Algorithmic management refers to the use of automated systems and algorithms to organise, monitor, and control workers' performance through the processing of large volumes of data and predefined, often opaque, rules. Although these systems are not exclusive to digital platforms, they take on a paradigmatic character in this context. This paper examines the specificities of algorithmic management within freelance labour platforms, an area less extensively studied compared to delivery platforms.
Freelance platforms function as global intermediaries, connecting clients and workers across geographical boundaries. Their main innovation lies in providing the infrastructure for matching supply and demand, establishing rules that enhance reliability, and ensuring compliance: for workers, the assurance of timely and secure payment without cross-border disputes; for clients, mechanisms to inform hiring decisions. Two central elements of these platforms are thus the payment systems and the evaluation frameworks, which include two-way ratings: clients assess workers, and workers assess clients, providing valuable information for others. Compared to delivery or ride-hailing platforms, more information is generally accessible here, including ratings and average prices for different types of projects.
We explore three dimensions of algorithmic management in these platforms: (1) the criteria for job offers and ranking of applications; (2) the mechanisms for monitoring and measuring working time; and (3) the evaluation and rating systems. Our analysis focuses on the specific criteria and tools employed by platforms to manage work, as well as the perceptions of workers and clients regarding these practices and the aspects they consider most opaque/discretionary.
Our empirical research focuses on three platforms widely used in Latin America —Upwork, Workana, and Fiverr— and draws on three types of data: (a) platform documents such as terms of use and tutorials; (b) a 2024 survey of Argentine design and IT freelancers addressing their perceptions of algorithmic management; and (c) a corpus of 22 interviews exploring workers' and clients' views on platform governance, with particular attention to areas perceived as opaque or discretionary.