In the past the man has been first; in the future the system must be first.

Frederick Winslow Taylor

Introduction

In 1911, Frederick Taylor released The Principles of Scientific Management, which presented ‘task management’. This approach replaced workers’ rules of thumb with procedures derived from scientific study, standardizing production to achieve its best form [1].

A century later, the same logic was encoded digitally. Lee, Kusbit, Metsky, and Dabbish first discussed in their research on Uber and Lyft [2]. Now, platforms using big data and artificial intelligence manage millions of workers who are independent contractors, often without any direct human supervision [3].

The changes linked to the Fourth industrial revolution have changed not only technology, but also social structures, impacting how professional work and labor are managed [4].

This leads to a key question of this paper: How have scientific management ideas changed into algorithmic control systems in the twenty-first century, and what managerial issues does this change create? This question matters both in theory and practice. In theory, it shows how managing labor scientifically has turned into controlling information [5]. In practice, algorithmic management has moved beyond just platform labor and is now present in manufacturing, logistics, retail, healthcare, and professional services [6].

Three things have come together to make digital Taylorism a reality: economic globalization pushing the need for new ways to be efficient [7], the technological basis of the Fourth Industrial Revolution [8, 9], and the COVID-19 pandemic, which sped up remote monitoring and digital control in all areas [10].

This paper aims to: show the ongoing link between scientific management and algorithmic control; apply this idea to examples, such as Amazon, Yandex Taxi, and Sber; and critically look at the managerial issues created by digital Taylorism. The study uses both theory and real-world examples, using ideas from labor process theory, agency theory, algorithmic management theory, and dynamic capabilities.

The paper is organized as follows. The first part develops the theoretical approach. The second part applies this approach to warehouse logistics and platform labor. The third part discusses the main tensions and issues created by algorithmic control systems. The paper ends with what these findings mean for organizations in the twenty-first century and suggests areas for future study.

Theoretical Framework

Historical preconditions of the digital Taylorism

To understand algorithmic control in management today, one must start with how management ideas have grown over time and what led to digital Taylorism.

Modern management theory comes from F. Taylor’s Principles of scientific management, which replaces workers’ informal methods with procedures derived from scientific study of work. The key idea is to break down production into small, simple steps and standardize them through ‘The one best way to do the job’ [1].

Harry Braverman, in his 1974 book Labor and Monopoly Capital, predicted this move toward digitizing labor control: ‘…the key element in the evolution of machinery is not its size, complexity, or speed of operation, but the manner in which its operations are controlled’ [11].

Max Weber thought that modern management involves strict rules and being impersonal, where rules matter more than personal feelings [12]. He called this ‘rationalization,’ where clear calculations drive all human activity. Researchers like Baiocco see current algorithms as the result of Weber's ideas: ‘Potentially, computer-based algorithmic management can be the culmination of Weber´s bureaucratisation process’ [6]. Instead of managers making choices based on experience, code now executes rules automatically. The ‘iron cage’ [12] of control is now a reality, not through bureaucracy, but through digital code.

For such widespread digitization to succeed, it needs strong support in both ideas and resources. Agency theory offers a framework for why organizations invest in monitoring and control [13]. The theory points out a problem in the relationship between a principal and an agent: agents (employees or managers) have information and interests that may differ from those of principals (employers or shareholders), which can lead to them avoiding work, acting selfishly, or taking advantage of situations. Employers need to set up monitoring and reward systems to make sure agents act in line with the principals' goals. Algorithmic management arises as a way to solve this principal-agent problem by putting in place automated monitoring and management tools [14].

The idea of algorithmic management entered academic discussion with Lee, Kusbit, Metsky, and Dabbish's study of Uber and Lyft drivers, which examined the managerial role of platform algorithms [2]. Baiocco et al. define algorithmic management as ‘the use of computer-programmed procedures for the coordination of labor input in an organization’ [6]. This definition includes the algorithmic system and its managerial goal.

Kellogg, Valentine, and Christin offer a good overview of algorithmic management in their review in Academy of Management Annals [15]. They list six ways algorithms control: restricting (limiting what workers can decide), recommending (guiding worker choices), recording (collecting data on worker activity), rating (evaluating performance), replacing (substituting workers who don’t perform well), and rewarding (motivating desired behavior). These connect to Fayol's management functions (planning, staffing, commanding, coordinating, and controlling) [16], showing a link between traditional and algorithmic management. The historical evolution of management control is presented in the Figure 1.

Fig. 1. The historical evolution of management control theories [28]

Strategic Tension of Digital Taylorism

A key issue arises when algorithmic management is considered with Dynamic Capabilities Theory. Dynamic capabilities mean an organization's ability to spot new chances, take them through strategic moves, and change resources in response to changes. These depend on human input: judgment, creativity, and problem-solving: ‘Entrepreneurial management has little to do with analyzing and optimizing. It is more about sensing and seizing-figuring out the next big opportunity and how to address it…’ [17].

Algorithmic control reduces these abilities. When workers must follow algorithmic methods without input, they cannot find and make improvements. When performance is judged only on numbers, actions that help learning and adapting but aren’t measured, like mentoring or experimenting, are not valued. This paper sees the tension between algorithmic control and dynamic capabilities as the main issue of digital Taylorism: systems that improve efficiency and predictability in the short term may weaken the abilities companies need to adapt and innovate in the long term.

Main Analysis: Digital Taylorism in Practice

Amazon Fulfillment Centers — The Algorithmic Factory

Amazon's fulfillment centers are a well-known example of algorithmic management in a workplace. The system closely follows Taylorism using digital tech, and data from journalism, labor studies, and investigations offer rich material for analysis.

At the center is a system of ‘norms’: a real-time performance metric that tracks each worker's output against an algorithm-set target. Workers who miss the norm get warnings; repeated failure leads to automatic dismissal without management review. This mirrors Taylor's idea: separating thinking from doing. The algorithm sets the norm, while workers perform tasks without choosing pace or method. The algorithm, not a manager, makes decisions [18].

The control system supports this. Employee movements are tracked using scanners that record every item touched, downtime, and any changes from set routes. The algorithm improves these routes, guiding workers through the warehouse along the best path — Taylor’s ‘one best way’.

The company's view is clear. Amazon encounters a principal-agent problem: how to ensure warehouse workers try hard at minimal cost. The algorithmic wage system solves this by reducing information differences — every movement is recorded. From an agency theory view, the system is effective in its goal.

The result matters. A report [19] showed that injury rates at Amazon warehouses in the US are about twice the industry average, suggesting that productivity is favored over worker health. This is because the system doesn't include injury costs in its optimization. The system works best with known methods in a stable setting and cannot adapt.

Platform Labor — Yandex Taxi as an Algorithmic Employer

Digital labor platforms represent advanced algorithmic management. Algorithms don't just assist managers they sometimes replace them. Algorithmic control removes contact, automating the traditional manager role: workers are managed, evaluated, and disciplined through tech, not a supervisor. This is management without managers — Taylorism without a foreman [15].

Yandex Taxi, a taxi platform in Russia, shows digital Taylorism, confirming models from Western platforms [2, 20]. Research shows algorithms match participants and motivate independent drivers. The platform controls transaction details, standards, and service conditions [21].

A key feature is the legal issue it creates. The platform classifies drivers as independent, but controls working conditions: pricing, standards, routes, bonuses, and sanctions [21].

From an agency theory view [14], Yandex Taxi's system solves the principal-agent problem: GPS tracks, ratings monitor, and data collection eliminates information differences. However, this causes issues. A study showed that the bonus system made drivers unable to reach targets, as the algorithm reduced ride requests near the threshold [22]. This results in long workdays, as workers extend shifts for bonuses [6].

Yandex Taxi mirrors the globalization aspect of digital Taylorism: the same logic from Uber in the US was used in Russia, Eastern Europe, Central Asia, and the Middle East, showing the spread of algorithmic management [23].

Algorithmic HR Management — Sberbank

The Amazon and Yandex Taxi cases show how digital Taylorism works in physical and platform labor. Back in 2025, Sberbank restructured its workforce which spread this idea to a different area: knowledge work for white-collar employees in a big financial business. This case matters for theory because Taylor’s original idea was for physical tasks that could be easily seen. Using computer control for thinking and managing jobs is the newest part of digital Taylorism. As of this writing, with the restructuring occurring in late 2025, there are no peer-reviewed academic papers about it. The below analysis relies on journalistic and business sources, which are cited.

In November 2025, at the AI conference, Sberbank's CEO, Herman Gref, said that the bank would cut 20 % of its main staff based on a performance review done by AI. The tool they used was a system that looked at how project teams across the business were doing and what they produced. It found projects that weren't doing well and pointed out the workers who should be fired or moved to other jobs. The size of the changes could be seen in the financial data: by September 30, 2025, the Sberbank Group had 294,578 workers, which was less than the 308,092 at the end of 2024. This means they cut about 13,500 workers in nine months [26]. Gref later said that the AI system’s reviews were over 80 % correct, which is why they started the process of making things more efficient [27].

From a business point of view, the idea is simple: a large business with many knowledge workers faces a problem where managers don’t always know what workers are doing. The AI system fixes this by giving more information about what thousands of workers are doing at the same time. The Kellogg idea fits the Sberbank case: the system keeps track of performance data, compares workers to standards, and makes decisions about who to replace, doing three of the ‘6 R’s’ of computer control in one automated process.

But the Sberbank case shows a problem that the Amazon case doesn’t. In a warehouse, it’s easier to measure how much work is being done because you can see how many items are handled per hour. Knowledge work, like creating financial products, handling client connections, and organizing projects, involves results that are harder to measure, and the value often appears over a long time. When a system says a project team is ‘not efficient’ and should be closed, it has to guess based on things like milestones reached, money spent, and short-term income. The knowledge that the team has, the relationships they've built, and the early work that hasn't yet shown results are not seen by the system. This is the Dynamic Capabilities problem in its clearest form: the things that Teece says give a business a lasting advantage are exactly the ones that computer review systems can’t measure [17].

German Gref didn’t say what the system used to review worker performance. This is a ‘black box’ problem that makes computer management decisions hard to understand for those who are affected by them [26]. By February 2026, Gref admitted that the changes had caused problems for the business: We cut 20 % of our staff last year, 20 % of our engineers. This was hard to imagine even three years ago. And we fired real people. It was a shock inside the company [28]. This result shows the hidden costs that the focus on doing things efficiently ignores: the loss of business knowledge, the destruction of worker skills, the loss of trust, and the problems that come with cutting many workers which don’t show up in the performance numbers that the system is made to increase.

Discussion: Tensions, Trade-offs, and Critical Reflections

Efficiency vs. Human Capital

The core tension of digital Taylorism is between short-term efficiency and long-term human capital loss. Oztemel and Gursev show that automation improves factory productivity, aligning with Taylor's idea of systematic optimization [24]. However, short-term gains may hide long-term costs.

Amazon’s injury and turnover data illustrate this temporal mismatch concretely. A system that maximizes short-term throughput by pushing workers to the edge of their physical capacity generates high rates of injury, absenteeism, and voluntary turnover — each imposing substantial costs. The Strategic Organizing Center data showing Amazon’s injury rates at approximately twice the industry average suggests that the algorithmic rate system, by optimizing for output without incorporating worker health as a constraint, generates significant hidden costs that do not appear in the productivity metrics the system is designed to maximize [19]. This is not merely a human cost; it is a managerial failure: the system measures the wrong things and therefore optimizes toward an outcome that, from a total-cost perspective, may not be optimal at all.

The Sberbank case extends this argument to knowledge work. When a multi-agent AI system evaluates project teams on quantifiable proxies — milestones met, budget consumed, short-term revenue generated — it systematically discounts the tacit knowledge, institutional relationships, and exploratory work whose value materializes over longer time horizons. Gref’s own acknowledgment that the restructuring caused ‘shock’ inside the organization [28] illustrates the category of cost that algorithmic efficiency metrics cannot capture: the erosion of trust, the loss of organizational memory, and the disruption of informal coordination networks that sustain day-to-day performance. Human capital, in this sense, is not merely a stock of skills; it is an embedded, relational asset that algorithmic workforce reduction can destroy faster than it can be rebuilt.

The Dynamic Capabilities Paradox

A key tension is between algorithmic control and innovation. Teece et al.’s [17] framework argues that sustained competitive advantage in complex environments derives from the ability to continuously sense new opportunities, seize them through strategic investment, and reconfigure organizational resources in response to change. These capabilities require organizational flexibility, worker autonomy, experimentation, and the exercise of tacit judgment — precisely the capacities that algorithmic management is designed to constrain.

Algorithmic management is, by design, an exploitation mechanism: it optimizes performance in known task environments according to defined metrics. The literature on organizational learning distinguishes between exploitation — the refinement and extension of existing knowledge — and exploration — the search for new knowledge and capabilities. Exploitation pursued without exploration leads to organizational rigidity and eventual obsolescence as environments change. When workers must follow algorithmically prescribed methods without the discretion to deviate, when performance is evaluated solely on quantitative indicators, behaviors that promote learning and adaptation — mentoring colleagues, experimenting with new approaches, building cross-functional relationships — are systematically undervalued and crowded out. Organizations that fully embrace digital Taylorism therefore risk optimizing themselves into irrelevance.

This paradox is particularly acute for organizations that simultaneously pursue efficiency through algorithmic management and innovation through knowledge-intensive strategy. The cases examined in this paper illustrate the tension across different sectors: Amazon optimizes warehouse throughput at the cost of worker health and retention; Yandex Taxi maximizes ride completion rates while structurally preventing drivers from accumulating earnings; Sberbank reduces headcount through algorithmic evaluation while acknowledging the organizational disruption this creates. In each case, the system achieves its proximate goal while generating costs that fall outside its optimization function. Digital Taylorism may, in this sense, be a strategic trap: effective at its stated objectives while systematically undermining the broader organizational capabilities that twenty-first-century competitive environments demand.

Challenging Classical Assumptions

Digital Taylorism questions foundational assumptions of classical management theory in ways that its originators could not have anticipated. Most fundamentally, it challenges the assumption that organizations require human managerial judgment to function effectively. Fayol’s five functions of management — planning, organizing, commanding, coordinating, and controlling — presuppose a human agent who interprets organizational context, exercises discretion, and adapts to circumstances [16]. The substitution of algorithmic systems for human managers in the direction, evaluation, and discipline of workers represents a structural transformation of the organization that has profound implications for what management, as a profession and a practice, actually means.

A second challenged assumption concerns transparency and accountability. Classical management theory, from Taylor through Weber, assumes that the rules governing worker behavior are knowable, stable, and legible to those subject to them. Algorithmic management systems frequently violate this assumption: as noted in the Sberbank case, Gref did not disclose the specific criteria by which the AI evaluated employee performance [26]. Kellogg et al. identify opacity as a structural feature of algorithmic control rather than an incidental flaw — algorithmic systems often function as ‘black boxes’ whose decision logic is inaccessible to the workers they govern, to the managers nominally responsible for them, and sometimes even to the organizations that deploy them [15]. This opacity creates accountability gaps that existing regulatory and organizational frameworks are not equipped to address: when an algorithm dismisses a worker, who is responsible?

What remains of management when the core managerial functions are automated? This question, which the emerging literature on algorithmic management is only beginning to address, may prove to be one of the defining organizational questions of the twenty-first century. The cases examined in this paper suggest a partial answer: what remains is the design of the system itself — the choice of what to measure, what to optimize, and whose interests to prioritize. These are not technical decisions; they are deeply political and ethical ones. Digital Taylorism does not eliminate managerial judgment; it displaces it upstream, into the design phase, where it becomes less visible and therefore less accountable.

Conclusion

This paper has argued that digital Taylorism — algorithmic management of work through automated systems — intensifies scientific management, not departs from it. The principles Taylor said in 1911 — task breakdown, knowledge transfer, eliminating worker choice, and monitoring — are copied in algorithmic management systems with accuracy and reach. Agency Theory explains the organizational reason; Weber's theory explains cultural resonance; and labor process theory explains its function. These theories show digital Taylorism as a choice about management and labor, not just tech.

The three empirical cases examined — Amazon’s fulfillment centers, Yandex Taxi’s platform labor system, and Sberbank’s AI-driven workforce restructuring — demonstrate the same underlying logic across different organizational contexts: the encoding of managerial functions in algorithmic systems that direct, monitor, evaluate, and discipline workers with minimal human intervention. Taken together, the cases reveal a pattern of escalation: from physical labor in warehouses, to platform labor among independent contractors, to knowledge work in a major financial institution. This progression suggests that digital Taylorism is not confined to low-skill, easily observable tasks; it is expanding into domains where the gap between what algorithms can measure and what organizations actually need is widest.

The analysis also reveals critical tensions that classical management theory was not equipped to anticipate. The Dynamic Capabilities paradox — that algorithmic control may systematically suppress the innovation capacity that competitive firms require — represents a fundamental strategic challenge. The accountability gap created by opaque algorithmic decision-making represents a governance challenge that regulatory frameworks are only beginning to address. And the human capital erosion documented across the cases — in injury rates, organizational shock, and the loss of tacit knowledge — represents a social and organizational cost that short-term productivity metrics systematically obscure.

The managerial implications of this analysis are several. First, organizations deploying algorithmic management systems should account for the full costs of these systems — including worker health, turnover, skill erosion, and long-term capability degradation — not merely the efficiency gains visible in short-term productivity metrics. Second, the design of algorithmic management systems should incorporate transparency mechanisms enabling workers to understand, question, and appeal decisions that affect their work and compensation; opacity is not a technical necessity but a design choice. Third, organizations operating in knowledge-intensive sectors should be alert to the cultural effects of algorithmic control: systems that optimize for measurable outputs may inadvertently suppress the unmeasurable but organizationally critical behaviors — creativity, collaboration, experimentation — that Dynamic Capabilities theory identifies as the true sources of sustained competitive advantage.

For future research, several directions appear particularly promising. Comparative institutional analysis of how the same algorithmic management systems operate differently across regulatory contexts — the EU, the US, and Russia represent meaningfully different regulatory environments — would advance understanding of the institutional determinants of algorithmic management outcomes. Longitudinal research tracking the effects of algorithmic management deployment on organizational innovation capacity over time would provide empirical evidence on the Dynamic Capabilities paradox identified in this paper. And as the Sberbank case illustrates, the application of algorithmic management to knowledge work is still in its early stages and largely undocumented in the academic literature, representing a significant gap that future research should address.

The fundamental question that digital Taylorism poses for twenty-first-century management is one Taylor would have recognized: who controls the knowledge of work, and in whose interests is that control exercised? The algorithmic form is new; the substance is as old as the organization of labor itself.

References:

1. Taylor, F. W. (1911). The principles of scientific management. Harper & Brothers.
2. Lee, M. K., Kusbit, D., Metsky, E., & Dabbish, L. (2015). Working with machines: The impact of algorithmic and data-driven management on human workers. Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI '15) , 1603–1612. https://doi.org/10.1145/2702123.2702548
3. Petrovskaya, I. A., & Demchenko, V. S. (2023). Algorithmic management: an empirical study. study. Lomonosov Economics Journal, 58(6), 109–132. https://doi.org/10.55959/MSU0130–0105–6–58–6–7
4. Popova, E.V., & Beibutova, A.A. (2025). The evolution of human resource management technologies in the context of digital transformation: Essence, content, and development trends. Economics: Yesterday, Today, Tomorrow, 15(11A), 379–387. https://doi.org/10.34670/AR.2026.96.60.037
5. Ball, K. (2022). Surveillance in the workplace: Past, present, and future. Surveillance & Society, 20(4). https://doi.org/10.24908/ss.v20i4.15805
6. Baiocco, S., Fernandez-Macías, E., Rani, U., & Pesole, A. (2022). The algorithmic management of work and its implications in different contexts (Background Paper No. 9). International Labor Organization. https://www.ilo.org/publications/algorithmic-management-work-and-its-implications-different-contexts
7. Manyika, J., Lund, S., Bughin, J., Woetzel, L., Stamenov, K., & Dhingra, D. (2016). Digital globalization: The new era of global flows . McKinsey Global Institute. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/digital-globalization-the-new-era-of-global-flows
8. Schwab, K. (2016). The fourth industrial revolution .
9. Günsel, A., & Yamen, M. (2020). Digital Taylorism as an answer to the requirements of the new era. In B. Akkaya (Ed.), Agile business leadership methods for Industry 4.0 . Emerald Publishing. https://doi.org/10.1108/978–1–80043–380–920201007
10. Sostero, M., Milasi, S., Hurley, J., Fernández-Macías, E., & Bisello, M. (2020). Teleworkability and the COVID-19 crisis: A new digital divide? (JRC Working Papers on Labour, Education and Technology No. 2020–05, JRC121193). European Commission Joint Research Centre. https://publications.jrc.ec.europa.eu/repository/handle/JRC121193
11. Braverman, H. (1974). Labor and monopoly capital: The degradation of work in the twentieth century. Monthly Review Press.
12. Weber, M. (2001). The Protestant ethic and the spirit of capitalism (S. Kalberg, Trans.). Blackwell. (Original work published 1904–1905)
13. Jensen, M.C. (2000). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure.
14. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360. https://doi.org/10.1016/0304–405X(76)90026-X
15. Kellogg, K. C., Valentine, M. A., & Christin, A. (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals , 14(1), 366–410. https://doi.org/10.5465/annals.2018.0174
16. Fayol, H. (2016). General and industrial management. Ravenio Books. (Original work published 1916)
17. Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350. https://doi.org/10.1002/smj.640
18. Delfanti, A. (2019). Machinic dispossession and augmented despotism: Digital work in an Amazon warehouse. New Media & Society, 23(1), 39–55. https://doi.org/10.1177/1461444819891613
19. Strategic Organizing Center. (2022). The injury machine: How Amazon's production system hurts workers. https://thesoc.org/resources/the-injury-machine-how-amazons-production-system-hurts-workers/
20. Ogunleye, O. S., & Kalema, B. M. (2020). Evaluation of algorithmic management of digital work platform in developing countries. In S. K. Sood & R. Kaur (Eds.), Digital work platforms. IntechOpen. https://doi.org/10.5772/intechopen.94524
21. Geliskhanov I. Z. Digital platform governance structures: A case study of Yandex Taxi. Voprosy Ekonomiki. 2024;(9):76–106. (In Russ.) https://doi.org/10.32609/0042–8736–2024–9–76–106
22. ILO. (2021). World employment and social outlook 2021: The role of digital labour platforms in transforming the world of work. International Labour Organization. https://www.ilo.org/wcmsp5/groups/public/---dgreports/---dcomm/---publ/documents/publication/wcms_771749.pdf
23. Manyika, J., Lund, S., Bughin, J., Woetzel, L., Stamenov, K., & Dhingra, D. (2016). Digital globalization: The new era of global flows. McKinsey Global Institute. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/digital-globalization-the-new-era-of-global-flows
24. Oztemel, E., & Gursev, S. (2019). Researching the effects of automation and digitalization on manufacturing companies’ productivity in the early stage of Industry 4.0. Procedia Manufacturing, 39(1), 886–893. https://doi.org/10.1016/j.promfg.2020.01.401
25. Sberbank to cut 20 % of central office staff based on AI performance evaluation. (2025, November 19). RBC. https://www.rbc.ru/finances/19/11/2025/691e05f39a7947863a9cc7d8
26. How to protect your job from an algorithm: Dismissals based on AI evaluation. (2026, February 10). RBC Education. https://www.rbc.ru/education/10/02/2026/69836f859a79472d09e0ce31
27. Gref, H. (2026, February 12). Statement at the Congress of Public Administration. Cited in: Sberbank pauses staff reduction after dismissing 20 % of employees. iXBT Live. https://www.ixbt.com/live/sw/sberbank-priostanovil-sokraschenie-personala-posle-vysvobozhdeniya-20-sotrudnikov.html
28. The historical evolution of management control theories. Generated by FigureLabs AI upon request by the author, 2026 March, 9. Promt: ‘1911 — Taylor Scientific Management: ‘the one best way’, standardization, separation of conception and execution. 1916 — Fayol Administrative Management: five functions of management (planning, organizing, commanding, coordinating, controlling) the theoretical foundation onto which algorithms will later be mapped. sociological grounding. 1922 — Weber Bureaucracy as rationalization: formal rules, impersonality, hierarchy. The iron cage of rational control. 1976 — Jensen & Meckling Agency Theory: the principal-agent problem. 1974 — Braverman Critique of Taylorism: digital realization. 2015 — Lee et al. First academic conceptualization of algorithmic management. 2020 — Kellogg et al. Six mechanisms of algorithmic control, the ‘6 R’s’. 2022 — Baiocco et al. Algorithmic management defined as ‘the use of computer-programmed procedures for the coordination of labor input in an organization’. 1997 / 2007 — Teece et al. Dynamic Capabilities Theory: sustained competitive advantage requires flexibility, autonomy, and experimentation.