Research Background
The rise of the on-demand economy in China has led to a significant increase in the use of food delivery platforms, with over 300 million users and a market revenue of US $37 million. This growth has been accompanied by intense competition among platforms, leading to the bankruptcy of many smaller players and the emergence of a duopoly dominated by Eleme and Meituan. The labor conditions of the delivery workers, who are essential to the functioning of these platforms, have become a critical area of study.
The global emergence of the “sharing,” “gig,” or “on-demand” economy has attracted increasing scholarly interest in how intermediary platforms build, connect, and reconstruct social relations among consumers, laborers, and companies. In China, the food delivery market is characterized by a highly concentrated population and decreased labor costs, leading to the proliferation of O2O (online to offline) food delivery platforms. According to Phoenix Television, the number of delivery workers exceeds three million, with migrant workers constituting the majority.
Previous research on technology and labor has documented the increasingly individualized control of entrepreneurism and algorithms in the on-demand economy. For instance, Rosenblat and Stark (2016) and van Doorn (2017) have highlighted the role of algorithmic management in shaping the work practices and performances of digital workers. However, these studies often overlook the human labor involved in the platform economy, particularly the lived experiences of gig workers whose lives and livelihoods are directly affected by these changes.
In this context, Ping Sun’s study aims to explore the labor conditions of delivery workers on food delivery platforms and how they experience and make sense of the algorithms. The study argues that delivery workers are not merely passive entities subjected to a digital “panopticon” but are active participants in creating their own “organic algorithms” to manage and, in some cases, subvert the system. This approach provides a more nuanced understanding of the dynamics of algorithmic labor in the platform economy.
The study is particularly relevant in the context of the Chinese food delivery market, which has seen a monumental increase in response to the surge in on-demand economic activity. In July 2018, the number of users of online meal ordering services had reached 300 million, generating a revenue of US $37 million. However, the market is also marked by instability, with IT giants like Alibaba, Tencent, and Baidu competing for market share, leading to the bankruptcy of many smaller platforms. For example, Baidu Deliveries, once the third-largest market operator, was sold to Ele.me in August 2017 after failing to gain an adequate market share. This consolidation has effectively created a duopoly, with Eleme backed by Alibaba and Meituan Dianping, whose main investor is Tencent.
The application designs and services provided by the three platforms—Baidu Deliveries, Eleme, and Meituan—are quite similar, allowing users to order restaurant food, supermarket products, vegetables, fruits, desserts, and even cake and flower deliveries. To attract more customers, these platforms often subsidize them through various promotions. The labor conditions of the delivery workers, who are essential to the functioning of these platforms, have become a critical area of study. These workers are distinct from other digital workers due to their embeddedness in the digitalized platforms, where mobile adoption and application usage are prerequisites for their employment.
Key Findings
The study reveals several key findings regarding the labor conditions of delivery workers and their interactions with algorithms. These findings are organized into three main categories: temporality, affect and emotional labor, and gamification.
Temporality: The Pressure of Time
The first key finding is the central role of temporality in the work of delivery workers. The platform algorithms constantly calculate and revise delivery times, creating a sense of urgency and pressure for the workers. Xiao Ji, one of the interviewees, noted that the delivery time for each order had been reduced from 45 minutes to 29 minutes, highlighting the increasing emphasis on speed and efficiency.
45 minutes to 29 minutes—this reduction in delivery time reflects the platform’s efforts to optimize labor resources and leverage the extra value generated by the delivery workers. However, the algorithmic design of timing the delivery workers’ labor does not account for the emotional and physical toll it takes on them. Li Feng, another interviewee, expressed frustration with the way Meituan calculates delivery distance and time, noting that the system often underestimates the actual distance and ignores real-world factors like traffic lights and winding roads.
The cut-off time has led to a dramatic spike in the number of traffic accidents experienced by delivery workers. According to The Paper, there were 76 casualties of traffic accidents in the first half of 2017, with half of them working for Eleme and Meituan. Sina News reported that there are casualties of delivery workers every 2.5 days in Shanghai, and in Nanking, there are 18 traffic accidents involving delivery staff every day.
The process of platformization has enabled algorithmic systems to dispatch orders and manage deliveries, leading to a loss of control over work time for the delivery workers. Although corporations promise flexible schedules and good payment, the reality is that full-time workers are required to comply with the compulsory eight-hour day, and most choose to work longer hours to get more orders. This “platform adhesion” has made them venture laborers who have lost control of their work time, blurring the demarcation between work and leisure.
The study also found that the algorithmic systems used by the platforms do not consider the emotional and physical well-being of the workers. For example, during rush hours, the workers have to compete against the limited delivery time set by the platforms, leading to high levels of stress and anxiety. The application installed on the delivery workers’ mobile phones periodically collects and accumulates data on the delivery time for the purpose of predicting, managing, and rearranging the delivery times increasingly precisely. However, this data collection does not include the emotional and physical toll it takes on the workers. As a result, the workers often feel overwhelmed and exhausted, leading to a decline in their overall well-being.
Affect and Emotional Labor: The Customer is King
The second key finding is the importance of affect and emotional labor in the work of delivery workers. The food delivery platforms are strongly oriented towards customer satisfaction, requiring workers to perform various kinds of emotional labor. For example, they are not allowed to enter the customer’s room, receive tips, or ask for good comments. Instead, they must smile, knock at the door, and hand over the change with both hands.
This emotional labor is a form of social performance centered on customers, contributing to the generation of a customer-oriented culture. The delivery workers must regulate and perform their emotions to make customers satisfied, even when they face marginalization and exploitation. The algorithmic governance of the platforms enforces these rules, making it difficult for workers to express their true feelings and frustrations.
The study found that the emotional labor performed by delivery workers is not only a requirement of the job but also a source of stress and dissatisfaction. Many workers reported feeling undervalued and disrespected by customers, leading to a sense of powerlessness and frustration. The lack of recognition and appreciation for their hard work further exacerbates the precarious nature of their employment.
The emotional labor required by the platforms includes maintaining a positive and professional demeanor, even in the face of challenging and sometimes hostile customer interactions. For example, delivery workers are expected to remain calm and polite, even when faced with unreasonable demands or complaints from customers. This can be particularly challenging during peak hours, when the pressure to deliver orders quickly and efficiently is at its highest. The study found that the constant need to perform emotional labor can lead to burnout and a decline in the quality of service, as workers struggle to maintain their composure and professionalism in the face of ongoing stress and pressure.
Gamification: The Knight System
The third key finding is the use of gamification in the management of delivery workers. The platforms employ a ranking system, known as the “knight system,” to incentivize and control the workers. The system assigns different levels to the workers based on their performance, with higher levels receiving greater subsidies for each order. For example, a Divine Knight (神骑士) receives a subsidy of 1.5 RMB (approx. 0.23 USD) per order, while an Ordinary Knight (普通骑士) receives only 0.1 RMB (approx. 0.015 USD).
The knight system creates a competitive environment where workers are motivated to complete more orders and improve their performance. However, this gamification also leads to a sense of alienation and dehumanization, as the workers are reduced to mere cogs in the machine. The algorithmic design of the system does not consider the human aspects of the workers, such as their well-being and job satisfaction. Instead, it focuses solely on maximizing efficiency and profit.
The study found that while the knight system can be an effective tool for managing and motivating workers, it also has negative consequences. The constant pressure to achieve higher levels and earn more subsidies can lead to burnout and a decline in the quality of service. Moreover, the system can create a sense of unfairness and inequality among the workers, as those who are unable to meet the high standards set by the platform may feel left behind and undervalued.
The gamification of the work process through the knight system is designed to motivate workers by providing tangible rewards for high performance. However, the study found that this system can also lead to a sense of competition and rivalry among workers, as they strive to outperform each other and achieve higher rankings. This can create a stressful and adversarial work environment, where workers are pitted against each other rather than working collaboratively. Additionally, the system can lead to a sense of unfairness and inequality, as workers who are unable to meet the high standards set by the platform may feel left behind and undervalued. The study found that the constant pressure to achieve higher levels and earn more subsidies can lead to burnout and a decline in the quality of service, as workers struggle to meet the demanding expectations of the platform.
Limitations
While the study provides valuable insights into the labor conditions of delivery workers and their interactions with algorithms, it also has several limitations that need to be addressed. These limitations include the sample size, the focus on a specific geographic region, and the potential bias in the data collection methods.
Sample Size and Geographic Focus
The study is based on a relatively small sample size of 45 in-depth, semi-structured interviews with food delivery workers in Beijing. While this sample provides a detailed and nuanced understanding of the workers’ experiences, it may not be representative of the broader population of delivery workers in China. The focus on Beijing, which has the largest number of delivery workers, may also limit the generalizability of the findings to other regions with different labor markets and economic conditions.
To mitigate this limitation, future research could expand the sample size and include workers from other cities and regions in China. This would provide a more comprehensive and diverse picture of the labor conditions and algorithmic practices in the food delivery industry. Additionally, comparative studies could be conducted to examine the differences and similarities in the experiences of delivery workers across different platforms and regions.
Data Collection Methods
The study relies on ethnographic fieldwork, including interviews, participant observations, and online ethnography, to collect data on the work practices and experiences of delivery workers. While this method allows for a deep and contextual understanding of the workers’ lives, it may also introduce potential biases and limitations. For example, the presence of the researcher may influence the behavior and responses of the workers, leading to social desirability bias or self-censorship.
To address this limitation, the study could incorporate multiple data sources and methods, such as surveys, focus groups, and secondary data analysis, to triangulate the findings and enhance the validity and reliability of the results. Additionally, the use of anonymous and confidential data collection methods could help reduce the potential for bias and ensure that the workers feel comfortable sharing their true experiences and opinions.
Algorithmic Transparency and Accountability
The study highlights the opaque and black-boxed nature of the algorithmic systems used by the food delivery platforms, which makes it difficult for workers to understand and challenge the decisions made by the algorithms. This lack of transparency and accountability can lead to a sense of powerlessness and mistrust among the workers, as they are unable to see or influence the factors that affect their work and income.
To mitigate this limitation, the platforms could implement measures to increase the transparency and explainability of their algorithms. This could include providing clear and accessible information about the criteria and processes used to calculate delivery times, assign orders, and rank workers. Additionally, the platforms could establish mechanisms for workers to provide feedback and contest the decisions made by the algorithms, ensuring that their voices and concerns are heard and addressed.
Practical Implications
The findings of the study have several practical implications for supply-chain and AI practitioners, particularly in the areas of algorithmic design, worker well-being, and platform governance. These implications can guide the development and implementation of more ethical and sustainable practices in the food delivery industry.
Algorithmic Design for Worker Well-Being
One of the key practical implications is the need for algorithmic design that considers the well-being and human aspects of the workers. The current algorithmic systems used by the food delivery platforms prioritize speed and efficiency, often at the expense of the workers’ health, safety, and job satisfaction. To address this, the platforms could incorporate human-centered design principles and metrics into their algorithms, such as the inclusion of rest periods, realistic delivery times, and fair compensation.
For example, the platforms could develop algorithms that take into account the physical and emotional toll of the work, such as the impact of long hours, heavy loads, and stressful interactions with customers. They could also provide tools and resources for workers to manage their workload and well-being, such as mental health support, training programs, and community-building initiatives. By prioritizing the well-being of the workers, the platforms can create a more sustainable and equitable labor model that benefits both the workers and the business.
Worker Representation and Voice
Another practical implication is the need for greater representation and voice for the workers in the decision-making processes of the platforms. The current governance structures of the food delivery platforms are often top-down and hierarchical, with little input or participation from the workers. This can lead to a sense of alienation and disempowerment, as the workers are unable to influence the policies and practices that affect their work and lives.
To address this, the platforms could establish mechanisms for worker representation and participation, such as worker councils, unions, or co-operative models. These mechanisms could provide a formal channel for workers to voice their concerns, suggest improvements, and negotiate better working conditions. Additionally, the platforms could involve the workers in the design and testing of new algorithms and features, ensuring that their needs and perspectives are taken into account. By empowering the workers and giving them a stake in the platform, the companies can build a more inclusive and democratic governance model that fosters trust and collaboration.
Regulatory and Policy Interventions
Finally, the study highlights the need for regulatory and policy interventions to address the challenges and risks faced by the delivery workers in the platform economy. The current legal and regulatory frameworks in China and other countries often do not adequately protect the rights and interests of gig workers, leaving them vulnerable to exploitation and abuse. To address this, policymakers could develop and implement new regulations and policies that recognize and address the unique characteristics and needs of platform workers.
For example, policymakers could introduce minimum wage and working hour regulations for platform workers, ensuring that they receive fair and adequate compensation for their work. They could also establish safety and health standards for the delivery industry, such as mandatory training programs, protective equipment, and accident insurance. Additionally, policymakers could promote the development of alternative and cooperative models of platform governance, such as platform cooperatives, that give workers more control and ownership over the platforms they work for. By addressing the structural and systemic issues in the platform economy, policymakers can create a more just and equitable labor market that benefits all stakeholders.
