Algorithm: riders' decisions and routes

Food delivery applications uses algorithms to determine routes for delivery riders. These optimization problems take into account a wide range of factors, and occurs on various stages of the delivery process.

It begins with the riders picking a location to start his work. Sometimes delivery platforms provide information on which regions are currently dense with orders, and which areas are expected to receive higher orders. The riders need to decide on whether to move to regions with denser demand, as a larger amount of riders reallocating could result in decreasing delivery fees when they arrive.

Next, the algorithm would plan the order and path to pick up orders from restaurants, as sometimes riders simultaneously process multiple orders. This takes into account the distance to the restaurants, the speed of food preparation, and sometimes the preferred routes of the riders.

Here, there is a great need for the accuracy in the algorithm. From the perspective of the restaurant, they need to know when exactly the food must be ready. From the perspective of the riders, they need to have the food by the estimated time, or else they risk not arriving on time and getting negative reviews from the consumers.

It can be seen that the need for accuracy imposed by the delivery systems changed the culture and power relations in rider-restaurant relations. Riders would often need to negotiate with restaurants to get their orders on time. In China, the largest economy for food delivery platforms, riders would need to “make friends” with restaurant owners to ensure that their orders would be cooked on time, hence delivered on time.

Thirdly is the rider’s route to delivering food to consumers. On one hand, the system can use external data from maps and real-time traffic to calculate for the fastest route, on the other hand, many systems are updating to keep up with preferences and habits of the riders. Here, familiarity of the riders with the region and the routes determines their speed of delivery.

By providing routes for riders, the delivery application
s actually reduce the amount of training that one needs to become a rider. With such a low bar of entry, sometimes students or other people looking for quick and part-time jobs would consider taking on a food-delivery job. This technology resulted in the riders often working for short periods, hence shaping the “digital culture” experience of riders.

Another issue is certain delivery apps provide predictions on arrival times based on the past performance of riders. In other words, if riders perform well, future orders will have a shorter estimated delivery time. Over time, this can increase the pressure put onto riders, as they are facing more and more stress to deliver the order on time.

However, other research shows that many arriving quickly is not the prioritized consideration. Psychological studies have shown that consumers actually emphasize more on predictability. In other words, as long as the time of arrival is predictable and accurate, a longer delivery time would just result in the consumer ordering slightly earlier.

Finally, riders’ decision to stop working is heavily influenced by the pay mechanism. Many platforms give riders goals to meet on the duration of work, number of orders or the number of good consumer feedback. Therefore, this mechanism of wage calculations could lead to riders being pushed to work for longer hours than expected.

It can be seen that in all stages of a rider’s work, the algorithm plays a significant role. Their entire work experience is shaped by the data and calculations behind food delivery platforms.