Chapter 5
The bitter aftertaste of app-based food delivery

Purchasing food from restaurants and grocery stores and having it delivered to your doorstep is convenient and has become popular in the Nordic countries. Foodora, Wolt and Just Eat are the platform companies dominating the Nordic market. These companies use the digital infrastructure of a platform to coordinate supply and demand (Srnicek, 2017) and usually classify their couriers as independent contractors or use other non-standard forms of employment (Jesnes, 2024). These working arrangements deviate from the standard employment relationship prevalent in the Nordic countries (Rasmussen et al., 2019), raising concerns about the couriers’ working conditions and their health and safety. 

In this chapter, we investigate working environment challenges that may arise with digitalized work arrangements like app-based food delivery in the Nordic countries. We are particularly interested in how app-based food delivery companies use algorithmic management techniques to control couriers, and in what ways this affects couriers’ occupational safety and health. We build on three case studies of app-based food delivery companies in Denmark and Norway, Just Eat, Foodora and Wolt, that use different forms of employment. We have conducted interviews with couriers, platform companies and union representatives, as well as desk research and document reviews. 

First, we present our theoretical framework, wherein we define how we understand algorithmic management and which dimensions of working environment challenges we focus on in our empirical analysis. Then we present our methods and describe the empirical cases and, finally, share our analysis and conclusion. We argue that app-based food delivery work is insecure, physically demanding and mentally exhausting, but we find differences in how couriers cope with these challenges. Furthermore, working environment challenges seem to be connected to type of employment arrangement, with some business models putting couriers more at risk than others.

In this chapter, we explore working environment challenges for couriers with different employment statuses working in app-based food delivery. We have a special interest in companies’ use of algorithmic management and its impact on couriers’ occupational safety and health (OSH). 

First, we define algorithmic management as “a system of control where self-learning algorithms are given the responsibility for making and executing decisions affecting labour, thereby limiting human involvement and oversight of the labour process” (Duggan et al. 2020: 119). The platform companies we are looking at are known for changing the way the algorithmic management system is set up or other elements of how they organize work, referred to as intraplatform algorithmic changes (Mendonça and Kougiannou, 2023), which may alter working conditions and pay from one day to another. We view algorithmic management as an extension of companies’ growing reliance on field technology (see Oppegaard and Bråten, Chapter 2) but with a shift towards automated decision-making in areas traditionally handled through human interaction. This implies that algorithmic management is not something entirely new but rather a continuation of a trend towards using field technology for recording and accessing information about couriers in remote work with the aim of controlling them.

Second, we define our understanding of working environment challenges. Here, we are inspired by Ropponen et al. (2019), who reviewed the existing literature to explore features of digitalized platform work and their influence on OSH. As a theoretical framework for their review, they used the job demands-resources (JD-R) model, which explains why stress, burnout and health-related problems occur (Demerouti et al., 2001). The JD-R model groups work characteristics into job demands – requiring sustained physical or mental effort such as noise, time pressure or a heavy workload – and job resources – aspects of the work that are functional for work goals or that stimulate personal growth, such as autonomy, participation in decision-making or task variation (Demerouti et al., 2001: 501). While being exposed to extreme job demands can lead to exhaustion and burnout in the long run, job resources can have the opposite effect, creating a motivational process that leads to work engagement and well-being at work (Demerouti et al., 2001).

Job demands and job resources vary across different occupations and contexts. Ropponen et al. (2019) applied this framework to platform work, exploring online platform work and on-demand platform work where jobs or tasks are assigned online but carried out physically (Ropponen et al., 2019). Based on a review of the existing literature they identified different job demands (e.g., job insecurity, a strenuous physical work environment with time pressure, harassment, isolation, competition etc.) and job resources (e.g., task variety and opportunities to develop competences) and discussed how they are connected to the health, well-being and safety of platform workers. They observed extreme variations in platform work, ranging from situations with high demands and many resources to situations lacking both. They concluded that there is a need to balance job demands and job resources for platform workers to stay healthy, well and engaged (Ropponen et al., 2019).

We structure our empirical analysis around several characteristics of platform work. Some of these were identified by Ropponen et al. (2019), but we also add some that we find pertinent based on our empirical analysis. The categories we use from Ropponen et al. (2019) are as follows: job insecurity, time pressure, physical work environment, harassment, isolation and competition. We add income insecurity (which we understand in connection to job insecurity) and waiting time, which we found important when analysing our data. 

The analysis of the cases from both Denmark and Norway is based on data collected through interviews with bike and moped couriers (not couriers with cars), management from the platform companies and union representatives and dialogue with representatives of the labour inspectorate, as well as a desk review of relevant documents (see Table 5.1). The analysis is primarily based on the courier interviews, and we use the other interviews and the document review to support it.

In 2022 and 2023, we conducted ten interviews in Denmark and 12 in Norway. Representatives from trade unions and management were recruited via e-mail, and we used various methods to recruit couriers for the interviews. In both countries, couriers were approached outside popular restaurants (so-called hot spots) in Aalborg (the fourth largest city in Denmark)

Aalborg was chosen because the Danish research team is based there. Several of the interviewees, however, had work experience from the capital, Copenhagen, or other cities and talked about these experiences in the interviews. The main implication of conducting interviews in Aalborg instead of the capital is that there might be less competition and less time pressure because there are fewer customers ordering takeaway than in larger cities.

and Oslo. We provided prospective interviewees with information sheets stating our research aims and asked if they would like to participate in the project. This process was time-consuming due to challenges with reaching the couriers: some did not speak the local language or English, others were in a rush, and some showed no interest. In Denmark, some of the couriers had time to do the interview when we met them, but most interviews were scheduled for later at a place and time convenient for the couriers. The couriers who also served as union representatives were recruited through e-mail, and we also used the snowball method to reach more couriers. In accordance with courier preferences, interviews were held face-to-face, over phone or video calls or via e-mail. Follow-up discussions were conducted via e-mail or in-person meetings as needed.

In the courier interviews we used the same semi-structured interview guide in the two countries covering topics such as the couriers’ background and previous labour market experiences, their motivations for being a courier, their working conditions and pay, their experience with and perceptions of algorithmic management, health and safety issues and representation/​union involvement. In the interviews with management, we used the same semi-structured interview guide with questions about the company and its business model and about their perspectives on working conditions, pay and OSH issues for the couriers. The use of semi-structured interview guides also made it possible to explore specific themes of particular importance to the interviewees. 

Seventeen of the 22 interviews were with couriers (eight in Denmark and nine in Norway), some of whom had experience with one platform company, others with several companies. In some of the interviews we therefore learned about more than one platform company. In Denmark, all the couriers interviewed were male. In Norway both female and male couriers were interviewed, although most were men, which also resembles the general picture of workers in app-based food delivery. We interviewed couriers of different ages, ranging from 20 to 50 years old, and with different national backgrounds. In both countries, a little less than half of the interviewees had native backgrounds, and the other half came from Southern Europe, Eastern Europe, Asia or Africa, and our sample seems to be relatively representative of the demographic composition of the general courier workforce in the two markets. To ensure the anonymity of the couriers, we have, in the analysis, chosen a strategy where we refer to what has been said in the interviews rather than using quotes; when we use quotes, we do not indicate which interview they are from. 

In addition to the interviews, we also collected and read documents and data from the companies and the couriers to support our analysis. To gain a detailed understanding of the algorithmic management systems of the companies, we explored Wolt’s registration and training programme. Furthermore, informants in both countries provided valuable visual materials, including app screenshots. We also examined company documents such as the Wolt courier protection policy, Wolt’s transparency reports and one of the courier contracts, as well as the collective agreements for Foodora and Just Eat (see Table 5.1).

Table 5.1 Data sources

This empirical material was subsequently carefully read through several times and coded according to the work characteristics identified in Ropponen et al. (2019). During this process new categories also emerged (for instance waiting time), and some of Ropponen et al.’s categories were nuanced (for instance, we learned that job insecurity is connected to income insecurity). 

In the data analysis process, we observed that working conditions and safety and health concerns were largely consistent between the two countries, although they varied by company (Wolt, Foodora, Just Eat) based on form of employment (see Table 5.2) and algorithmic management practices. Consequently, we chose not to conduct a comparative analysis between the countries but rather to treat each company as a distinct case. 

Our three cases are Wolt, Foodora and Just Eat. Wolt operates in both Norway and Denmark, and we have empirical data from both countries. Foodora also operates in both countries, but most of our empirical data is from Norway because at the time of the data collection Foodora had just entered the Danish market and was in the beginning of organizing their operations. However, we have included information from Foodora Denmark’s website about its business concept and we have some information from couriers who have experiences with Foodora.

In May 2024, after operating in Denmark in 18 months, Foodora announced that they are closing their operations in Denmark due to “challenging macroeconomic developments” (Eriksen 2024).

Just Eat operates only in Denmark. While Foodora and Wolt have expanded their services to include both restaurant and grocery deliveries from their own storage facilities during the pandemic, Just Eat only offers restaurant deliveries.

In this section, we provide a brief overview of each case, the form of employment used, working conditions and the algorithmic management of the couriers.

Table 5.2 App-based food delivery platforms selected as analytical cases

Foodora, a subsidiary of Germany-based Delivery Hero, entered the Norwegian market in 2015 and expanded to Denmark in October 2022 (by acquiring the existing Danish company Hungry.dk, which had operated in Denmark since 2013). In Norway, the company initially employed part-time couriers and signed a collective agreement with Fellesforbundet in 2019 after a significant mobilization amongst the couriers. The agreement was renegotiated in 2020 and 2022. Since 2019, Foodora Norway has also relied on independent contractors and freelancers – all car-based couriers are freelancers/self-employed. The freelancers are facilitated through third-party umbrella companies known as EasyFreelance and Manymore (Jesnes and Oppegaard, 2023). Foodora Denmark uses freelancers.

Foodora couriers are required to sign up for shifts that typically last between two and a half and eight hours in Norway and two and five hours in Denmark. They must be physically present at a specific location five minutes before their shift starts. Employed couriers are assigned orders while freelance couriers generally have to actively accept delivery requests. Orders are offered or assigned to couriers based on location and order size. Couriers, who can decline orders, must accept them within 60 seconds. If not, the order will be offered to another courier. During deliveries, a dispatch centre follows the courier’s route from the restaurant to the customer on a digital map. If couriers need help during deliveries, they can contact the dispatch centre through the app. 

The employed couriers in Norway are paid per hour and earn an additional commission per delivery, figures which are set in the collective agreement, while freelancers/self-employed are paid only per delivery – a price they cannot negotiate. According to the Foodora Denmark website, prices for each delivery are calculated based on factors such as distance, number of orders and time of the day and week. 

During the pandemic, Foodora Norway also introduced a productivity measure system, which is an example of how the company uses algorithmic systems to manage its couriers. The system was introduced for all couriers, employees and freelancers/self-employed (Jesnes, 2024). Through this system, which also exists at Foodora Denmark, Foodora assesses various aspects of couriers’ performance, including adherence to scheduled shifts, app logins, order acceptance rates, active app usage during shifts, acceptance of orders during high-demand hours and more. Based on their scores, the couriers are compared with each other and ranked into batches (one to ten), with priority shift selection given to those in batch number one. Couriers in lower batches must choose less attractive shifts with lower demand and therefore earn a lower income in the following weeks. Couriers receive requests for deliveries, which they can decline, but declining orders may result in being assigned to a lower batch.

The second case, Just Eat, was founded by Danish entrepreneurs in 2000. The company remained Danish-owned until 2020, when it merged with the Dutch company TakeAway.com. Just Eat entered the Norwegian market in 2021 but exited in 2022 after failing to gain a significant market share. The couriers working for Just Eat used to be employed on zero-hour contracts and had no guarantee of either a minimum or maximum number of weekly hours. In 2018, Just Eat became part of the employer organization Dansk Erhverv and entered a collective agree­ment for its office clerks with the union HK. In 2019, Just Eat was approached by the union 3F which wanted to negotiate a collective agreement for couriers. With Dansk Erhverv’s help, the first collective agreement for food delivery couriers in Denmark was signed in 2021 (Madudbringningsoverenskomsten, 2021–2023). It was renegotiated in 2023 and will run until 2025. All 850 Just Eat couriers are therefore now employed under this collective agreement (Ilsøe and Madsen 2022: 70). 

The collective agreement allows for part-time and full-time employment, but the former is most common. Couriers sign up for specific shifts with a minimum duration of four hours and they get paid per hour. When they start their shift, they need to log on to the app, but this is only possible when they are close to the city centre. They get a wage supplement when they work during unsocial hours. Just Eat uses different bonus systems and competitions to incentivize couriers to work faster. According to the collective agreement, they must receive their shifts four weeks ahead. It is possible to get extra shifts of at least two hours, but these are voluntary (Madudbringningsoverenskomsten, 2023–2025). The couriers are assigned orders and do not have the opportunity to decline them.

Couriers type in their availability for shifts, and the shifts are assigned automatically. If they are sick, they can switch shifts through different channels. The couriers know little about how orders are assigned but suspect that proximity to pick-up location and type of vehicle used play a role. Upon assignment, couriers receive a pick-up location and time. After pick-up, they register in the app and receive detailed information about the delivery. The app uses GPS for route calculation monitored by the live-support team.

Wolt is a Finnish company present in both Norway (from 2017) and Denmark (from 2018). In 2022, Wolt was acquired by the American company DoorDash. In contrast to Foodora and Just Eat, Wolt relies exclusively on independent contractors and freelancers in both Norway and Denmark. In Wolt’s terminology, couriers are called “partners”. They choose when and where they log on and off the app and are paid only per delivery. When receiving a delivery request, Wolt couriers have a certain amount of time to accept or reject the order. Our interviewees reported having 60 seconds, although according to the Wolt transparency report from 2023, 30 seconds is set as the default; however, it is possible for the couriers to change this (Wolt 2023: 17). The requests contain detailed information about pickup location, delivery destination, delivery distance and the proposed fee. It is possible to “bundle” orders, which means that the courier can pick up and deliver more than one order at a time and thus earn more. While delivering, the courier is monitored by a dispatch centre that can follow their route. The dispatch centre can also be contacted if the courier needs assistance during deliveries.

Since 2022 Wolt has published so-called algorithmic transparency reports describing how their algorithmic management system works (Wolt, 2022, 2023). The company uses, among other things, an algorithm to connect couriers with customers to “offer the most suitable courier partner a delivery task between the merchant and the customer” (Wolt, 2023: 16). In this calculation, Wolt uses information on the location of couriers, their availability (online or not) and type of vehicle to estimate how large an order can be placed and how fast it can be delivered (Wolt, 2023: 16).

Wolt also uses algorithms to calculate the price for each delivery. In the transparency report, Wolt explains that they “offer a delivery fee for each delivery task that is worthwhile accepting by pricing them based on the individual factors that could potentially impact the delivery” (Wolt, 2023: 19). In this calculation, Wolt uses estimated route distance (not a straight line), the courier’s distance to the pick-up location, customer and merchant location, weather conditions, type of order and other factors, for instance difficult of terrain (Wolt, 2023: 19). The fee that the courier gets includes a pick-up fee and a fee for the distance travelled from the pick-up location to the customer or customers (Wolt, 2023: 19). Wolt argues that no delivery is the same and therefore the fee will vary.

To sum up, we have selected three platform companies with different employment arrangements: Wolt, with the freelancer model; Just Eat, with the employee model where employees work under a collective agreement; and Foodora, with a combination of both forms of employment (in Norway). In our analysis, we will also highlight findings concerning working environment challenges that are related to different employment arrangements.

Our analysis concerns work environment challenges for couriers in app-based food delivery, with a special focus on how algorithmic management techniques affect couriers’ health and safety. As mentioned in the theory section, our empirical analysis is structured around the findings of Ropponen et al. (2019), who have already identified several health and safety challenges in digitalized work arrangements. However, we nuance and expand their framework based on findings from our empirical analysis. The following characteristics structure our empirical analysis:

Ropponen et al. (2019) identified job insecurity as an important aspect of job demand affecting the health and safety of platform workers. It is relatively easy to become a courier – download the courier app, provide identification, undergo an online training programme – but it is demanding to earn and maintain an income, which is why this type of work is considered insecure. Ropponen et al. (2019) only consider job insecurity, but we argue that income insecurity is also an issue, and that job insecurity and income insecurity are two sides of the same coin.

In our interviews, both job and income insecurity were central themes in the experiences of couriers. However, the extent of insecurity varied between the couriers and the platform companies due to distinct work arrangements (see Table 5.2 and case description above). Hence, while all platform workers experience job and income insecurity, it is more severe for some than for others, which we elaborate on below.

First, the employed couriers at Foodora in Norway and Just Eat in Denmark, who work under a collective agreement and get an hourly wage, have greater job security than freelance and self-employed couriers at Wolt and Foodora, who are compensated solely per delivery. However, the employed couriers often work part-time which can cause income insecurity if they are not able to secure enough weekly hours, and we did learn from the interviews that part-time employment is more common among these couriers as there are not enough available shifts to make a living as a full-time courier. Some couriers work other jobs in addition to their job as a food delivery courier. We have also come across couriers who work for several platforms at the same time to maximize their income.

Employed Foodora couriers in Norway, have a minimum guarantee of ten hours per week in their contracts, with the option of taking on more shifts when available. Part-time couriers fall under the protection of the Working Environment Act (Arbeidsmiljøloven), which includes a provision that allows workers to demand that the hours stated in their contract align with their actual working hours over a longer period of time. Many couriers have used this provision to obtain more than ten hours in their contracts and hence more job and income security (Jesnes et al., 2021).

In contrast, freelancers/independent contractors working for Foodora experience reduced job security since they do not have an employment relationship and they are also not protected by the Working Environment Act. However, in comparison to Wolt couriers, Foodora freelancers/independent contractors are still offered shifts, which gives a certain job security. These shifts are scheduled during peak demand hours, ensuring a guaranteed workload when couriers have a shift. During periods of reduced demand, such as the summer season, the number of available shifts may decrease, impacting the couriers’ ability to choose shifts, and hence decreasing job security. This model still provides a degree of predictability that is not present at Wolt (Jesnes and Oppegaard, 2023).

There is no job security at Wolt as couriers are only compensated on a per-delivery basis. A Danish Wolt courier explained what it was like to be a courier at Wolt as follows: “We decide ourselves how much we work. It’s just an app on the phone and if you turn it on, you’re at work, and if you turn it off, you’re not at work anymore”. A self-employed courier based in Norway explained job and income insecurity as follows: “Well, there is no holiday pay, for example. So, what you are earning is what you are getting and there is no extra. And if anything happens to you it is your own responsibility. That’s the risk you have in this”.

When we asked Wolt couriers about their perception of income insecurity, we heard two different perspectives. Some described working long hours with little income. Others also described working long hours and completing many deliveries – often between 50 and 60 – per day but expressed satisfaction with their earnings and did not experience income insecurity. Hence, not all freelancers/independent contractors are concerned with insecurity related to income. Nonetheless, and as we will explore in detail in the next section, we learned in the interviews that earning this much often requires working at a very fast pace, which can increase the risk of accidents and affect health and safety in a negative way.

Ropponen et al. (2019) identified time pressure as an important characteristic of platform work. For couriers, time pressure is connected to making deliveries fast. Through the interviews, we also identified time pressure as a characteristic of platform work, but we also found waiting time to be both a job demand and a resource for couriers. We start with waiting time and then turn to time pressure.

Most of the couriers we interviewed described waiting for orders. Waiting time can occur if the demand for takeaway is low or if too many couriers are online at the same time, which makes competition between couriers fierce. Waiting time can also occur if the restaurant is slow in making the food or if the customer is not present at the delivery location. The interviews revealed different perspectives on waiting time. Some of the couriers found waiting time challenging, stressful and demotivating, while others did not.

First, we did not find any employed couriers paid per hour that considered waiting time a job pressure since they get paid while waiting anyway. When asked about how they experience waiting time, some of the couriers said that they find it relaxing, including a Just Eat courier in Denmark, who described it in the following way: “Relaxing. If we don’t have orders, we can come and sit in here [the interview took place at a Burger King]. This table is actually for us”. One could argue that this waiting time could decrease motivation since the job is perceived as boring, but none of the interviewees at Just Eat mentioned being bored or losing motivation when they spend time waiting for orders. They all accepted that waiting time is a part of the job.

On the other hand, several other couriers from both countries who worked as freelancers experienced the waiting time as more stressful and they linked it to income insecurity: they were frustrated that they had to wait because then they did not earn money. It is not entirely clear in our data, but it seems that there is a connection between whether or not a person experiences waiting time as exhausting and employment arrangement, with our data showing that more freelancers than employees experience waiting time as a job demand even though this was not the case for all the freelancers we interviewed. This point about the relation between the experience of waiting time and work arrangement is elaborated in this citation from a Danish courier at Just Eat in which he imagines what it is like to work at Wolt. He also reflects on time pressure:

If it’s the same way [at Wolt] that there are days when you have nothing to do, then you don’t get any money from working for Wolt. I don’t think I would be able to. I’d rather have a flat salary and not have to stress. It’s not a job worth stressing yourself over …that is the thing about thinking that if I am really fast, I can run the next order. I would sit all the time thinking about being able to reach this and that. And as soon as you stand inside a restaurant and the food is not finished, you stand there and lose your earnings. And then you get angry. We can also see that from the other couriers. They are significantly more impatient in restaurants. I just sit down and wait. It doesn’t matter to me. I get the salary I get.

A freelance courier based in Norway working for both Wolt and Foodora had a similar interpretation. He also reflected on the relation between waiting time and motivation: “I prefer shifts [with Foodora], because when you have shifts […] your motivation is like ‘Ok, I work from three to seven.’ But with Wolt, if I don’t get the order fast enough, my motivation goes down and I don’t want to work”. 

However, we did also interview freelance couriers who were not worried about waiting time or who had strategies to cope with or avoid it. For instance, one Wolt courier in Denmark had made the decision only to work during peak times when the waiting time is shorter so that he simply avoided long waiting times. Another courier from Denmark used longer waiting times to make other shorter deliveries. This last courier is an example of something we also encountered in our interviews, namely that some freelance couriers seemed to be very keen on time optimization, to the point that they were constantly thinking about how to minimize waiting time and make deliveries very fast. They tended to perceive delivering fast as an individual skill that makes the job more motivating, like this Danish Wolt courier who did not experience time pressure as a job demand: “I like cycling a lot and I do it very fast”. And he continued by explaining why he likes working as a courier:

It’s about understanding how to find your way around the city without GPS. I know where most of the customers are. I can open and close a bag quickly. I know where to position myself at traffic lights so that I can go before everyone else. These are some small things you pick up.

Other freelance couriers did experience time pressure, which they sometimes linked to the fact that the platform company bundles orders that are too far away from each other. Here is one example, shared by a Wolt courier:

Sometimes the system [the app] is so bad. For example, it was so stressful yesterday, because they gave me three orders – two orders in X and one order far away in Y ….and I lost half an hour writing to support to ask them to drop the one in Y and give me the one in X. The customer ended up waiting ten minutes more because I lost time writing to support.

In sum, waiting time and time pressure during deliveries seem to be central characteristics of food delivery work, but couriers understand and cope with it differently. However, it would appear that the type of work arrangement (employed paid per hour or freelancer paid per delivery) has an impact on how severe or challenging waiting time and time pressure is perceived, with couriers who are employed and assigned orders (the case of Just Eat in Denmark) perceiving both waiting time and time pressure during deliveries as less problematic because their earnings do not depend on the number of deliveries they make.

Ropponen et al. (2019) found that that the physical work environment is a job demand in delivery work because couriers are specifically prone to riding fast on bikes, the risk of accidents and tough working conditions due to weather. These themes are also prevalent in our interviews.

Accidents in the food delivery industry in Norway and Denmark are common, according to the couriers we interviewed. In both countries, nearly all interviewed couriers had experienced accidents, ranging from minor to major incidents, either on bikes or on scooters. Several interviewees told us about more severe accidents, resulting in longer periods of income loss, and some were now afraid of working in bad weather due to the risk of accidents. 

The high risk of accidents could partly be attributed to the payment structure for couriers, where freelancers and self-employed individuals earn based on each delivery, incentivizing faster biking or driving to increase income. While employees follow a similar model, with an hourly wage, in some cases supplemented by delivery commissions, the pressure is less pronounced, as we demonstrated above.

Several couriers also described harsh working conditions. In the following quotation, a courier who was working as both a freelancer at Wolt and an employed courier at Foodora told us about how he experienced the physical work environment:

It is tiring, yeah. It is not mentally tiring, but it is physically tiring. In winter, if you are not moving, then you get cold, really cold. If you are not wearing gloves, if it’s just for a few minutes, then you your hands also freeze. So you have to keep warm inside the buildings, waiting until you get your next order. Because if not, if you are still outside, then you get really cold.

Similarly, a Danish employed courier working for Just Eat told us about working in the cold Danish winter. He was provided with gloves, but the temperature got so low that the gloves were not warm enough:

They [the gloves] cover everything reasonably well, but when temperatures get extremely low, there is really nothing that can keep the cold out. The problem is that it is progressive the longer you are out… It's just cold. I lost a lot of weight during that period. Two to three kilos in ten days. 

Furthermore, he expressed doubt about whether he would work as a courier next winter because the cold weather is too much for him to handle. These accounts highlight the profound impact accidents and harsh weather conditions can have on couriers’ well-being and income security.

The previous citation also touches upon another relevant theme, namely proper equipment and safety equipment such as clothes, helmets etc. Here we find differences between the platform companies. Just Eat provides employed couriers with the necessary equipment for free while the other companies rent out equipment (Foodora) or have the couriers buy it (Wolt). However, as we saw in the quote above, this equipment is not always sufficient: the gloves provided for the courier were not warm enough to keep the cold out.

In terms of safety equipment, employers can instruct their employees to wear helmets, for example, but this is not the case for the food delivery companies that use the freelance model. Considering the high risk of accidents in the industry, this is unfortunate. In a consultation response to a government-appointed committee on the Norwegian model and the future of work (NOU 2021:9), Foodora stated that they regretted their inability to enforce helmet-wearing among freelancers/​independent contractors:

Because this [setting a requirement for wearing a helmet] could be considered an exercise of management rights on our part, we have had to remove it from our agreements. We would like to emphasize that this is a very unfortunate result of the current legal situation and that we want to take the same safety considerations towards all our couriers, something that will be able to be taken care of largely through an extended obligation for the company to ensure a fully responsible working environment – including for self-employed contractors. (Foodora, consultation response, June 2021, our translation)

Ropponen et al. (2019) also highlight harassment as a characteristic of platform work that can affect workers’ safety and health. In our data, we found examples of couriers who had been subjected to harassment. Our data do not show harassment to be a daily occurrence, but we get the impression from the interviews that most couriers do experience it from time to time. For instance, one courier based in Denmark described how he had been pushed off his bike more than once by random people on the street, and a courier based in Norway explained that he had witnessed harassment at work: 

Yes. I was working for Foodora and I was in the restaurant and there was a guy from Wolt, but the time was overdue, the food should have been ready. And the cook was telling him: ‘Wait two minutes more, two minutes more.’ After a while, the Wolt rider got mad at him and started screaming, so the cook took a knife and threatened him. 

However, we also have several examples where food couriers described how they believe that they are being exposed to unfair treatment, often by the platform companies. This unfair treatment is often connected to intraplatform algorithmic change, which is where the platform companies change the way the algorithmic management system works or change other aspects of how they organize work in ways that are unfavourable to working conditions and couriers’ OSH. For instance, in Denmark in 2021 Wolt removed a weekend bonus for couriers having completed a certain number of deliveries and lowered the fee per delivery. Couriers argued that these cutbacks made their working situation even more insecure, and the grassroots organization Wolt Worker Group organized a demonstration against Wolt in Copenhagen (Hau and Savage, 2023). Couriers again mobilized in the spring of 2023 when Wolt launched a new payment model based on what they called more “dynamic price setting”. According to the couriers, the new model made the price setting opaque and was, in practice, a camouflaged payment reduction. 

Another example is from Foodora couriers in Norway, where the productivity assessment system has led to feelings of unfair treatment among Foodora couriers because they receive messages about their performance in the app and via e-mail. One informant, who worked as a freelance courier for both Wolt and Foodora, provided evidence of these messages through screenshots and communications from the company. Even when couriers have a high order acceptance rate, declining a single order triggers messages from the company stating that couriers are declining half their orders and warning that this will reduce their overall earnings, scores and performance numbers and if more orders are declined, then it will result in a temporary suspension from the app. In this sense, there is a perceived unfairness in the productivity measurement system, which penalizes couriers for declining orders, affecting their overall performance rating and income. This is also the case if couriers choose to take a break: “You do get a break, but it also gives you a lower performance rating because it counts as time you are not working”, said one Foodora courier. An employed courier based in Norway explained how he and other couriers perceived the app as harassing them: “I experience the app as harassing when I am put on pause because I am not quick enough to respond due to the weather conditions in Norway making it difficult to respond quickly in the app. I perceive this as a punishment because I am not fast enough”. “Put on pause” in this context means that the courier is temporarily suspended from the app. 

Additionally, informants also discussed their attempts to address similar instances of mistreatment, emphasizing the stressful and time-consuming nature, and often fruitlessness, of filing complaints. These complaints rarely resulted in any meaning­ful changes to their performance levels within the app, leading one courier based in Norway to remark: “It is not worth complaining”. Thus, couriers find it difficult to take actions that lead to change in the platform companies’ business models.

Finally, Ropponen et al. (2019) identify isolation and competition as two central work characteristics for platform workers that represent a significant job demand. They describe how digitalized platform work is often performed alone or separately from other platform workers with no face-to-face interaction with colleagues or supervisors, and in some cases also in competition with other platform workers. Ropponen at al. (2019) argue that this enhances the social isolation of this work situation, making communication with other platform workers, clients/customers and the platform company challenging, and couriers can feel anonymous and as though they do not have their own voice.

Since app-based food delivery is carried out in the physical world, the theme of isolation is not so pronounced in our empirical data. Even though work is carried out alone, couriers did not report feeling isolated or lonely. Most of them described a friendly atmosphere among the couriers, including talking while waiting for orders, and some of them know each other so well that they help each other out if needed. For instance, if someone is in an accident, others will help the courier and also deliver the order. In the case of the employed couriers at Just Eat in Denmark, they seem to have a more collegial relationship where they can engage in conversations with other employees while waiting for orders (as in the earlier quotation where they have a table at a Burger King where they can wait for orders). This gives them an opportunity to discuss work concerns in a social and physical environment.

In terms of communication with the platform companies, the couriers in all three companies can contact a dispatch centre if they have problems during their deliveries. Normally this is done through a chat function in the app. Issues range from having to stop the shift/​delivery due to mechanical problems with bikes to having trouble locating a customer, experiencing a long wait at a restaurant/pick-up place or finding that the food does not meet proper standards or that some­thing was damaged during the delivery. The call centre is in a way the only human intervention in the labour process the couriers’ experience. On the one hand, the couriers know that it is important to be in contact with the support team in case something happens as this courier from Denmark explained: “It’s important to be in contact with the chat, because if you don’t, you lose time”. He therefore under­stands that he depends on communication with the support team. Another courier shared that the support team gives him a sense of security during deliveries. He feels safe because he knows that the people providing live support at the head­quarters are monitoring him and that he can easily get in contact with them. On the other hand, couriers also reported instances of poor and frustrating communi­ca­tion with the support team and these seemed to be related to the fact that they most often communicate with the dispatch centre when something has gone wrong and they feel that the communication is about these problems, even though they are not always to blame. This is probably also connected to the physical distance between the couriers and the headquarters; the couriers found that the people working in support did not understand their perspective on how something happened when they were not physically present. The couriers did not mention feeling anonymous or alone like Ropponen et al. (2019) report, but we got the impression that they feel distanced from the people working in the support team.

Ropponen and colleagues also highlight competition as a characteristic of digitalized platform work, arguing that it can enhance feelings of isolation if workers are competing with others for employment and they don’t know who these others are. In our data, competition between food delivery couriers was not a theme as such, but in the case of Foodora Norway, which has a business model with both employees and freelancers, competition between couriers with different employment arrangements was an issue and was perceived by some couriers as discrimination or unfair treatment. One freelancer explained the competition as follows: “The employed couriers are not happy because they know how much we make, and they know that we try to get shorter distances to make more deliveries. Every time the freelancers get the short distance, employees get the longer distances”. One of the employees explained that removing the freelancers would be the best way to improve the working environment and OSH:

Get rid of the freelancers. It ruins the work environment and the working conditions. […] The conditions for employees are too poor, and people are forced into freelance contracts. It's a safety risk and it creates competition for assignments, which can also increase the risk of accidents.

In this chapter, we have explored working environment challenges for couriers working in different forms of employment in app-based food delivery (Wolt, Foodora and Just Eat) in Denmark and Norway. We have had a special focus on investigating the connection between algorithmic management and the couriers’ occupational health and safety.

In structuring our empirical data around working environment challenges, we took inspiration from the analytical framework from Ropponen et al. (2019), who identify a number of job characteristics of digitalized platform work that threaten occupational health and safety. However, we have also added job characteristics from our empirical analysis, including job and income security, waiting time and time pressure, the physical work environment, harassment and unfair treatment and isolation and competition. 

First, we found that job and income insecurity emerged as significant topics among couriers. All the interviewees, regardless of type of employment face job and income instability in a more formal sense, but its severity varies depending on what company they work for and their form of employment, with employed couriers having more security than freelance/independent workers. However, not all freelancers experienced insecurity. Those freelancers who managed to earn enough money did not seem to be concerned about the insecure nature of the job.

Second, we found that waiting time and time pressure were defining features of food delivery, but the interviewees coped with them differently. However, freelancers seem to be most exposed to time pressure, which is linked to their employment arrangement and payment structure, according to which they earn more the more deliveries they make. We also found that couriers have a physically demanding job, working outside in all types of weather and with equipment that is not always sufficient. Again, the physical demands of the job are connected to the type of employment the courier has and the payment structure that the companies use: some couriers have more incentives to bike or drive faster to increase their income.

Third, we found evidence of couriers experiencing harassment, and we also encountered cases where couriers felt they were being treated unfairly, especially by the platform companies. This perception of unfair treatment seems to be linked to algorithmic management and intra-algorithmic change, which makes it difficult for couriers to understand the rationale behind a function of the app or the payment structure.

Finally, while Ropponen et al. (2019) highlight isolation and competition as key job characteristics for platform workers, isolation does not seem to be a prominent job characteristic for food delivery couriers. Competition appears to be most prominent in the case of Norwegian Foodora, which uses different employment models that create dissimilar working conditions for couriers working for the same platform company. 

All in all, our empirical analysis reveals that couriers working in app-based food delivery in Denmark and Norway are exposed to health and safety challenges. App-based food delivery work is insecure and physically demanding and can be mentally stressful, but we do find differences between couriers in how they cope with these challenges. Furthermore, algorithmic management and intraplatform algorithmic change (for instance the productivity measurement system at Foodora and changes in the payment structure at Wolt) seem to encourage a more insecure working environment; couriers also find the system unfair and untransparent, but they have limited opportunities to get the system changed. Moreover, the work arrangement in place seems to impact the severity of the challenges: couriers employed under a collective agreement are less exposed to these challenges compared to freelancers, which can be explained by the fact that employers have a greater focus on the working conditions and health and safety challenges of their employees compared to the freelance model.

Eriksen, A.A. (2024). Madudbringer dropper Danmark. TV2. Accessed from: https://nyheder.tv2.dk/business/2024-05-27-madudbringer-dropper-danmark, 28. May 2024.

Demerouti, E., Bakker, A. B., Nachreiner, F., & Scaufeli, W. B. (2001). The job demands resources model of burnout. Journal of Applied Psychology, 86, 499–512. DOI: https://doi.org/10.1037/0021-9010.86.3.499.

Duggan, J., Sherman, U., Carbery, R., & McDonnell, A. (2020). Algorithmic management and app‐work in the gig economy: A research agenda for employment relations and HRM. Human Resource Management Journal, 30(1), 114–132. DOI: https://doi.org/10.1111/1748-8583.12258.

Hau, M. F., & Savage, O. G. (2023). Building coalitions on Facebook: ‘Social media unionism' among Danish bike couriers. New Technology, Work and Employment, 38, 529–547. DOI: https://doi.org/10.1111/ntwe.12261.

Ilsøe, A., & Larsen, T. P. (2022). Digitale arbejdsmarkeder. Fra jobs til opgaver. DJØF Forlag.

Jesnes, K. (2024). Shifting gears: How platform companies maintain power in app-based food delivery in Norway. Transfer: European Review of Labour and Research, 29(4), 507–524. DOI: https://doi.org/10.1177/10242589241228199.

Jesnes, K., Neumann, D., Trappmann, V., & de Becdelièvre, P. (2021). The role of worker collectives among app-based food delivery couriers in France, Germany and Norway: All the same or different? In J. Drahokoupil & K. Vandaele (Eds.) A modern guide to labour and the platform economy, pp. 258–273. Edward Elgar Publishing.

Jesnes, K., & Oppegaard, S. M.N. (2023). Plattformmediert gigarbeid i Norge. Fleksibilitet, uforutsigbarhet og ulikhet. In T. Fløtten, H. C. Kavli, & S. C. Trygstad (Eds.) Ulikhetens drivere og dilemmaer, pp. 137–153. Universitetsforlaget. DOI: https://doi.org/10.18261/9788215065403-23-08.

Madudbringningsoverenskomsten. (2021–2023). Madudbringningsoverenskomsten. Dansk Erhverv and 3F.

Madudbringningsoverenskomsten. (2023–2025). Madudbringningsoverenskomsten. Dansk Erhverv and 3F. 

Mendonça, P., & Kougiannou, N. K. (2023). Disconnecting labour: The impact of intraplatform algorithmic changes on the labour process and workers’ capacity to organize collectively. New Technology, Work and Employment, 38(1), 1–20. DOI: https://doi.org/10.1111/ntwe.12251.

NOU 2021:9. Den norske modellen og fremtidens arbeidsliv — Utredning om tilknytningsformer og virksomhetsorganisering. Ministry of Labour and Social Affairs.

Rasmussen, S., Nätti, J., Larsen, T. P., Ilsøe, A., & Garde, A. H. (2019). Nonstandard employment in the Nordics – Toward precarious work? Nordic Journal of Working Life Studies, 9(S6), 7–31. DOI: https://doi.org/10.18291/njwls.v9iS6.114689.

Ropponen, A., Hakanen, J. J., Hasu, M., & Seppänen, L. (2019). Workers’ health, wellbeing, and safety in the digitalizing platform economy. In S. Poutanen, A. Kovalainen, & P. Rouvinen (Eds.) Digital work and the platform economy: Understanding tasks, skills and capabilities in the new era. pp. 56–73. Routledge. DOI: 10.4324/9780429467929-4.

Srnicek, N. (2017). Platform capitalism. John Wiley & Sons.

Wolt. (2022). Algorithmic transparency report 2022. Wolt. Accessed from https://assets.ctfassets.net/23u853certza/5G5O7KFnwzDGWzE1JFwCN/b4f9307af92b4ad9dc8a32ae09d7a9e0/Wolt_Algorithmic_Transparency_Report_2022.pdf, 10 May 2024.

Wolt. (2023). Algorithmic transparency report 2023. Wolt. Accessed from https://downloads.ctfassets.net/23u853certza/6zSHjJiE1XtTrD5TzMtvh1/cef4f51b2670b56df900e626c6d43015/Wolt_Algorithmic_Transparency_Report_2023.pdf, 10 May 2024.