Behind the takeaway delivery service: couriers are stuck in algorithms
With the rapid development of the mobile Internet, there are more and more start-up companies providing instant takeaway delivery services, such as Uber Eats, Deliveroo, Meituan and Ele.me. Chinese people seem to be increasingly relying on this online mode of food order and receive. In 2019, the scale of the takeaway delivery industry reached 90 billion USD, which experienced an increase of 39.3% compared to 2018 and 347.7% compared to 2015. Apart from food and drink, everything can be delivered from one place to another, such as life necessities, flowers, nonprescription drugs, have made a great contribution to the transaction volume. There are 421 million Chinese users placing orders on takeaway delivery platforms and 7 million couriers shuttling through streets and roads. Behind such a huge industry, algorithms play a crucial role in resource allocation.
How the algorithm works behind the takeaway delivery platforms？
Takeaway delivery involves a quite complicated operation model, which includes both online and offline business. The orders made on takeaway platforms not only directly reflect the demands of each user but also strictly assess the supply chain’s capability. The traditional dispatch mode is called manual management. Supervisors allocated delivery tasks according to distance and couriers’ workload. However, this method is far away from achieving a balance between demand and supply. To save operation costs and bring good experiences to customers, the takeaway delivery platforms need to maximize the overall efficiency of orders distribution. One of the most critical points is to allow the couriers to deliver more orders per unit time and generate more value.
Various platforms have made a lot of efforts and achieved remarkable results in using algorithms and big data to dispatch orders. For example, in 2017, Chinese famous takeaway delivery service provider Meituan launched an intelligent system called “Super Brain”, which used trajectory big data and algorithms to estimate the delivery time for each order, assigned the suitable couriers, and designed the optimal route between stores and consumers. With the deepening of machine learning, the system has been continuously optimized and has introduced voice assistance, automatic order acceptance and other functions to help solve trivial things in the delivery process. The effect is that the average delivery time has been shortened from 41 minutes to 28 minutes, and the punctuality rate has improved to 98%. The per capita daily order volume has increased by 46% compared to before the launch. In 2020, Meituan’s average daily orders exceeded 40 million, with a market share of nearly 70%.
For couriers, the speeding of “life and death”.
The continuously shortened delivery time is commendable progress for the creators, those platforms with brilliant programmers, and it is the embodiment of the deep learning ability of AI intelligent algorithms. However, for couriers, this can be crazy and terrible. Along with the delivery tasks comes the delivery time calculated by the intelligent system based on algorithms. The delivery time accurate to the minute will not only appear in couriers’ dispatch equipment, but also on the customers’ mobile phone screens and it is the most important indicator to measure the quality of work. If they spend more time than set, it will lead to negative feedbacks and comments from customers’, deductions from wages, or even being fired.
“I have 20 minutes to arrive at the restaurant, wait for food ready, and deliver them to consumers 1 kilometre away.” A courier shared his real experience, and unfortunately, he is not the only one. To achieve these incredible goals, couriers have to violate traffic rules such as speeding, running the red light, reverse driving. The direct consequence is the sharp increase in the number of couriers who encountered traffic accidents. According to the Traffic Police Corps, on average, 1 courier was injured or killed every 2.5 days on the way in Shanghai. Ironically, the emotionless algorithms continue to analyse the returned data and again shorten the delivery time to the extent they think it is possible. Algorithms never know what happened on the way. In the face of huge amounts of data, couriers turn into a variable role which was just providing input and any attempt to protect themselves is so weak and pale under the data monitoring and efficiency-oriented algorithms.
What does not match the highly dangerous working environment is couriers’ income. The commission for each order is 0.75–0.9 USD. According to Professor Zheng Guanghuai from Central China Normal University, the average monthly income of a courier in Wuhan is 879.6 USD, which is below the average social income in Wuhan (1221.8 USD). According to Feng Xiangnan from Capital University of Economics and Business, the monthly income of couriers in Beijing is concentrated in the range of 747.7–1,196.3 USD, which is below the average social income in Beijing (1601.9 USD). According to Lv Xuanru from East China Normal University, the average monthly income of couriers in Shanghai is 937.8 USD, which is below the average social income in Shanghai (1171.3 USD).
Who is responsible for saving couriers?
Recently, the tragedy of couriers has been exposed to the public, under pressure from public opinions and requirements of social responsibility. Meituan and Ele.me, the two major takeaway delivery platforms in China, have successively introduced new policies to protect the rights and interests of couriers. Ele.me launched a new function called “5 minutes”. When consumers place orders, they can choose to give 5 more minutes. Meituan seems to be a bit more generous because it gave 8 more minutes. Do this really work? The facts are not as expected. Whatever 5 or 8 minutes, couriers still seize every second to finish more orders so that they can get more income. In my perspective, what they need to change is the payment model. Set a basic salary and increase the commission ration may slow down the pace of couriers. Besides, the algorithms still have room to improve, but the main goal may not always be efficiency-oriented but includes more social elements. Apart from programmers, the process of algorithms establishment should involve sociologists, merchants, consumers and government.
Couriers should also be responsible for saving themselves. They should be awakened in terms of protecting their rights in the workplace. There could be some social platforms where they can loudly speak out their own opinions.
There is still a long way to go to save couriers from algorithms.
