BOOK REVIEWS

The 996 Working Pattern in Chinese Internet Firms: How Hegemonic Despotism Promotes Long Working Hours for Employees

Xiaojing Zheng is an assistant researcher of labour and social security in the School of Public Administration at South China University of Technology, No. 381 Wushan Road, Tianhe District, Guangzhou, China, 510640 (zxj612523@163.com).

Zitong Qiu (corresponding author) is a research assistant at the Labour Dispute Office, Chinese Academy of Labour and Social Security, No. 17 Huixin West Street, Chaoyang District, Beijing, China (qzt@ruc.edu.cn).

Introduction

The rapid growth of China’s internet economy over the past decade has been accompanied by increased work time and work intensification. Between 2014 and 2017, work overtime and overtime pay took the highest percentages in the total number of labour disputes accepted by Beijing’s Haidian District, an internet industry hub in China. From March until April 2019, the online “996 ICU” campaign rallied more than 200,000 Chinese internet firm employees complaining about the 996 working schedule,[1] epitomising the escalation of ubiquitous workplace dissatisfaction towards long working hours. A series of sudden deaths among young information technology (IT) engineers from internet firms in recent years[2] has painfully raised the alarm about the health and life risks of excessive working hours. Some firms have begun to react to the phenomenon by launching various work-life programmes, such as requiring people to leave work at 6:00 p.m. and cancelling the big-small workweek.[3] Yet the results are disappointing,[4] and long working hours are still highly persistent, especially after the outbreak of the Covid-19 pandemic. This study examines the mechanisms through which excessive working hours, typically the 996 working pattern in Chinese internet firms, are generated and persist, and identify what it essentially means in the Chinese context.

Scholars have proposed various reasons to account for employees’ long working hours. They have largely emphasised social and economic determinants (McGovern et al. 2007: 7; Golden 2009), management control approaches (Williams, Berdahl, and Vandello 2016), or self-chosen modes (Mazmanian, Orlikowski, and Yates 2013). Although these approaches capture important facets of long work hours, they do not fully and exactly explain the development of the idiosyncratic 996 working pattern with fixed minimum extra working hours in Chinese internet firms. Building on theoretical work derived from the labour process tradition and more recent developments, we identify hegemonic despotism as the underlying mechanism that internet firms use to elicit excessive work effort from employees and foster their consent.

Originally, hegemonic despotism was a hybrid regime identified by Burawoy when neoliberalism began to dominate external sociopolitical conditions since the 1980s (Burawoy 1983). Here, it responds to our research questions. We first uncover one prominent and coercive mechanism behind the 996 working pattern that we call informal-flexible-allied despotism, in which the firm manages to make the 996 working schedule a base line by implementing top-down informal supervision that places employees at risk of termination if their attitudes and behaviour are inconsistent with the requirement. Besides weakened job stability, employees are constantly exposed to the risk of being dismissed due to institutional settings such as the forced performance distribution system. The Covid-19 pandemic further intensified the downsizing risk, and workers perceive few alternatives in the difficult wider economic environment but to increase their effort as a means of reducing the chances of job loss. While atomised individuals identify more strongly with their career, employers form alliances to set barriers by sharing their work records among employers and making comments, which forces individuals to make self-adjustments to the overwork requirements for fear of being denied employment in the industry.

Second, we identify a hegemonic mechanism that supplements the internet firms’ coercive controls, signifying management’s attempts to coordinate interests with workers and to mobilise their consent to increase performance. These include normative controls, in which internet firms establish ideal worker norms, enrich rewards for putting in more work effort, and enhance employees’ internal drive to increase work effort; career identification defines internet firm employees’ subjectivities under neoliberalism and coercive controls. The orientation to pursue career development suggests a positive alignment with dominant hegemonic discourse in which striving and learning are presented as key approaches to realise career expectations.

By conducting a comparative case study in six Chinese internet firms, this paper makes two contributions to the existing literature. First, our analysis enriches the literature on long working hours by uncovering the mechanism of hegemonic despotism through which coercive practices facilitate the 996 working pattern and hegemonic mechanisms elicit employees’ consent. Second, we contribute to labour process studies by constructing an updated version of hegemonic despotism in internet firms, through which informal-flexible-allied despotism suggests a multifaceted and cumulative coercive effect, while normative control and employees’ career identification reinforce the hegemonic coordination of interests. We conclude by discussing the practical implications of making changes to the 996 work regime.

Existing explanations for long working hours

One central question in the literature on long working hours asks why do employees work long hours? Numerous studies have proposed theories to address this question, with relevant research covering inter alia the topics of extreme work, overemployment, work intensification, extensive work effort, long work hours, and overtime work.

Research at the macrolevel focuses on the social and economic determinants underlying the length of working hours and intensity of overwork. Analysts examine long working hours and associate them with a list of causes: social and economic environment (Gascoigne, Parry, and Buchanan 2015), labour supply and demand conditions (Golden 2009), institution and labour regulation (Green et al. 2022), trade union strength (Gallie 2009), the intensity of market competition (McGovern et al. 2007), the fast update of advanced technology (Ayyagari, Grover, and Purvis 2011), work and employment conditions (Burchell 2011), cultural characteristics (Wang 2020), consumerism (Schor 1999), and the work-and-spend cycle (Raphael, Douglas, and Morris 2006). This stream of literature suggests that long working hours reflect the effect of the social and economic environment on work and employment relations.

It is well established in the literature that employers impose long working hour constraints on employees (Bryan 2007). Firms are incentivised by increasing profit and reducing costs to lengthen the hours of existing employees rather than hire new ones (Golden 2009). A traditional account of management practices to obtain maximum effort from employees, drawing on a range of literature, sees its origins in incentive-and-restraint mechanisms such as the performance appraisal, which largely relies on working hours to determine employees’ promotions and salaries (van Echtelt, Glebbeek, and Lindenberg 2006). Modern management frames long hours as the very nature of work. Ideologically, organisations encourage workers to uphold the “ideal worker” norm, which equates overwork with loyalty and commitment (Williams, Berdahl, and Vandello 2016). By instiling in employees a profound sense of personal commitment to the goals and values of the company, management successfully induces employees to engage in self-surveillance and monitor their own behaviour to align with company interests and requirements (Anteby 2008). Regarding management practices, high-performance work practices, including teamwork, quality circles, and project schemes are implemented to urge employees to prioritise work over all other responsibilities (Wynn and Rao 2020). This has been found to be negatively associated with workload, working hours, and mental and physical health (Burke, Singh, and Fiksenbam 2010) given the inherent requirements of rapid workflow, short deadlines, and work uncertainty.

Why do employees choose to work overtime? Existing studies propose two mechanisms. According to the first, self-motivation is the reason individuals make and sustain work effort, which includes extrinsic motivation and intrinsic motivation (Ryan and Deci 2000). The extrinsic motivation for prolonging working hours is to achieve an external objective, such as salary increases, career promotions, and status advancements (Gicheva 2013). Intrinsic motivation refers to the intrinsic rewards, process benefits, satisfaction, and amenities acquired through working long hours. According to the second mechanism, employees’ autonomy and control over their work lead to long working hours. Although the “job demands-control model” suggests that increased job control decreases job-related strain and psychological distress, higher job control is also related to extensive working hours. As scholars have indicated, apparently flexible work arrangements and greater responsibilities are effectively leading to employees being available 24/7 (Mazmanian, Orlikowski, and Yates 2013). Increasing job control and job autonomy help to explain why most employees insist that their sustained pattern of overwork is “self-chosen.”

There is a lack of research into the 996 working pattern in internet firms, yet some internet firm-based work-related studies offer insights into this subject. Critically, Liang (2019a) suggests that since corporate culture in internet firms has been replaced by the external market competition philosophy, a working environment that cultivates long working hours has emerged. Yan (2020) identified a new form of management control: self as enterprise, under which R&D employees comply with the market mechanism and put their best effort into their own professional work. Hou and He (2020) point out that working overtime is the outcome of inter-construction between management control and labour autonomy. Wang (2020) investigates how management in the hospitality and manufacturing sectors use Confucian culture and controls to impose a 996 work regime that constitutes modern slavery. From the management efficiency perspective, Jenkins and Delbridge (2013) argue that a “soft approach” emphasising the “human” dimension of management is conducive to promoting employee engagement. In contrast, a “hard approach” focusing on the “resource” features of human resource management leads to high levels of employee disengagement. Tran (2017) provides a case study on Google to examine its management practices in improving employee engagement, such as creating a more comfortable and friendly working environment, tolerating mistakes and helping staff correct them, and providing attractive workplace perks.

To sum up, the existing literature has been fruitful in analysing long working hours, but little has focused on the 996 working pattern in internet firms. This gap is particularly problematic given the ubiquity of the 996 working pattern among internet firms and the prominent negative effects. The purpose of this study is to explore how the idiosyncratic 996 working pattern with fixed minimum extra working hours in Chinese internet firms developed and is sustained – in other words, a theorising of the management control approaches and mechanisms that shape employees’ attitudes and behaviours. We do so by reviewing Burawoy’s concept of hegemonic despotism and recent developments in labour process research in the overview that follows.

Hegemonic despotism and internet firms

Hegemonic despotism is one of the production regimes identified by Burawoy (1983, 1985). Based on two types of political apparatuses, “despotic” versus “hegemonic,” Burawoy constructed the production regimes theory in accordance with the degree to which the regime relies on coercion versus consent. For Burawoy, the capitalist labour process is successively characterised by market despotism and hegemonic regime. In explaining variations among these types, he placed weight on the degrees of market competition, state intervention, and workers’ access to means of subsistence apart from wage labour itself. With respect to contemporary capitalism since the 1980s, Burawoy (1985: 263-5) observed how the external sociopolitical conditions for consent within the labour process were being eroded by the trend of neoliberalism. Instead of a hegemonic regime, he recognised a hybrid regime that he called “hegemonic despotism,” through which globally oriented firms take advantage of the long-term process of deregulation, financialisation, and weakening of employment to emphasise coercive workplace practices and shape how workers define their work, leaving workers even more exposed to ongoing bouts of unemployment until they perceive no alternative but to align their interests with those of the firm (Dörflinger, Pulignano, and Vallas 2021).

Accordingly, more recent efforts to advance the hegemonic despotism theory have put their emphasis on more coercive control methods (Wood 2021). For example, Vallas, Johnston, and Mommadova (2022) found that Amazon uses highly coercive mechanisms that they called techno-economic despotism to control the labour of its warehouse employees. This despotism is supplemented by three hegemonic mechanisms to foster workers’ consent. With observations of Silicon Valley’s high-tech assembly workers, Chun (2001) identified flexible despotism as the mechanism that intensifies insecurity and uncertainty among this group of workers. The trend to emphasise coercive control is explicitly manifested in platform work, where algorithmic technology-driven surveillance systems lead the platform control toward an algorithmic despotism (Griesbach et al. 2019). Realising hegemonic despotism’s limitations in capturing workers’ consent, Purcell and Brook (2022) proposed a new version of hegemonic despotism, in which the authors provide an explanation of how platform-mediated gig workers’ consent is constructed by shaping individual subjectivity.

Internet firms are representative of the new business models that have emerged in the digital economy. They rely on network effects (Rahman and Thelen 2019) and the winner-takes-all mode (Zysman and Kenney 2018) to create and capture value. To compete for the market, internet firms pay more attention to management practices that enhance employees’ work effort and improve job performance and innovation outcomes. As the question labour process theory pursues is how management seeks to extract the full value of labour power, typically by maximising workers’ efforts (Vallas, Johnston, and Mommadova 2022), it is meaningful to investigate working time regimes in the digital economy driven by more effort-biased technologies (Green 2004), given that the labour control of internet firms is largely reflected in their effort to elicit higher levels of worker effort.

Research design and method

This study presents a qualitative study of six organisational cases to examine the 996 working pattern in internet firms. Following the replication logic and comparative guidelines that are essential to multicase analysis (Yin 2009), this study aims to discover the causal relationships, process dynamics, and mechanisms behind the research questions in an inductive way.

The six samples in this study were selected with a theoretical sampling strategy. Table 1 summarises the major characteristics of these case firms. The companies and individuals are disguised to ensure confidentiality. We considered several factors in selecting the cases. First, these six firms are typical of the long working hours problem and representative of the internet industry, making them illustrative samples for the research questions (Eisenhardt 1989). Second, our study purposefully sought to select internet firms with different industry statuses and ownership patterns to ensure a large degree of variability among different cases for developing theories (Yin 2009). A third factor considered for case selection was access to informants. The case firms eventually included in this study were among the first in which an effective quantity of employees agreed to interviews. All informants were gained through the snowball sampling techniques.

Table 1. A summary of the major characteristics of the case firms

No.	Firm	Industry status[5]	Ownership pattern	Firm description
1	Firm A	Leading	Privately owned	Firm A is one of the leading internet firms in China and has developed corporate values and management institutions with its own characteristics. Firm A is representative in implementing the 996 working schedule.
2	Firm B	Top-ranked	Privately owned	Firm B is one of the most influential e-commerce type internet firms in China. Its employees were observed to be one of the latest ones to leave work (after 10 p.m.) in the industry.
3	Firm C	Front-ranked	Privately owned	Firm C is among the front-ranking firms in the e-commerce sector. As a transitional firm from a traditional sales company, it also has the salient problem of long working hours.
4	Firm D	Top-ranked	Privately owned	As a new e-commerce internet firm, Firm D offers industry-leading salaries to attract talents, but issues of long work hours have damaged its reputation in recent years.
5	Firm E	Top-ranked	Privately owned	Firm E is representative of the mobile internet field, which is characterised as both growing fast and working hard.
6	Firm F	Moderate	State-owned	Firm F is an e-commerce firm specialising in food supply. It is now encountering business difficulties, but also has problems with overtime work.

Source: authors.

This study is based on field research in the above six firms between 2018 and 2021. The fieldwork entailed semi-structured interviews and informal talks with 102 employees and 26 managers. Each interview lasted on average of an hour; some informants were interviewed through online instant messaging apps such as WeChat due to personal preference or Covid-19 pandemic controls. Field observations and archival research were also conducted to triangulate the findings. An overview of data structure and data collection is shown in Table 2.

Table 2. A summary of data collection among case firms

Firm	Number of informants	Gender (male/female)	Age (<30/30-35/>35)	Position rank (grassroots/ midlevel)	Education background (≤undergraduate/ graduate)
Firm A	29	17/12	13/14/2	21/8	13/16
Firm B	30	14/16	8/16/6	25/5	16/14
Firm C	25	12/13	9/12/4	23/2	15/10
Firm D	18	10/8	11/7/0	15/3	10/8
Firm E	16	10/6	4/7/5	10/6	7/9
Firm F	10	4/6	4/4/2	9/1	5/5
Total	128	67/61	49/60/19	103/25	66/62

Source: authors.

 

The data were analysed by utilising NVivo 12.0 software, following the steps of grounded theory methodology. Replication and comparison were carried out in two stages. First, we selected a typical case that could serve to develop a conceptual framework. Firm A was chosen for its pioneering role in implementing the 996 pattern and its supportive institutional settings. Second, we included cases that could confirm, disprove, and extend our concept categories identified earlier. We added firms with different industry status and then different ownership, and repeated the comparative process until commonalities and divergences across the six cases were identified and an integrative framework was established.

To assure reliability and validity, we first examined the theoretical saturation by adding new data and new case firms, and we didn’t observe new concepts and relations. Then we measured intercoder reliability to further verify whether different coders would code the same data the same way. Following Campbell et al. (2013), we calculated the proportion of intercoder agreement, which means we divided the number of coding agreements by the number of agreements and disagreements combined. For instance, with three coders encoding 20 original texts, we saw 48 pairwise agreements out of a total of 60 possible. Thus, intercoder reliability is 0.8. According to the intercoder reliability range of [0.7, 0.94] (Fahy 2001), our result is acceptable. Due to space limitations, we do not include the intercoding process, but it is available upon request.

Findings

Understanding the 996 working pattern

The working schedules of the case firms exhibit a marked 996 working pattern, which is manifested in two characteristics, as our field work revealed.

• Implicit coercion

Implicit coercion was coded in all interviews and matched with interviewees’ statements stressing what the 996 working pattern is. Fixed extra working hours and informal requirements were subsumed under this facet. Fixed extra working hours refers to requiring employees to complete a minimum of extra working hours at the workplace. Take the 996 working schedule as an example, people are supposed to work six days a week, with daily work starting at 9 a.m. and lasting until at least 9 p.m., beyond the declared working schedule such as 965. In this way, overtime work is guaranteed by covertly renewing the working schedule. Given the demand of internet firms for increased work effort, this method is efficient in unifying the length of off-work time and work time. This quantifying of employees’ increased work effort makes it necessary for employees to be at work within the 996 schedule, leaving most employees with no alternative but to comply. Given the legal risk of violating labour regulations on working hours and the consequent damage to corporate reputation, internet firms never make any official formal arrangements for long working hours. Instead, management forces employees to follow their renewed working schedule through informal coercive means such as verbal demands, monitoring, and penalising employees who resist managerial control.

• Weak discretion

Weak discretion means employees have very little control over their work hours. This is another key facet most interviewees mentioned when talking about their understanding of the 996 work pattern. This facet encompasses two subcategories: involuntary overtime and lack of flexibility. Voluntary overwork refers to hours that employees willingly work beyond contractual demands (Avgoustaki and Canibano 2020), encouraged by incentives such as promotions, career opportunities, or self-achievement. Employees working under the 996 working pattern are forced to complete the fixed minimum of extra hours regardless of their willingness. There is the possibility that people work within the 996 working pattern voluntarily for purposes of personal striving or learning, but they are still constrained by the fixed overtime framework. Due to pressure from supervisors, colleagues, and cultural factors, individuals often work longer hours than they would have initially preferred (Golden and Wiens-Tuers 2005). Regarding flexibility, management always uses the flexible work schedule as an excuse to explain employees’ long working hours (Liang 2019b), but the flexibility applies only to extended working hours. As we observed, unless they have worked in a 996 arrangement, employees are largely deprived of schedule discretion, such as when or where to work extra hours and how long they should stay at the office. In this respect, the 996 working pattern is by no means a flexible working pattern.

In sum, the 996 working pattern in internet firms can be defined as an informal working hour norm that forces employees to work fixed extra hours at the workplace. It is a pattern with varied manifestations across cases, and not only the 996 working schedule, but also the 10107 or 007 schedule[6] can fall into this category. Prior studies have documented some types of working hours, such as the flexible working system, voluntary overtime, or mandatory overtime, but while these share some similarities with the 996 working pattern, implicit coercion and weak discretion are the characteristics that set the 996 working pattern apart from these other forms. Table 3 provides an overview of these various types of working times.

Table 3. An overview of working time types

Working time type	Working length	Coerciveness	Willingness	Discretion
Flexible working	Optional	Non-coercive	Voluntary	Discretionary
Mandatory overtime	Fixed extra hours	Formal coercion	Involuntary	Non-discretionary
Voluntary overtime	Optional	Non-coercive	Voluntary	Discretionary
996 working pattern	Fixed extra hours	Implicit coercion	Involuntary	Weak discretion

Source: authors.

 Hegemonic despotism in internet firms

Our data suggest that hegemonic despotism is the underlying mechanism through which the 996 working pattern is generated and persists in Chinese internet firms. In what follows, we uncover two distinct sources of hegemonic despotism: informal-flexible-allied despotism, and normative control-career identification-driven hegemony. The remainder of this section uses data from the case firms to describe each source in detail.

• Informal-flexible-allied despotism

Hegemonic despotism theory and recent development have increasingly emphasised firms’ growing use of coercive labour control. For the coercive mechanism of the 996 working pattern, our data speak to the confluence of surveillance, job insecurity, and employer alliances, yielding what we term informal-flexible-allied despotism through which employees have no apparent alternative but to comply with management’s working hour demands.

• Informal coercion

Internet firms never declare official or formal demands for long working hours. We characterise firms’ informal requirements and top-down surveillance as three distinct types: leaders’ covert requirements, indirect promotion through informal norms, and management’s direct supervision. First, corporate leaders, especially the founders, are in a critical position in setting the tone for long working hours (Blair-Loy 2003). Leader’s behaviour and attitudes convey what tends to be valued, and what yields rewards and avoids punishment. Given the role of signal and model in organisation, the leaders of firms use multiple methods to shape employees’ working hours. For example, the founder of Firm A used to walk around the office building after work, checking on whether employees were still working and then commenting on employees’ overtime work behaviour. He also publicly expressed encouragement of the 996 working schedule. The founder of Firm E once distributed red pockets in a WeChat group late at night to check whether employees were still working online. These informal and covert requirements signal leaders’ expectations of longer working hours, which would promote the spread of overtime culture throughout the whole firm. When market competition intensifies, firm leaders strengthen their supervision even further. As an important battlefield of e-commerce in China, “Singles’ Day”[7] always witnesses people from internet firms work through night with leaders at the scene. As one marketing employee from Firm A said: “Management is always at the scene, waiting to take photos and celebrate sales performance after the event” (A-19, 12 May 2019).

Second, establishing informal norms is also one of the approaches internet firms use to promote longer working hours for employees, especially when the firm encounters rigorous market competition. Norms are not self-enforcing (Drago, Black, and Wooden 2005); they are a presentation of the firm’s requirements and are advocated by the firm, like the ideal worker norm. Internet firms create informal norms to induce employees’ intensified work effort and commitment by emphasising the corporate competition crisis and individual values. For example, in order to increase their market share in community-level competition, the founder of Firm D proposed a hardcore striving work pattern calling for employees’ overinvestment in work, along with the introduction of a new big-small workweek schedule, which required employees to work seven days on the big week, and six days on the small week, leaving them with only two days of rest every month. As more tasks are consistently assigned on a big-small workweek schedule, it becomes harder for employees to refuse working longer and harder. There are also some other norms in different fields that substantially facilitate the formulation of fixed overtime patterns, such as allowing reimbursement of taxi fares if you leave work after 10 p.m., scheduling the late regular bus at 9 p.m., or offering an 8 p.m. meal subsidy.

Third, informal requirements by management, especially the immediate supervisor, are another factor that forces employees to work fixed extra hours. Management exerts its control predominately through its authority within the organisational structure (Hughes, Ginnett, and Curphy 1993). In internet firms, due to encouragement from the firm leader and corporate culture, working overtime has increasingly become part of the career system, and management at each level turns into the direct source enforcing longer working hours on subordinates. They are even able to exercise a range of punishments, such as blaming, allocation of extra duties, and lower grading in evaluation, to alert employees who don’t follow the overwork requirement or work fewer hours than expected. Top-down supervision therefore provides favourable conditions for the formulation of the 996 working pattern. The following words from a R&D engineer in Firm B are typical:

My team leader texted me when I worked fewer extra hours than he expected: “I don’t mean to blame you. You often leave before 9 p.m. while other colleagues are still working. Even some of our colleagues who live far from the office work longer than you. Their work attitude is what I think you should learn from” (B-10, 10 October 2018).

• Flexible coercion

Two conditions – external economic environment and internal institutional settings – generate the flexible employment norm and the imposition of coercive controls, depriving workers of the resources needed to resist and forcing them to accept a ratcheting up of work demands.

In the context of globalisation and technological change, the competitive pressures faced by companies have intensified. The aim to reduce labour costs while responding more rapidly to changing markets leads firms to base more on market-driven employment arrangements, offering less secure jobs, less predictable wages, and poorer welfare benefits (McGovern et al. 2007). By such means, employers shift the market risks onto employees. Internet firms rely on the winner-take-all business strategy to grow and gain profits (Zysman and Kenney 2018), which means that the firms need to fight for market dominance, or they might be eliminated by their competitors and disappear quickly in the market. In such an intensified competitive environment, internet firms tend to obtain high-level worker effort and produce greater performance given their intellectual effort-biased business essence. They use more flexible arrangements to stimulate employees and cope with market challenges, such as implementing diversified dismissal and exit mechanisms. Downsizing and job loss are also common in recent years due to the crisis caused by the Covid-19 pandemic.

Although most employees of internet firms hold standard employment, they are no strangers to job insecurity. Crucially, when employees describe their fear of being dismissed, they almost always reference the firm’s forced performance distribution system, which management uses to evaluate each employee and make elimination decisions. Take the 361 principle as an example. The 361 principle means employee performance is ranked into three categories: top-ranked, middle-ranked, and last-ranked according to fixed proportion ranges of 30%, 60%, and 10%. People who fall into the last 10% are deprived of the year-end bonus and promotion, and those in the last 10% for two consecutive years will be dismissed. Therefore, in order to avoid being assessed in the last 10%, people work harder and prolong their working hours to gain higher performance marks and win better evaluations. As one product manager of Firm A said: “No one wants to be the first to leave work, and 996 is the base line for the competitive overtime race” (A-20, 4 July 2019).

The forced performance distribution system is thus a deterrent to employees’ job security due to its close association with the risk of being dismissed, which causes deep anxiety among the workforce and escalates peer competition. Moreover, the 35-year-old crisis presents a similar pattern in that people who are unable to reach a higher position by the age of 35 might face the threat of dismissal.

• Allied coercion

Allied coercion means employers form alliances and force employees to comply with management by creating threats to employees’ career development interests. Allied coercion imposes a powerful management control on employees given that internet firm employees have increasingly prioritised boundaryless career development over other interests.

Workers have become continuously individualised and atomised in recent decades, experiencing shrinking collective labour power (McGovern et al. 2007). Under these conditions, they generally lack leverage against employer power, or to challenge coercion and resist management’s demands. Compared with workers from traditional industries, internet firm employees seem to have gained more professional power against employers as they are generally highly educated, and have higher professional abilities and stronger labour market competitiveness. Indeed, they are career-identified and would be able to switch to alternative jobs, which helps release them from being trapped in management coercion.

While employees value this outlet, employers impede their job transfers by formulating employer alliances. Associations such as HR Think-tank Alliance (HR zhiku lianmeng, HR智庫聯盟)[8] and Trust and Integrity Enterprise Alliance (yangguang chengxin lianmeng 陽光誠信聯盟)[9] have emerged in the internet industry and are absorbing more members. These associations declare themselves to be platforms aimed at promoting management and anticorruption, but employees perceive them as channels for employers to share information about employees and jointly enforce employee-antagonistic practices. Indeed, under the employer alliance, not only employees’ work attitudes, performance, salary levels, and reward and punishment records can be shared among member firms, but employees can also be marked as risky candidates in the shared system for any uncompliant ideas or behaviours from the employers’ perspective, such as complaints against management or labour disputes. As these comments would be considered by firms when making hiring decisions, this practice would negatively influence employees’ job transfers and harm their career development. Once marked as risky, the employee might be completely denied employment by all member firms. These associations have been called “employers’ siege strategy against workers.” As one software testing engineer from Firm C commented:

Once working attitude, performance, salary, and working hours are all recorded (by the HR Think-tank Alliance), there is the possibility of employees being marked viciously, which would make it difficult for them to find a new job (C-15, 18 February 2021).

Another post we extracted from online social media Zhihu (知乎) even reveals some shocking underlying operations:

I was threatened to give up my statutory redundancy payments for N+1.[10] Management said that if I disagree, they will add me onto the Trust and Integrity Enterprise Alliance (online post, 22 February 2020).

Given the harmful consequence, employees are coerced into making their attitudes and behaviours adhere to the requirements of the current employer. As enhancing work effort lies at the core of management practices in internet firms, employees work on a 996 working pattern basis to show compliance and obedience. While employees who deviate from the requirement face severe penalties, dismissal threats, or even unemployment risks, a good 996 pattern work record would help break the job transfer barrier.

For internet firm employees, any one of the above coercion methods is insufficient to produce despotism. It’s the cumulative effect of informal coercion, flexible coercion, and allied coercion that imposes despotic management control on employees. Under informal-flexible-allied despotism, informal overtime requirements and penalties exert excessive working pressures; the dismissal mechanism intensifies employees’ vulnerability in the labour market, making the competitive overtime race an important approach to reducing the risk of being fired; although high-tech employees can use job transfer as an outlet, employer alliances establish commenting and information sharing mechanisms to enforce employee compliance. Together with the high risk of job loss and permanent unemployment facilitated by informal-flexible-allied despotism, powerless internet firm employees have no alternative but to comply with management’s 996 working pattern demands.

Hegemonic mechanisms

Our data indicate that informal-flexible-allied despotism is not the only labour control mechanism on which management relies. As observation shows, even employees who suffer from overwork or health impairment at the workplace were willing to keep striving, which reminds us that there must be other mechanisms that foster employees’ consent to managerial authority. Indeed, our analysis identified two analytically distinct but empirically overlapping mechanisms – normative control and career identification – that seemed to secure employee compliance by eliciting their consent.

• Normative control

Normative control means that employers entice employees to identify with the firm via structural arrangements, enveloping them in an organisational culture that leaves little space for dissent (Kunda 1992). Our data suggest that the establishment of ideal worker norms, enrichment of rewards, and enhancement of employees’ motivation are management practices that help internet firms engage employees.

In internet firms, the ideal worker norm is usually closely connected with corporate goals. Firms are adept at presenting their business as positive and lofty, highlighting the firm’s morality, mission, and responsibility, such as “Making doing business easy,” or “Being the coolest company in the user’s mind.” These articulations are inspiring and encouraging to passionate people, making them desperate to join and work for these goals. In answering how to achieve the goals, firms formulate corporate values to illustrate what the firm needs and what employees should do, delivering the firm’s expectations and informing people of what thoughts and behaviours are encouraged and what are unwelcome. “Customers first,” “working hard,” “devotion to work,” “fast pace,” and “pursuit of the ultimate” are typical corporate values promoted by the firms. Through these concise words, firms empirically set the work guidelines and performance evaluation criteria for every specific position throughout the firm, which actually serves as the firm’s ideal worker norm. With the ideal worker norm from corporate values, employees are expected to exhibit an internalised belief in hard work and prioritisation of work above all other responsibilities. Indeed, employees who have been chosen as model employees are those loyal and diligent people. The ideal worker norm in this way efficiently links corporate goals with detailed position duties, providing a plausible reason for employees to invest in hard work, encouraging them to align with the firms’ goals, and generating consent despite management control and multifaceted coercion.

Adopting practices aimed at rewarding people who adhere to the ideal worker norm and disciplining those who do not is a necessary approach to strengthening job requirements and corporate values. Internet firms also make full use of motivation systems to enrich employees’ rewards for complying with the ideal worker norm (Wynn and Rao 2020). By associating employees’ working hours with evaluation results, internet firms make employees believe that employees who are fully engaged in work are the ideal workers in the firm and deserve being rewarded with scarce resources and opportunities such as pay raises, promotion, training programs, housing subsidies, and education subsidies for their children. Conversely, employees who deviate from the ideal worker norm face severe penalties. Under this system, employees are encouraged to work harder and longer to get pay raises or win other rewards. Here we refer to Firm B’s performance evaluation, which emphasises employees’ loyalty and hours worked, leaving people with the impression that longer working hours mean better salaries and promotions. With this guidance, employees even compete with their colleagues on the number of hours worked. Thus, by implementing practices that capitalise on material motivations, internet firms not only manage to reinforce the principle of hard work and enhance employees’ understanding of the ideal worker norm, but also induce employees to engage in performance-enhancing activities and lead them to adjust their behaviours to align with corporate interests.

Our data show that a substantial number of employees expressed little interest in getting rewards but maintained their work effort at the firm’s bottom line or in the middle range. This is inconsistent with the spirit of corporate value and contradicts the ideal worker norm. How can employees’ behaviour be shaped to obtain maximum discretionary effort from them? Internet firms see the power of perceived job insecurity as motivating employees’ internal drive to work harder. They use corporate values to warn employees of job insecurity and occupational crisis, urging them to make changes to keep their current jobs. Their methods can range from highlighting intensive labour market competition to stressing rapid skill obsolescence, or even instructing employees to believe they are very fortunate to hold a position in a firm that offers the 996 working pattern. When employees are immersed in job insecurity anxiety and a heightened sense of occupational crisis, the significance of internally driven work effort metrics becomes even more evident. Take Firm A as an example: the firm added some new items to their corporate values in 2019, among which “Today’s best performance is tomorrow’s baseline” is the very aphorism that acts on employees’ internal drive. It stimulates people by warning them of immediate job insecurity risks and persuades them to adopt solutions by making every effort to perfect their current tasks. With this method, internet firms subtly connect employees’ interests with the corporate interest, extracting additional effort in the name of securing employees’ jobs and positions.

• Career identification

Normative control represents the firm’s effort to encourage employees to align their attitudes and behaviours with the firm’s goals. Our data also reveal that the work time investment of internet firm employees is more than purely instrumental; career identification is what employees construct to make sense of their work and drive themselves towards purposeful work engagement.

Identification is a process by which people define themselves and is a root construct in predicting people’s behaviours (Ashforth, Harrison, and Corley 2008). Career identification views career development as essentially self-interested. Individuals facilitate the conception of self in their work context, meeting their own need for safety, affiliation, uncertainty reduction, and personal meaning through the realisation of career opportunities. Internet firm employees have constructed their identification with their careers in the context of a demanding labour market and competitive professional skills, and seek to secure their jobs and pursue possible career prospects through boundaryless career development under neoliberalism and coercive management controls. Our qualitative data analysis suggests that career development pressure and career development drives are two elements that shape employees’ career identification.

Career development pressure refers to the challenging occupational environment in which employees are required to be equipped with highly professional capacities such as problem solving, continuous learning, good performance maintenance, etc., in order to keep or gain their jobs; moreover, the implicit “35-year-old crisis” in the internet industry shortens career life covertly and intensifies the career development anxieties of professionals. Yet, internet firm employees continue to pursue their career development despite these pressures. For them, the internet industry is both challenging and rewarding, and people who work really hard have a greater chance of gaining higher earnings, quick promotions, and self-fulfilment. Even if they may have limited career opportunities in their current firm, they are confident that they can get better offers by transferring among firms. As one data analyst of Firm A said: “Outside firms always offer much better packages” (A-05, 17 June 2020).

These attractive career prospects and feasible career development paths provide internet firm employees with strong internal drives to remain engaged in the current job, which is the most rational approach to developing employability for meeting the target firms’ recruitment requirements on technical capacity, project experience, and employment record. As one R&D engineer from Firm F told us:

The target firms appreciate candidates who have developed technical skills and accumulated project experience in prior jobs because these candidates can directly contribute to the new firm. For example, someone who has experienced a billion-class service[11] knows the underlying risks of the project and how to avoid them (F-01, 8 April 2021).

Career identification therefore plays an important role in shaping internet firm employees’ work experience and the meaning attached to the labour process, with the confluence of career development pressure and career development drive. In order to pursue better careers, employees proactively engage in their current jobs, enhancing performance by creatively solving task problems, improving their employability via a self-directed learning process. Career identification in this way is not only strongly linked to employees’ demands for fulfilling material rewards and benefits, such as salary, promotion, and status, but also provides employees with the spiritual experience of self-control in some aspects of their work under worsening employment conditions. Drawing upon employees’ subjectivities, career identification thus corresponds more to what underpins the hegemonic discourse, providing analytical and empirical insights into the explanation of employees’ willingness to invest in overtime work and their alignment with managerial goals. Therefore, career identification, together with normative control, provides answers for understanding why internet firm employees consent to the extraction of additional work effort under neoliberalism and coercive management controls.

Conclusion and discussion

This study seeks to examine the mechanisms through which excessive working hours, typically the 996 working pattern with fixed extra hours in Chinese internet firms, are generated and persist. By drawing upon labour process analysis, we found that Burawoy’s theorisation of hegemonic despotism provides a response to this question. In this study, the most prominent and coercive mechanism behind the 996 working pattern is that of informal-flexible-allied despotism, which combines to generate the cumulative effect of high-risk of job loss and permanent unemployment, leaving powerless internet firm employees with no alternative but to comply with management’s 996 working pattern demands. Alongside the despotic apparatus, hegemonic mechanisms that rely on normative control and career identification can account for employee compliance and willingness to keep striving. This hegemony is reinforced via ideological means within conditions of neoliberalism and despotism as employees construct career identification to secure a dignified sense of work that provides them with a rational reason to engage in self-exploitation under corporate values. Based on the analysis, we suggested that hegemony and despotism relying on consent versus coercion work together to exert a powerful effect. As the two sources that frame hegemonic despotism, hegemony and despotism partially explain our research question. Although firms emphasise coercive practices, the normative control and emerging subjectivities constitute a reinforced hegemony that also strengthens despotism by eliciting employees’ consent. By synthesising these two mechanisms, this study highlights a rearticulation of hegemonic despotism to account for the 996 working pattern in internet firms.

This comparative multicase study summarises an integrated theoretical framework to uncover the underlying mechanisms of the 996 working pattern, but the case firms exhibit variations in each mechanism. Table 4 provides a comparative evaluation of the degree of each mechanism across cases and their contribution to the intensity of overtime work. As the table shows, case firms with typical 996 working pattens are those with medium to high-level despotic mechanisms and at least medium-level hegemonic mechanisms.

Table 4. An overview of variations across cases

	Despotic mechanism				Hegemonic mechanism		996 working pattern
	Informal coercion	Flexible coercion	Allied despotism		Normative control	Career identification
Firm A	High	High	High		High	High	High
Firm B	High	Medium	High		High	High	High
Firm C	Medium	High	Low		Medium	Medium	Medium
Firm D	High	High	Low		Medium	High	High
Firm E	High	Medium	High		High	High	High
Firm F	Low	Medium	Low		Low	Medium	Low

Source: authors.

Our findings add to research on long working hours by focusing on the 996 working pattern in internet firms, which continues to receive little attention. We answered the research questions of what exactly the 996 working pattern is, what differentiates it from other overtime types, what underlying mechanisms foster it, and how to understand its nature. Based on our findings, we constructed a theoretical framework to account for the 996 working pattern in internet firms. We suggested that the 996 working pattern is essentially an outcome of coercive management control on powerless employees, in which management coercion operates by integrating top-down bureaucratic supervision and inside-outside occupational pressures to force employees to rebuild values and reshape self-meaning to comply with management. To our knowledge, this study is among the first to explain the 996 working pattern. It offers an integrative insight into understanding the problem of excessive working hours, which has become the most salient workplace concern in Chinese internet firms. However, given the fast-paced development of the emerging digital economy and the immature corporate management system in China, the 996 working pattern and its underlying mechanisms are characteristics idiosyncratic to Chinese internet firms, and the generalisation of these findings to firms in other industries or countries remains to be seen.

Our analysis holds theoretical value in that it provides an updated version of hegemonic despotism in internet firms. On the basis of Burawoy’s concept of hegemonic despotism, we first introduced informal-flexible-allied despotism to uncover the approaches that internet firms adopt to impose comprehensive and profound coercive controls on their employees. This point aligns with the recent research trend that emphasises coercion under neoliberalism (Griesbach et al. 2019; Vallas, Johnston, and Mommadova 2022), and our discussion considerably extends prior studies by suggesting a novel form of despotism imposed on high-tech professionals in emerging platform firms. As for the firms’ hegemonic mechanisms, we identified normative control and career identification, the latter of which plays a significant role in motivating internet firm employees to exhibit consent to management controls. We summarised it as employees’ reshaping of their subjectivities under despotism, which is consistent with the recent vein of hegemony research (Purcell and Brook 2022). Therefore, with the traditional theoretical framework of hegemonic despotism, this study explained the emerging 996 working pattern in internet firms. As the original concept of hegemonic despotism is described as limited in capturing workers’ consent under neoliberalism, this study helps to fill this theoretical gap by examining workers’ reconfigured subjectivities and their effects on hegemony.

Our analysis sets some directions for future research by providing an analytical framework for the examination of overtime regimes and management-labour relations in internet firms. Existing research has already largely examined the adverse effects of working long hours, so future research should explore what might change at work to address the unsustainable toll of the 996 working pattern for reducing damage to employees’ well-being and creating a healthier work environment. Within a context marked by the rise of the digital economy, increasing competition between firms may continue to intensify exploitation and deterioration of working conditions, so research questions especially important for scholars and policy makers to consider are: Would work intensity and work length be reduced if management used lower-level coercive practices? Would workers’ collective representatives be effective in voicing and affecting workplace decision-making? What are the effects of labour regulations on reducing long work hours? More broadly, can industry competition be moderated to alleviate the stress that employees are exposed to? That said, we should take multiple actors, including workers and workers’ collective defenders, firms, and the state, into consideration for making any beneficial change to the 996 working pattern.

Acknowledgements

We thank the editors and two anonymous reviewers for their extremely thorough and constructive guidance throughout the review process. We fully acknowledge funding support from the Social Science Foundation of Guangdong Province of China (GD21YSH01), and the 70^th China Postdoctoral Science Foundation (2021M701252).

Manuscript received on 12 July 2022. Accepted on 31 March 2023.

References

ANTEBY, Michel. 2008. “Identity Incentives as an Engaging Form of Control: Revisiting Leniencies in an Aeronautic Plant.” Organization Science 19(2): 202-20.

ASHFORTH, Blake E., Spencer H. HARRISON, and Kevin G. CORLEY. 2008. “Identification in Organizations: An Examination of Four Fundamental Questions.” Journal of Management 34(3): 325-74.

AVGOUSTAKI, Argyro, and Almudena CANIBANO. 2020. “Motivational Drivers of Extensive Work Effort: Are Long Hours Always Detrimental to Well-being?” Industrial Relations 59(3): 355-98.

AYYAGARI, Ramakrishna, Varun GROVER, and Russell PURVIS. 2011. “Technostress: Technological Antecedents and Implications.” MIS Quarterly 35(4): 831-58.

BLAIR-LOY, Mary (ed.). 2003. Competing Devotions: Career and Family among Women Executives. Cambridge: Harvard University Press.

BRYAN, Mark L. 2007. “Free to Choose? Differences in the Hours Determination of Constrained and Unconstrained Workers.” Oxford Economic Papers 59(2): 226-52.

BURAWOY, Michael. 1983. “Between the Labour Process and the State: The Changing Face of Factory Regimes under Advanced Capitalism.” American Sociological Review 48(5): 587-605.

——— (ed.). 1985. The Politics of Production: Factory Regimes Under Capitalism and Socialism. London: Verso Books.

BURCHELL, Brendan. 2011. “A Temporal Comparison of the Effects of Unemployment and Job Insecurity on Wellbeing.” Sociological Research Online 16(9): 1-20.

BURKE, Ronald. J., Parbudyal SINGH, and Lisa FIKSENBAM. 2010. “Work Intensity: Potential Antecedents and Consequences.” Personnel Review 39(3): 347-60.

CAMPBELL, John L., Charles QUINCY, Jordan OSSERMAN, and Ove K. PEDERSEN. 2013. “Coding In-depth Semi-structured Interviews: Problems of Unitization and Intercoder Reliability and Agreement.” Sociological Methods & Research 42(3): 294-320.

CHUN, Jennifer Jihye. 2001. “Flexible Despotism: The Intensification of Insecurity and Uncertainty in the Lives of Silicon Valley’s High-tech Assembly Workers.” In Rick BALDOZ, Charles KOEBER, and Philip KRAFT (eds.), The Critical Study of Work: Labour, Technology, and Global Production. Philadelphia: Temple University Press. 127-54.

DÖRFLINGER, Nadja, Valeria PULIGNANO, and Steven P. VALLAS. 2021. “Production Regimes and Class Compromise Among European Warehouse Workers.” Work and Occupations 48(2): 111-45.

DRAGO, Robert William, David BLACK, and Mark WOODEN. 2005. “The Existence and Persistence of Long Hours.” IZA Discussion Paper No. 1720. Bonn: Institute of Labor Economics.

EISENHARDT, Kathleen M. 1989. “Building Theories from Case Study Research.” Academy of Management Review 14(4): 532-50.

FAHY, Paul. 2001. “Addressing Some Common Problems in Transcript Analysis.” The International Review of Research in Open and Distance Learning 1(2). https://doi:10.19173/irrodl.v1i2.321

GALLIE, Duncan (ed.). 2009. Employment Regimes and the Quality of Work. Oxford: Oxford University Press.

GASCOIGNE, Charlotte, Emma PARRY, and David BUCHANAN. 2015. “Extreme Work, Gendered Work? How Extreme Jobs and the Discourse of ‘Personal Choice’ Perpetuate Gender Inequality.” Organization 22(4): 457-75.

GICHEVA, Dora. 2013. “Working Long Hours and Early Career Outcomes in the High-end Labour Market.” Journal of Labour Economics 31(4): 785-824.

GOLDEN, Lonnie. 2009. “A Brief History of Long Work Time and the Contemporary Sources of Overwork.” Journal of Business Ethics 84: 217-27.

GOLDEN, Lonnie, and Barbara WIENS-TUERS. 2005. “Mandatory Overtime Work in the United States: Who, Where, and What?” Labour Studies Journal 30(1): 1-26.

GREEN, Francis. 2004. “Why Has Work Effort Become More Intense?” Industrial Relations 43: 709-41.

GREEN, Francis, Alan FELSTEAD, Duncan GALLIE, and Golo HENSEKE. 2022. “Working Still Harder.” ILR Review 75(2): 458-87.

GRIESBACH, Kathleen, Adam REICH, Luke ELLIOTT-NEGRI, and Ruth MILKMAN. 2019. “Algorithmic Control in Platform Food Delivery Work.” Socius 5. https://doi.org/10.1177/2378023119870041

HOU, Hui 侯慧, and HE Xuesong 何雪松. 2020. “‘不加班不成活’: 互聯網知識勞工的勞動體制” (“Bu jiaban bu chenghuo”: Hulianwang zhishi laogong de laodong tizhi, “Working overtime is life”: Labour ecology of internet knowledge workers). Tansuo yu zhengming (探索與爭鳴) 5: 115-23.

HUGHES, Richard L., Robert. C. GINNETT, and Gordon. J. CURPHY (eds.). 1993. Leadership: Enhancing the Lessons of Experience. New York: McGraw Hill.

JENKINS, Sarah, and Rick DELBRIDGE. 2013. “Context Matters: Examining ‘Soft’ and ‘Hard’ Approaches to Employee Engagement in Two Workplaces.” The International Journal of Human Resource Management 24(14): 2670-91.

KUNDA, Gideon (ed.). 1992. Engineering Culture. Boston: MIT Press.

LIANG, Meng 梁萌. 2019a. “996加班工作制: 互聯網公司管理控制變遷研究” (996 jiaban gongzuo zhi: Hulianwang gongsi guanli kongzhi bianqian yanjiu, 996: Research on the change of management control of internet companies). Kexue yu shehui (科學與社會) 9(3): 67-86.

———. 2019b. “彈性工時制何以失效? 互聯網企業工作壓力機制的理論與實踐研究” (Tanxing gongshi zhi heyi shixiao? Hulianwang qiye gongzuo yali jizhi de lilun yu shijian yanjiu, Why does the flexible working hour system fail? Research on the theory and practice of work stress mechanism in internet companies.) Shehuixue pinglun (社會學評論) 7(3): 35-49.

MAZMANIAN, Melissa, Wanda J. ORLIKOWSKI, and JoAnne YATES. 2013. “The Autonomy Paradox: The Implications of Mobile Email Devices for Knowledge Professionals.” Organization Science 27(5): 1337-57.

McGOVERN, Patrick, Stephen HILL, Colin MILLS, and Michael WHITE (eds.). 2007. Market, Class, and Employment. Oxford: Oxford University Press.

PURCELL, Christina, and Paul BROOK. 2022. “At Least I’m My Own Boss! Explaining Consent, Coercion and Resistance in Platform Work.” Work, Employment and Society 36(3): 391-406.

RAHMAN, K. Sabeel, and Kathleen THELEN. 2019. “The Rise of the Platform Business Model and the Transformation of Twenty-first Century Capitalism.” Politics & Society 47(2): 177-204.

RAPHAEL, Snir, Evan J. DOUGLAS, and Robyn MORRIS. 2006. “Workaholic, or just Hard Worker?” Career Development International 11(5): 394-417.

RYAN, Richard M., and Edward L. DECI. 2000. “Self-determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-being.” American Psychologist 55(1): 68-78.

SCHOR, Juliet B. 1999. The Overspent American: Why We Want What We Don’t Need. New York: Harper Perennial.

TRAN, Sang Kim. 2017. “Google: A Reflection of Culture, Leader, and Management.” International Journal of Corporate Social Responsibility 2(10): 1-14.

VALLAS, Steven P., Hannah JOHNSTON, and Yana MOMMADOVA. 2022. “Prime Suspect: Mechanisms of Labour Control at Amazon’s Warehouses.” Work and Occupations 49(4): 421-56.

van ECHTELT, Patricia E., Arie C. GLEBBEEK, and Siegwart M. LINDENBERG. 2006. “The New Lumpiness of Work: Explaining the Mismatch Between Actual and Preferred Working Hours.” Work, Employment and Society 20(3): 493-512.

WANG, Jenny Jing. 2020. “How Managers Use Culture and Controls to Impose a ‘996’ Work Regime in China that Constitutes Modern Slavery.” Accounting & Finance 60: 4331-59.

WILLIAMS Joan C., Jennifer L. BERDAHL, and Joseph A. VANDELLO. 2016. “Beyond Work-life ‘Integration.’” Annual Review of Psychology 67(1): 515-39.

WOOD, Alex. 2021. “Workplace Regimes: A Sociological Defense and Elaboration.” Work in the Global Economy 1(1-2): 119-38.

WYNN, Alison T., and Aliya Hamid RAO. 2020. “Failures of Flexibility: How Perceived Control Motivates the Individualization of Work-life Conflict.” ILR Review 73(1): 61-90.

YAN, Xia 嚴霞. 2020. “以自我為企業: 過度市場化與研發員工的自我經營” (Yi ziwo wei qiye: Guodu shichanghua yu yanfa yuangong de ziwo jingying, Self as enterprise: Over-marketization and self-management of R&D engineers.” Shehuixue yanjiu (社會學研究) 6: 136-59.

YIN, Robert K. (ed.). 2009. Case Study Research: Design and Methods. Thousand Oaks: Sage Publications.

ZYSMAN, John, and Martin KENNEY. 2018. “The Next Phase in the Digital Revolution: Intelligent Tools, Platforms, Growth.” Communications of the ACM 61(2): 54-63.

[1] The 996 working schedule refers to a working schedule that starts at 9 a.m. and ends at 9 p.m., lasting six days a week.

[2] For example, on 29 December 2020, a 23-year-old employee from Pinduoduo died of a heart attack when she left work at 1:30 a.m.; on 7 February 2022, the sudden death of a 25-year-old man occurred in Bilibili; on 23 February 2022, a 28-year-old employee of Byte Dance died suddenly after work.

[3] Big-small workweek is one of the work schedules set for internet employees, requiring them to work seven days during the big week, six days the following week, then called small week. With big week and small week proceed alternatively, employees only have two days for rest every month.

[4] “‘大小周’取消之後: ‘打工人’是皆大歡喜還是變相降薪?” (“Daxiao zhou” quxiao zhi hou: “Dagongren” shi jieda huanxi haishi bianxiang jiang xin?, Cancelling the “big-small workweek”: Is it a happy news for “workers” or a disguised way of pay cut?), CNR.cn (央廣網), 17 September 2021, https://baijiahao.baidu.com/s?id=1711146793443411010&wfr=spider&for=pc (accessed on 20 September 2021).

[5] Inspired by the ranking report released by the Chinese Internet Association in 2021, we classified the case firms into four categories: leading, top-ranked, front-ranked and moderate. Respectively, the leading firms are typically the most influential internet firms in China; top-ranked firms are inferior to leading firms but still among the top ten in terms of the comprehensive strength; firms rank between the tenth and the 50^th are defined as front-ranked; moderate-level firms are those of medium-sized and medium-performed firms.

[6] Modeled on the term 996, “007” indicates the possibility of working from midnight to midnight, seven days a week. If the work is not effective for all these hours, it implies that the employees can be asked to work at any time.

[7] 11 November, also called “double 11” or “Singles’ Day” was initially a day set for creating opportunities for single people to gather and find their other half, it then was manipulated by e-commerce companies and evolved into an online shopping day.

[8] HR Think-tank Alliance in Binjiang District (Zhejiang) was established on 9 November 2020. The first members were chief human resource officers or human resource directors from more than 30 firms such as Alibaba, Hikvision, Netease, and Geely. The alliance aims to create a human resource information-sharing and capability-enhancing platform best featured with mission, cohesion, and vitality.

[9] Trust and Integrity Enterprise Alliance was founded in February 2017. It was initiated by Jingdong and promoted by a group of well-known enterprises such as Tencent, Baidu, Walmart China, Xiaomi, etc. According to the official description, constructing an honest and clean business environment by fighting corruption, crime, and fraud is the purpose of the alliance. The membership is now more than 173.

[10] N denotes one’s working years in the current company according to Chinese Labour Contract Law.

[11] Billion-class service means the market value of one product/service, or estimated value of one upcoming product/service has reached at least one billion RMB. It signals the importance and potential influence of the project.