In March 2021, a customer and a delivery agent from a popular food delivery company grabbed the attention of social media in India. The customer accused the delivery agent of assaulting her, and the driver was subsequently arrested and suspended (Express Web Desk 2021). However, a few days later, it emerged that the customer might have assaulted the agent first and she, in turn, had a complaint issued against her (ibid.). The founder of the company stepped in and said that the company wanted to get to the bottom of the case and offered both parties support. In describing the delivery agent, the founder said that the agent had made 5,000 deliveries so far and had a 4.75/5 rating on the platform after 26 months (ibid.). The founder went on to say these are “facts, not opinions, or inferences” (ibid.).
This incident was another example of how the use of ratings in platform work becomes a tool of evaluation, negotiation and marginalization. Ratings have a significant impact on the lives of workers, their capacities to build relationships with their customers and the company, and the terms of their employment. Yet these ratings are still regularly perceived as neutral, even while they have the potential, through their design, to institutionalize data violence on workers.
Indeed, one of the ways in which work and relationships are defined in the platform economy is through the use of ratings. These ratings (Gupta and Natarajan 2020) permit customers to evaluate delivery agents for their work and, therefore, are the basis upon which delivery agents get future business. Ratings also become a medium for monitoring and surveillance by companies.
As in the example above, these ratings are wrongly interpreted as facts, suggesting a sense of neutrality (Porter 2020). They are assumed to be uncontested and a means of getting a good idea of the performance of delivery agents and their capacities, trustworthiness and efficiency. These ratings also assume a social world (Espeland and Sauder 2007), as they become a key consideration for the relationship between company and delivery agent, and delivery agent and customer, as well as an important source of communication between the parties.
There is a distance (Merry 2016) that numbers provide from the complexity of opinion, circumstance and context. As a result, cases where the delivery agent, for example, needed to deal with an unpleasant customer are reduced in this rating to a mark. The rating system also does not fully capture the agency of the delivery agent. For instance, if delivery agents wish to challenge or counter the ratings they have received, do they have a space to do so? Do the ratings capture the precarious conditions under which an agent has to complete a task? Does being rated incentivize fewer breaks and more trips, and how far agents are compelled to be at the behest of the customer to ensure their ratings do not go down? Ratings have become a self-referential system that companies, agents and customers continue to adhere to because of the narrative power to influence different behaviours and functions (Espeland 2015).
Very often, having one bad review attracts more attention from the company than several good reviews on the platform.
Ratings also affect the material lives of delivery agents on multiple levels. These ratings are the basis upon which agents can continue to work on the platform, garnering them greater attention and therefore more work, or in the case of poor ratings, also the risk of being deplatformed. Ratings are also the basis upon which, in negotiations with companies, their profiles are scrutinized. Very often, having one bad review attracts more attention from the company than several good reviews on the platform. Delivery agents are also nudged into making sure they get the highest rating. A recent study showed how a company in India used ratings to encourage competition between delivery agents by sharing information and data wherein delivery agents could find out how they fared compared to others within their location (Barik 2021).
Ratings are also the basis of negotiation. Agents use them to engage and interact with the company and the customer. But because companies do not often consider the ratings that the agents provide on the customer, agents are locked in an unequal relationship with the customer, with nearly negligible scope for complaint. Shaikh Saluddin, president of the Telangana Gig and Platform Workers Union in India, has argued that customers have the power to determine the lives of platform workers because the ratings have punitive impacts on the agents (Naraharisetty 2021).
In several platform companies, drivers, or delivery agents, are deplatformed or deactivated from the system if their ratings are low. There is no scope for discussion and often no termination notice. In these instances, the legal contract (de Souza 2020) is lopsided in favour of the company (Fairwork 2020) marginalizing workers, who, once removed from the platform, find it hard to work independently due to price undercutting by platform companies.
The precarity of workers who are dependent on these ratings indicates the technocratic and managerial ways in which work is evaluated. There is an emphasis on efficiency and scale as well as a presumption that numbers will tell the whole story. In another case from India, a company promised to deliver products in less than 10 minutes. When this marketing ploy was critiqued on social media for its lack of consideration of workers and their well-being, the founder justified the move by saying that it was disappointing that people were being cynical instead of celebrating the innovation of more efficient services (Mint 2021).
This narrative of innovation, speed and efficiency is used to continuously reinvent how platform companies offer customer satisfaction. However, it also creates systems where the ratings affect the capacity of agents to work and lead a dignified and flourishing life. Ratings trigger reactivity (Espeland and Sauder 2007) and impact not just the work that people are able to do, but also their capacity to be able to organize, negotiate and resist. Ratings also have the power of sanction, punitive action and erasure. They inflict data violence (Hoffmann 2018), where the collection of data not only fosters bias and discrimination, but also restricts access to services. Delivery agents are excluded from spaces, institutions and communities where they can live and work. As a result, ratings are not just devices to monitor efficiency but are also tools of oppression when used to monitor and control the lives of workers.