At present, the international political and economic situation is complicated, and the global economic recovery is weakening, creating an increasingly volatile and uncertain external environment for business. To survive this crisis, organizational managers are paying increasing attention to training team agility and employee agility. Enterprise social media (ESM), an indispensable digital technology in today’s enterprises, is recognized as an effective enabler of communication, collaboration, and social relationships within organizations. It also offers enterprises a greater possibility of enhancing the capabilities of teams and employees. Despite this, there is still a lack of in-depth discussion on how ESM usage affects team agility and employee agility. In order to effectively promote agility, it is necessary to explore whether and how the use of ESM affects team agility and employee agility in the context of ESM. Based on the team cohesion theory, we propose a multilevel model to examine how team ESM usage affects team agility and employee agility by promoting team cohesion and how team interdependence moderates the impacts of team cohesion on team agility and employee agility.

To test the proposed theoretical model and research hypotheses, we conducted a multi-wave and multi-source questionnaire survey through a professional survey agency. At Time 1, team subordinates reported on team ESM usage and team interdependence and diversity, and provided demographic information. At Time 2, team subordinates rated team cohesion. At Time 3, team leaders rated their subordinates’ agility and team agility. All measurement items were adapted from prior related studies. We received 386 team leader-team subordinate valid paired samples in 84 work teams for hypothesis testing. We calculated Cronbach’s α, average variance extracted, and conducted confirmatory factor analyses to evaluate the reliability, convergent validity, and discriminant validity, respectively. The multilevel structural equation modelling method was used to assess the structural model in Mplus 7.4 software.

The results of the data analysis support our research model and hypothesis. Results showed that: (1) Team ESM use had a significantly positive effect on team cohesion (r =.47, p <.001); (2) Team cohesion has a significantly positive effect on both team agility (r =.35, p<.05) and employee agility (r =.58, p <.001); (3) Team interdependence positively moderates the effects of team cohesion on team agility (r =.27, p<.01) and employee agility (r =.17, p <.01); (4) Team cohesion mediated the the influence of team ESM use on team agility (r =.18, 95%CI = [.02,.40]) and employee agility (r =.19, 95%CI = [.03,.35]).

Our study made contributions to the theory. First, this study explores in depth the impact of ESM use on team agility, thus effectively expanding the research on the antecedents of team agility. Second, this study integrates ESM research into agility at both the team and individual levels and adopts the team cohesion theory to reveal the mediating mechanism of ESM use affecting team agility and employee agility, so as to advance the existing research on the relationship between ESM and agility. Third, this study enriches the research on team interdependence to a certain extent by exploring the moderating effect of team cohesiveness on team agility and employee agility.

This study has several limitations. First, the data in this study is collected in China. Considering the essential differences in the cultural backgrounds of China and the West, the use of ESM and its impact may differ. Therefore, future studies should be conducted under different research backgrounds to compare the influence of ESM use on employee and team agility in different cultural backgrounds. This can improve the external validity of research conclusions. Second, although this study conducted a multi-wave and multi-source questionnaire survey to collect data, the questionnaire survey method is inevitably subjective. Third, since the time interval between waves in this study was only one week, future studies should extend it to better predict the causal relation between variables.