took longer to find workers. The national agency in this case also had a contract with Amazon, and sales reps from this agency claimed in an interview that Amazon was asking for up to 1200 workers a shift. This provides some reason to believe that the conditions pushing management to directly hire more workers may be replicating themselves at larger scales throughout the industry. In conclusion, I will consider the implications of the warehouse managers’ strategy in this article’s examples for our understanding of subcontracting’s role in the production of flexible labor.

Conclusion

As big box retail’s growth has stalled, the construction of e-commerce markets has emerged as a new outlet of value realization for retail capital. These markets are new in that they sell the commodity of delivery along with manufactured goods. To make these e-commerce markets possible, warehouses have had to restructure their operations and organization of work. This restructuring has created new demands on the amount and quality of work in warehousing, creating frictions with this sectors heavy reliance on labor market intermediaries like temp agencies and 3PLs (Gonos and Martino, 2011; Gutelius, 2015). Interrogating this restructuring of the increasingly crucial logistics sector in the US economy, this article has attempted to understand how warehouse labor is being reproduced to make these e-commerce markets possible. My research demonstrates that one response to these demands of e-commerce warehousing has been to reduce its reliance on intermediaries and directly hire more workers. Despite the fact that this strategy moves closer to the “standard” employment relationship, my research shows that workers disposability is being reproduced even as they are directly hired by lead firms. This article has illustrated at least two mechanisms for reproducing warehouse workers’ disposability through direct employment. First, warehouse management can mobilize discourses of “unskilled” labor to degrade the kinds of work that are increasingly core to their e-commerce operations. In the first warehouse, this was happening as management relied on workers unspoken knowledge of the warehouse’s organization but reserved the right to change their shift length halfway through the shift. Second, warehouse management can mobilize conceptions of retention that seek to maintain enough trained workers in their warehouses rather than invest in individuals. Disposability in this instance is reproduced in that capitalist firms do not attach particular value to any individual worker, only to the turnover rates of the entire workforce.

Significantly, these cases develop our understanding of the production of flexible labor within both the logistics industry and more generally in the post-Fordist economy. In the warehousing and logistics industry, it is well documented that temp agencies and 3PLs play a significant role in offloading the cost and risk of lean distribution on to workers (Cho et al., 2012; Gonos and Martino, 2011; Gutelius, 2015; Tsing, 2009; Warehouse Workers for Justice, 2010). A commodity chain analysis of these relationships has become a significant frame of reference in both academic and labor organizing circles, providing a means to map, identify, and target key relationships in the supply chain (De Lara, 2012; Miles and Jamieson, 2012; Quan, 2008; Theodore, 2016). This article suggests that such approaches must proceed with caution in prioritizing labor market intermediaries’ role in degrading workplace conditions. My research suggests that this caution is warranted because logistical restructuring is putting new strains on existing contracting relationships, and employers have plenty of mechanisms beyond labor intermediaries to degrade working conditions. While this article interrogates emerging tensions within warehousing’s labor strategies, what remains to be understood are the limits of these new strategies, the limits of e-commerce’s growth, and how these limits might be exploitable for workplace organizing. Future research may therefore find it useful to look at these tensions and limits of e-commerce in order to understand potential leverage points for labor organizing in this changing industry (Doussard, 2013: 42).

In addition to these insights for research on logistics, this article’s conclusions are significant for labor geography more broadly. In analyzing the production of flexible labor, geographers have understandably prioritized the role of labor market intermediaries and subcontracting relationships (Coe, 2013; Coe et al., 2009; Peck and Theodore, 2001). However, my research draws into question analyses of subcontracting as a paradigmatic form of employment (Wills et al., 2009). While it is clear that the use of subcontractors and temp agencies was critical in undermining the gains made by worker’s under Fordism, it is not clear that they remain as tactically and functionally significant for employers now that the United States’ labor movement has been significantly weakened (Hudson, 2001: chapter 4; Peck, 2016). My research suggests that while these intermediaries provide a convenient method of producing disposable labor, they are not necessary. Further, in cases like e-commerce where the task of industrial restructuring is about the production of new markets and commodities rather than cost-reduction (Silver, 2003: chapter 3), I would argue that the effectiveness of these tools for management is by no means given.