Introduction
While digital platforms like Uber, Didi, and Rappi maintain that their algorithms for order allocation are impartial, internal documents filed with the Mexican Federal Center for Labor Mediation and Registration (CFCRL) reveal that a rider or driver’s category can influence access to orders, incentives, and rewards.
Algorithmic Management Policies
Following a labor reform regulating digital work, these platforms were obligated to publish algorithmic management policies explaining the rules of engagement for riders and drivers. These documents detail how tasks are assigned, evaluation criteria, access to incentives, and penalties.
Order Allocation
Algorithms consider various factors for order allocation, with distance, vehicle type, and geographical location being the most common. Uber’s system also takes into account traffic conditions, payment methods, and historical data.
Rappi considers the number and state of products on its platform when displaying orders, which can affect task notifications. Didi acknowledges geolocation, accumulated wait time, average rating, and acceptance/cancellation rates as influential factors.
The document from Didi clarifies that repeated task rejection or failure to comply can negatively impact priority or limit future task assignments, as the algorithm aims to optimize task allocation by considering multiple factors.
Evaluation and Categorization
Application systems have evaluation mechanisms that influence a category assigned to riders and drivers. Some algorithms consider user ratings, while others incorporate criteria like task acceptance and completion rates.
Incentives and Rewards
Although categorization does not always affect task allocation, it can influence access to incentives and rewards. For instance, Uber links reward program access to working in high-demand zones, specific days, or completing a minimum number of trips. Uber Pro, a recognition and benefits program, considers cumulative criteria like completed tasks, average ratings, acceptance and cancellation rates, along with other specific factors outlined in the terms and conditions.
Didi’s reward access is subject to supply and demand, geographical characteristics of the area, and user factors. Although not the sole determinant, category can positively influence task or service allocation during promotional events or high demand.
Rappi explicitly states that based on a rider’s internal category, order delivery can be prioritized, and the same criterion applies to incentive allocation by the algorithm.
Penalties for Non-Compliance
The digital platform reform also introduced penalties for failing to adhere to algorithmic management policies. Fines range from 113,140 to 282,850 pesos per affected worker, making it the second-highest penalty under the Federal Labor Law.
This requirement is not exclusive to Mexican regulation; countries like Spain, Croatia, and Malta have incorporated similar obligations for apps to explain their order allocation and incentive access criteria in legislation.
Key Questions and Answers
- What are algorithmic management policies? These are documents explaining the rules of engagement for riders and drivers, including task allocation, evaluation criteria, access to incentives, and penalties.
- How do algorithms allocate orders? Factors like distance, vehicle type, geographical location, traffic conditions, payment methods, and historical data are considered.
- How do algorithms categorize riders and drivers? User ratings, task acceptance and completion rates, and other criteria are used for categorization.
- Can a rider’s category affect incentives and rewards? Yes, categorization can influence access to incentives and rewards, although not always affecting task allocation.
- What penalties exist for non-compliance with algorithmic management policies? Fines ranging from 113,140 to 282,850 pesos per affected worker have been established.
