Fulfillment of the Work Games: Warehouse Workers' Experiences with Algorithmic Management

Abstract: The introduction of algorithms into a large number of industries has already restructured the landscape of work and threatens to continue. While a growing body of CSCW research centered on the future of work has begun to document these shifts, relatively little is known about workers' experiences beyond those of platform-mediated gig workers. In this paper, we turn to a traditional work sector, Amazon fulfillment centers (FC), to deepen our field's empirical examination of algorithmic management. Drawing on two years of ethnographic research, we show how FC workers react to managers' interventions, imposed productivity rates, and quantified objectification when subjected to labor-tracking systems in their physical work environments. Situating FC workers' resistance to algorithmic systems and metrics within the current CSCW literature allows us to explicate and link the nuanced practices of FC workers to the larger discourse of algorithmic control mechanisms. In addition, we show how FC workers' resistance practices are emblematic of 'work games'--a long-studied means by which workers agentically configure ("trick") their engagement within work systems. We argue that gaining a more nuanced understanding of workers' resistance and consent in relation to algorithmic management expands our ability to critique and potentially disassemble the economic and political forces at the root of these sociotechnical labor systems.

• The paper uncovers how workers employ innovative 'tricks' to both resist and comply with algorithmic oversight in Amazon warehouses.
• It uses a two-year multi-sited ethnographic study with constructivist grounded theory to analyze digital surveillance, performance tracking inaccuracies, and managerial intervention.
• The study critiques simplistic views of algorithmic control by highlighting the nuanced interplay between human agency, system errors, and managerial mediation.

Ethnographic Inquiry into Algorithmic Management: Warehouse Workers’ Resistance and Consent

Introduction and Motivation

The paper "Fulfillment of the Work Games: Warehouse Workers' Experiences with Algorithmic Management" (2508.09438) offers an extensive ethnographic investigation into the lived experiences of Amazon Fulfillment Center (FC) warehouse workers subject to algorithmic management. Contrasting prior CSCW and HCI work that privileges platform-mediated gig work, this study turns to a datafied but conventional labor site, capturing the nuanced interplay between physical labor, digital surveillance, algorithmic control, and managerial intervention.

At the core, the research critiques the binary framing of algorithmic management as either purely oppressive or empowering, instead presenting workers as agentic actors capable of negotiating, resisting, and subverting algorithmic impositions. The focus on worker-devised “tricks”—strategies for gaming productivity metrics and labor-tracking systems—advances a more granular understanding of consent, resistance, and the lived realities inside algorithmically managed workplaces.

Methodological Framework

Fieldwork spans two years, leveraging multi-sited ethnography in the absence of direct physical site access. The study triangulates data from three online worker-run communities, 16 in-depth interviews (spanning several roles and sites), community artifacts, and union archival materials. Data analysis follows constructivist grounded theory, centering workers' own narratives and vernacular (e.g., “tricks” in preference to research-initiated labels like “coping strategies”).

This methodological stance foregrounds both the empirical constraints inherent to studying surveilled environments and the ethical challenges in reporting worker “tricks” without further endangering participants. Notably, workers remain highly sensitive to managerial and algorithmic scrutiny, complicating both access and disclosure.

Findings: Algorithmic Control, Managerial Mediation, and Worker Numerification

Assembly and Function of Labor-Tracking Systems

Amazon FCs deploy dense assemblages of digital and analog tracking technologies—badges with barcodes, handheld scanners, item and location barcodes, workstation software, pervasive cameras, and the A-to-Z app. Key metrics such as rate, UPH, stow/pick rate, and TOT (time-off-task) dominate workers’ performance evaluations, with the expectation to “make rate” being normalized and rigorously enforced via the labor-tracking system.

Crucially, not all labor is directly tracked—distinctions between “direct” and “indirect” roles create differential susceptibilities to algorithmic oversight. For direct labor, surveillance is granular and constant; for indirect labor, manual reporting and managerial discretion mediate computational representation.

Data Discrepancies, Human Mediation, and the Role of Managers

Algorithmic representations of labor are rife with inaccuracies—scanner logouts, network disruptions, and inflexible metrics regularly yield misaligned data, frequently penalizing workers without recourse. Human managers thus act as “context probers” and partial algorithmic intermediaries: workers must pro-actively advocate for corrections, relying on managerial discretion to resolve data discrepancies.

The study finds that, contrary to visions of fully automated managerial displacement, FC managers retain significant power as interpreters, enforcers, and sometimes manipulators of the system. Managerial relationships become instrumental for workers’ survival and accuracy within the algorithmic regime, introducing human unpredictability atop computational rigidity.

Numerification and Dehumanization

Algorithmic management engenders affective alienation among workers, with the prevailing experience described as being “treated as a number.” This numerification is both structural and psychological—break times, error tolerance, and reward systems are subordinate to productivity metrics, with even peer-to-peer interactions increasingly subsumed by abstractions produced through data. Amazon’s self-aware introduction of engagement surveys that directly ask about “being treated like a number” further underscores the ubiquity of this affective state.

Worker Resistance: “Tricks” and the Emergence of Work Games

Taxonomy of “Tricks”

The most original empirical contribution is the detailed cataloguing of warehouse worker “tricks”—from “cherry-picking” fast tasks to manipulating “scan to scan” metrics, exploiting menu screens for longer breaks, delegating scan events to colleagues, or subverting clock-in using misleading geolocation. These are understood both as micro-resistances and re-assertions of creative agency in a context otherwise dominated by machine logic.

Importantly, most “tricks” are individualized, cultivated within the constraints of solitary and surveilled work, but disseminated through online communities. Collective resistance—such as coordinated strikes—remains largely absent given the strict anti-union protocols and high surveillance.

Theoretical Perspective: Work Games and Consenting Resistance

By reactivating Burawoy’s “work games” analytic [burawoy1979manufacturing], the authors frame these “tricks” as simultaneously acts of resistance and compliance. Playing the game, by the rules or through creative subversion, necessarily entails some degree of consent to the structure itself. The authors advance a typology:

1. Active consenting resistance: Exploiting system flaws (e.g., timing, order of operations) but still engaging with the productivity regime.
2. Passive consenting resistance: Breaking explicit or implicit rules when undetected, but optimizing for performance.
3. Non-consenting resistance: Disengaging, evading, or seeking to remove oneself from algorithmic management (e.g., faking clock-ins).

Thus, the dialectic of consent and resistance emerges, with workers’ agency never absolute but navigated within structural constraints.

Implications for Theoretical and Practical Developments

Critique of Algorithmic Management

The research demonstrates that algorithmic management within traditional workplaces like warehouses is neither a replication of platform gig work control nor a simple extension of Taylorism. Instead, it is a hybrid configuration characterized by the coexistence of omnipresent digital surveillance, persistent managerial power, and the conditional agency of workers. The warehouse site is a paradigmatic locus for observing future trajectories of algorithmic control in broader labor contexts.

Prospective Directions in Algorithmic Labor Studies

This work points to several areas for further inquiry:

• Transparency and sensemaking: Ongoing opacity in algorithmic systems prompts the development of online communal sensemaking, even among non-remote, physically co-present workforces.
• Limits of datafication: Systemic errors and the necessity of managerial mediation highlight that totalizing algorithmic control is technically and organizationally incomplete.
• Labor agency: Worker tactics, while not collective in the traditional unionist sense, illuminate persistent creativity and resistance at the interface of human and machine decision flows. The granular description of game-playing provides a new frame for future studies of algorithmic power and subjectivity.

Pragmatically, interrogating the boundary between consent and resistance is essential for the design of future labor regulation, organizational policy, and sociotechnical interventions.

Conclusion

Through in-depth ethnographic engagement and fine-grained analysis, this work demonstrates that algorithmic management in Amazon FCs constitutes a contested and negotiated terrain shaped not just by technical systems but by managerial actors and, critically, by workers themselves. The nuanced depiction of “tricks” and work games revises the discourse on algorithmic control, foregrounding the dialectic of compliance and resistance enacted through everyday practices. This analysis broadens the empirical and theoretical horizon for understanding algorithmic management, worker subjectivity, and the politics of the datafied workplace.