Navigating the maze: the effects of algorithmic management on employee performance

The increasing integration of algorithms into workplace management, termed algorithmic management, is reshaping organizational practices and employment patterns. While research exists on algorithms' impact on employee rights and autonomy, limited understanding exists regarding its effect on creative and adaptive performance, critical for organizational success in dynamic environments. This study addresses this gap by examining how algorithmic management influences employee creative and adaptive performance. The study utilizes the Ability-Motivation-Opportunity (AMO) theory, positing that algorithmic management impacts employee performance by affecting their improvisation capability – the ability to creatively and spontaneously adapt to unexpected challenges. Algorithmic dependence, reflecting employee motivation and reliance on the system, is hypothesized to moderate this relationship. The study aims to contribute to the literature by providing a nuanced understanding of how algorithmic management affects individual employee performance in a complex interplay between technology and human agency.

Existing research on algorithmic management primarily focuses on employee rights, autonomy, and psychological well-being, neglecting its impact on creative and adaptive performance. Studies have shown that algorithmic systems can influence employee behavior, potentially leading to reduced autonomy and flexibility. AMO theory, emphasizing the interaction of ability, motivation, and opportunity to influence performance, provides a suitable framework for this study. The researchers highlight the importance of improvisation capability, defined as the capacity to spontaneously reorganize resources to adapt to novel situations, as a mediator linking algorithmic management to creative and adaptive performance. The concept of algorithmic dependence, reflecting the extent to which employees rely on algorithmic systems, is identified as a crucial moderator.

This study employed a quantitative approach using a survey questionnaire administered to 327 employees of an information technology service firm in northern China. The firm utilizes a well-developed algorithmic management system. Data collection was conducted in three separate rounds over a three-month period to mitigate common method bias. The survey measured algorithmic management using a scale adapted from Parent-Rocheleau et al. (2023), encompassing algorithmic monitoring, goal-setting, scheduling, performance management, and compensation. Improvisation capability was measured using an adapted scale from Vera and Crossan (2005), creative performance using an adapted scale from Gong et al. (2009), and adaptive performance using an adapted scale from Han and Williams (2008). Algorithmic dependence was measured using a modified scale from Shu et al. (2011). Control variables included age, gender, education, and job tenure. Confirmatory factor analysis was conducted to assess the discriminant validity of the constructs. Structural equation modeling (SEM) using Mplus 8.3 was employed to test the hypotheses, including mediating and moderating effects.

The study's key findings support the hypothesized relationships. Algorithmic management was significantly and negatively related to improvisation capability (β = -0.491, p < 0.001), indicating that algorithmic management inhibits employees' ability to improvise. Improvisation capability was significantly and positively related to both creative performance (β = 0.381, p < 0.001) and adaptive performance (β = 0.410, p < 0.001), confirming its mediating role. Improvisation capability significantly mediated the relationship between algorithmic management and both creative (indirect effect = -0.187, p < 0.001) and adaptive performance (indirect effect = -0.201, p < 0.001). Algorithmic dependence significantly moderated the relationship between algorithmic management and improvisation capability (β = -0.245, p < 0.05). Furthermore, algorithmic dependence moderated the indirect effects of algorithmic management on both creative and adaptive performance through improvisation capability. The negative indirect effects were stronger at higher levels of algorithmic dependence, indicating that high dependence amplifies the negative impact of algorithmic management on performance. The differences in indirect effects between high and low algorithmic dependence were significant for both creative performance (difference = -0.187, p < 0.05) and adaptive performance (difference = -0.201, p < 0.05).

The findings demonstrate a detrimental effect of algorithmic management on employee creative and adaptive performance, mediated by reduced improvisation capability. This effect is amplified by high algorithmic dependence, highlighting the importance of employee motivation and reliance on the system. The study's theoretical contributions lie in expanding the understanding of algorithmic management's impact beyond autonomy and well-being to include performance outcomes, identifying improvisation capability as a crucial mediating mechanism, and uncovering algorithmic dependence as a critical moderator. These findings provide a more nuanced understanding of the complex interplay between technological implementation and employee behavior in the workplace.

This study provides valuable insights into the impact of algorithmic management on employee performance. The negative relationship between algorithmic management and performance, mediated by improvisation capability and moderated by algorithmic dependence, suggests a need for a more balanced approach to algorithmic implementation. Future research should explore the generalizability of these findings across diverse organizational settings and cultures, utilize multiple data sources to reduce reliance on self-reported data, and investigate other potential moderators such as transparency and fairness perceptions. The study contributes to the growing body of research on algorithmic management by offering a more nuanced and comprehensive understanding of its implications for employee capability and performance.

The study's limitations include its reliance on self-reported data from a single organization in China, potentially limiting the generalizability of the findings. While efforts were made to mitigate common method bias, the use of self-report measures could introduce some response bias. The cross-sectional design restricts the ability to make definitive causal inferences. Future studies should consider longitudinal designs and incorporate data from multiple sources, including supervisor evaluations, to enhance the robustness of the findings. The focus on algorithmic dependence as a moderator excludes other potentially relevant factors, highlighting an area for future exploration.