Growing together and drifting apart: on the emergence and fragility of social bonds among food delivery riders

Research on platform labour often portrays food delivery work as spatially dispersed and socially isolating, framing fragmentation as a key obstacle to solidarity and collective organising. While this literature has generated important insights into the complex conditions of workers' resistance, less attention has been paid to the social significance of shared logistical spaces such as hubs and warehouses. This article addresses this gap by examining how social bonds among food delivery riders emerge, evolve and disintegrate in everyday work settings. The paper critiques conflict-centred accounts of solidarity and instead develops an interactionist perspective on the topic. Based on 22 autobiographical narrative interviews with food delivery riders in Hamburg, it reconstructs different forms of social bonding that are, however, fragile and subject to erosion through organisational change; competition; ageist, ethnic and class-based boundary making; and gendered vulnerabilities. By uncovering how intersectional inequalities shape riders' everyday interactions, this article contributes to the feminist understandings of platform labour and care.

1 Introduction

Food delivery work is often depicted as an isolated form of platform labour. Riders spend substantial parts of their working day alone: travelling through the city, waiting at restaurants and delivering orders to customers' doors. Orders are assigned through digital systems, and communication with supervisors is increasingly mediated by applications and algorithmic management. This spatial and organisational fragmentation has consequently been identified as an important obstacle to the development of workers' collective identities and solidarity structures (Heiland, 2021).

Yet the image of the isolated rider does not capture the full range of social relations that characterise food delivery work. Riders encounter one another while waiting for orders (Richardson, 2020; Ecker and Strüver, 2022), communicate through WhatsApp groups (Popan, 2024) and, particularly in grocery delivery, spend considerable periods of time in shared hubs and warehouses. These spaces are not merely logistical sites. They can also become places where workers rest, exchange information, joke, compete, get into conflicts, develop reputations and friendships, and construct memories of a shared workplace (Orth, 2022).

Taking this observation as a starting point, this article focuses on ways in which social bonds among food delivery riders emerge, develop and disintegrate over time: what can be learned about (a) the conditions and processes of developing trust and solidarity vs (b) the conditions and processes in which the development of solidarity is impeded and prevented, or ultimately collapses? Rather than treating solidarity only as a collective response to exploitation or as a precursor to organised resistance (Tassinari and Maccarrone, 2020), it pays attention to seemingly inconspicuous forms of socialising such as learning from experienced colleagues, helping one another, spending time together and discovering aspects of co-workers' biographies that exceed their current ways of making a living. At the same time, the analysis does not romanticise these relationships. The same conditions that allow for new we-groups to emerge can be exclusionary as well as inclusive. Workers draw boundaries along the lines of nationality, legal status, race, age, class and gender, while their different migration histories, economic obligations and positions in the labour market shape what they have at stake in their work (van Doorn and Vijay, 2024; Baril, 2023).

The article makes three contributions. First, it shifts attention from the question of whether platform workers are socially fragmented to the processes through which social worlds (Strauss, 1978) nevertheless emerge through interactions within fragmented labour arrangements. Second, it demonstrates that social bonding is ambivalent: cooperation and competition, belonging and exclusion, and solidarity and boundary making can coexist within the same workplace. Third, it shows how these relationships need to be understood biographically. Riders who appear to occupy the same occupational position may attach very different meanings to their work because of differences in migration histories, education, (transnational) family responsibilities and economic circumstances (Riemann et al., 2023).

The article also contributes to a feminist perspective on food delivery work. Food delivery is part of the wider platformisation of activities associated with social reproduction: cooking, shopping and the organisation of everyday household life can increasingly be outsourced through digital platforms (Dowling, 2022; Lentz et al., 2025). Yet this outsourced reproductive labour is performed under highly unequal conditions by a workforce that is predominantly male and migrantised. A feminist analysis therefore draws attention not only to the gender composition of the workforce but also to the racialised, classed, ageist and gendered relations through which reproductive labour is reorganised (Strüver, 2024). The social bonds that develop among riders may provide resources for coping and recognition, but they can also reproduce gendered and racialised hierarchies.

The empirical material consists of 22 autobiographical narrative interviews (Schütze, 2008) with food delivery riders in Hamburg, conducted in 2023 and 2024. The open interview format allowed participants to reconstruct how they entered platform work against the backdrop of their life histories – and how their experiences changed over time. The analysis focuses particularly on spontaneous narratives concerning relationships with colleagues, many of which centred on hubs and warehouses. Because these interactions were not observed directly, the article analyses workers' narrated experiences and their own retrospective interpretations of social encounters rather than claiming to provide ethnographic observation of interaction itself. Although this limitation is important, particularly when interpreting accounts of relations between different groups of workers, the open interview format allowed interviewees to openly share how they felt in certain situations – inner thoughts that may not be instantly “observable” from an outsider's perspective.

The article proceeds as follows: the next section discusses existing research on solidarity, fragmentation and platform labour, and develops an interactionist and feminist perspective on social bonding. The methodology section then explains the biographical approach and its implications. The empirical sections examine how workers grow together through shared practices and how relationships become divided through competition, racialised and classed boundaries, and gendered vulnerabilities. A final section analyses organisational change, workforce re-composition and collective disenchantment before discussing the broader implications of the findings.

2 From labour fragmentation to everyday solidarities

There is a significant body of research that has demonstrated how platform workers face multiple forms of labour fragmentation (Della Porta et al., 2022): legal arrangements can outsource risks and responsibilities towards workers through, for example, (bogus) self-employment (Ametowobla and Kirchner, 2025) and subcontracting (Wiese and Grieger, 2025). Technological systems can intensify surveillance, forms of semi-automated management and algorithmic control, which add to a sense of alienation (Heiland, 2025; Bonini et al., 2024). Organisational arrangements can individualise performance and heighten competition; and the spatial dispersion can make collective organising more difficult (Graham et al., 2017; Della Porta et al., 2022). Furthermore, the sector is known for its high turnover rate, as many riders only stay for short periods of time. The heterogeneity of platform workers adds another level of complexity, as riders differ in nationality, migration status, language, religion, gender, age, education, class background and financial dependency (van Doorn et al., 2023; Stingl and Orth, 2024; Altenried, 2024; Baril, 2023). These differences can affect workers' willingness and ability to participate in collective action (Popan and Jesnes, 2025; Mendonca and Kougiannou, 2025).

The literature on different forms of labour fragmentation is useful for understanding why platform workers may struggle to develop collective power. These obstacles should, however, not be mistaken for an absence of social relations. Research on WhatsApp groups and other forms of digital communication (Bonini et al., 2024; Kwan, 2022), for example, shows how riders use shared communication infrastructures to exchange information, provide support and address practical problems as well as coordinate strikes. Reoccurring topics of conversation include dangerous zones, restaurants with long waiting times and payment problems (Popan, 2024). Social interactions among riders therefore form part of the everyday infrastructure through which platform work is made possible.

This observation is particularly relevant for food delivery because the sector occupies an intermediate position on a spectrum of social isolation. Unlike domestic workers who may work alone in clients' homes (Rodríguez-Modroño et al., 2022; Lentz, 2025) or online freelancers who work remotely (Riemann et al., 2023), riders regularly encounter colleagues in urban and logistical spaces. The extent of this interaction differs according to platform organisation. Meal delivery riders such as those working for Lieferando may visit hubs mainly at the beginning and end of their shifts to pick up and return their bikes and other equipment. Grocery delivery riders at, for example, Flink return to warehouses between deliveries to pick up products. The spatial organisation of work consequently shapes the opportunities available for social interaction (Orth, 2022). The relevant question is therefore not whether platform labour is fragmented but how workers create and negotiate relationships within fragmented arrangements.

Given the adverse conditions, it is quite remarkable how food delivery platforms have been subject to multiple forms of riders' collective resistance, including public protests, coordinated mass logoffs and the formation of (grassroot) unions in multiple countries (Cini et al., 2022; Borghi and Murgia, 2022). Some initiatives have contributed to tangible legal and political changes such as the EU Platform Directive (Fredman et al., 2025) and the “Riders Law” in Spain (Waeyaert et al., 2022). These types of workers' movements have attracted the attention of social scientists (Woodcock and Cant, 2022; Cant, 2019), who commonly approach the phenomenon from a Marxist perspective. They are interested in solidarity formation primarily in resistance to oppression and exploitation, and as a “means” to achieving success in industrial disputes (Gandini, 2019; Braverman, 1994). This is particularly visible in, for example, Tassinari and Maccarrone's (2020) concept of “embryonic solidarity”, which refers to a certain “preparedness to act” based on collective feelings of reciprocity and responsibility. It hereby serves as the basis for the emergence of “active solidarity” in the form of strikes, the formation of workers' unions and so forth.

This literature is very important for understanding organised struggles. It is nevertheless less well suited to analysing everyday encounters and inconspicuous forms of solidarity that do not lead to collective action. Workers may share information, offer practical assistance or develop friendships without forming a political collective. At the same time, they may also experience forms of exclusion depending on, for example, their nationality, language skills, religion, age and gender. These forms of relationality matter because they shape how people experience platform work, how they interpret their position within a workplace and how they evaluate their current occupation in the context of their life histories. They also contribute to a more nuanced understanding of “intersectional precarity” in platform economies (Popan and Anaya-Boig, 2021). Rather than opposing everyday social bonding to political solidarity, this paper therefore treats them as analytically connected but empirically distinct. The point is not to replace a conflict-centred account with an idealised account of community but to examine how social relationships can emerge under adverse conditions and how their meanings change over time.

Taking this observation as a starting point, this paper focuses on everyday interactions between riders from a bottom-up perspective to discover (in the sense of Glaser and Strauss, 1967) how rider communities emerge and disintegrate over time. This is something that has been examined in a complex manner by sociologists in the research tradition of symbolic interactionism (Blumer, 1969) and in early Chicago sociology (Chapoulie, 2020). The data revealed how warehouses and hubs play a significant role in workers' day-to-day lives, as they serve not only as logistical sites and workplaces but also as recreational spaces, places of surveillance and a “home territory” in which workers build reputations and relationships. The same place can therefore sustain multiple and sometimes contradictory meanings, which change over time.

By adopting an interactionist lens, the paper focuses on the situated processes through which workers construct shared meanings. The analysis is sensitised by the “social world perspective” (Strauss, 1978), which is a sociological framework for understanding how collective action, meaning and organisation emerge through interaction. At its core, the theory suggests that society is made up of multiple social worlds, i.e. groups of people who share a common activity or set of practices, shared meanings and resources, and continuously communicate about what they do. Social worlds are not fixed organisations but fluid, overlapping and negotiated through interaction. In this context, Strauss (1978) also introduced the concept of “arenas”, where multiple social worlds meet, clash, negotiate or cooperate. These are spaces where issues are debated and decisions are contested (e.g. a hospital ethics committee). This perspective takes long-term social processes into account and does not remain fixated solely on “micro-phenomena”. While Strauss's “social worlds” concept is helpful in analysing the emergence of “we” groups and processes of prevention and decline of solidarity among workers, it is important not to simply subsume such phenomena under this perspective. It remains necessary to discover the different meanings of hubs in workers' narratives beyond the functions they have as logistical sites (Glaser and Strauss, 1967). It does not make sense to, for example, reduce them to “arenas” in which participants negotiate over issues that are relevant. At the same time, a hub can acquire such a meaning in certain circumstances if, for example, the topic of a “riders' world” emerges in labour struggles.

A feminist perspective adds two further dimensions. First, food delivery is part of the wider platformisation of social reproduction (Strüver, 2024). Tasks such as grocery shopping, preparing meals and organising everyday household consumption (e.g. doing the dishes) can be outsourced to platform workers. The apparent convenience of platform-mediated consumption is therefore connected to the organisation of labour elsewhere. Second, the work is performed within intersectionally unequal relations. The workforce in the food delivery sector is predominantly young and male, which is also reflected in this study's sample composition. Against this backdrop, it is important to study how social bonds formed at work can provide support to some while also reproducing exclusionary hierarchies concerning masculinity, respectability and belonging (Kwan, 2022). Staying sensitive to intersectional inequalities (Popan and Anaya-Boig, 2021) also requires attention to workers' biographies. The interviewees differed significantly in terms of their age, health, economic responsibilities, class backgrounds, migration histories, education levels, family constellations and legal status. A salary that represents supplementary income for one rider may be essential for other workers' livelihoods and that of their dependants (Schor et al., 2020). The same workspace can therefore be experienced quite differently, which has an impact on how workers socialise and relate to each other.

3 Methodology and fieldwork

The sample consists of 22 autobiographical narrative interviews (Schütze, 2008) with food delivery riders in Hamburg, which were conducted online in 2023 and 2024. Most of the interviews were conducted in English and some in German. They were then transcribed and partially translated if necessary. The interviewees were recruited in several ways, including an open call on flyers, direct messages on LinkedIn, and through snowball sampling. When approaching certain prospective informants, I was also – as far as possible – guided by the principle of theoretical sampling (Glaser and Strauss, 1967): trying to explore certain issues that had emerged during the fieldwork.

The participants worked for different meal (e.g. Lieferando) and grocery delivery platforms (including Flink, Gorillas) on full-time and part-time bases. I also included former riders, who had left the food delivery sector because their company (e.g. Gorillas or Getir) went bankrupt or left Germany, because they were fired, because they reached a point of physical or mental exhaustion, or because they found a better position elsewhere. The purpose was to capture how a platform's gradual decline was reflected in the day-to-day experiences of riders. Most interviewees were male (20 men and 2 women), between 20 and 46 years old and had a migration background. The countries of origin were extremely diverse, as the sample included, for example, many international students from Ghana, Türkiye, India and Bangladesh. There were a few older German nationals in the sample, who had a long history of working in precarious jobs and experiencing underemployment and unemployment.

The interviewees were first encouraged to narrate openly “how one thing led to another” in their lives. After finishing their introductory narrative, which usually entailed detailed narrative and descriptive sequences on delivery work, they were then asked different spontaneous and pre-formulated follow-up questions. The open format of the interview gave the interviewees space to narrate in detail how their lives evolved prior to entering the food delivery sector in Germany. These narratives were very rich and shed light on numerous different collective crises, including poverty and violence in western and southeast Africa, gendered violence in the Middle East, familial pressures and expectations in southern Asia and mass unemployment in southern Europe. What they shared is that they eventually found themselves in a situation where becoming a rider appeared as one of the few options – or often as the only option left – to be economically active due to the platforms' inclusive hiring strategies.

The analytical approach chosen for this paper was first developed in Germany (Schütze, 2008) in the late 1970s and early 1980s, and has been mainly used in biographical research in sociology and other social sciences. It is characterised by an intensive socio-linguistically-based interpretation of narrative interviews in combination with comparative analyses in the style of grounded theory (Glaser and Strauss, 1967). The basic step of a detailed sequential structural description of the interviews includes the consideration of formal textual features relevant to the analysis, such as the sequential structure of the narrative, the alternation between different schemes of communication (narration, description, argumentation) and phenomena of disorder (self-interruptions, background constructions) (see Riemann, 2018). In sequentially analysing and comparing the interviews, I also discovered that interviewees spontaneously referred to phenomena (in narrative and descriptive sequences) which revealed something of how they experienced the relationship with their co-workers (especially in the context of hubs and warehouses).

The interviews provide access to workers' memories of their experiences and how they make sense of them, including how they remember different encounters and how these relationships affected their evaluation of platform work. They do not provide direct observation of interaction, as would be possible in ethnographic studies (Timko and van Melik, 2021). They do, however, reveal their past experiences and their current ways of theorising. Adopting a biographical approach has the advantage of making temporal changes visible, as participants could reconstruct how relationships developed and sometimes failed, how workplaces changed and how their own circumstances affected their interpretation of those changes. The resulting analysis is therefore concerned with processes rather than a snapshot of current working conditions.

4 Findings

4.1 Growing together: the emergence of social bonds

4.1.1 Hubs as relational spaces

The spatial organisation of work strongly influences the opportunities for social interaction. Meal delivery riders move more independently throughout the city, as they are automatically assigned orders from different restaurants based on their current location. At Lieferando, for example, workers tend to visit the hub only twice a day to start and end their shifts. This is different in the case of grocery delivery platforms such as Gorillas or Flink, as workers need to return to their respective warehouses after each delivery. In addition to storing electric bikes and other equipment, these hubs are also designed as a form of supermarket – with additional staff (“pickers”) in charge of preparing orders. As Ernesto explained:

At Lieferando, you're much more independent and you're just by yourself. If you don't get any order, you can just chill. […] Whereas in Gorillas, you always have to come back to the hub. […] At Lieferando, the city is your home, your workplace. At Gorillas it's the hub and the surrounding houses.

Orth (2022) argues that due to the spatial configuration of grocery delivery, Gorillas workers can co-produce social infrastructures, which in turn make or break organising efforts at companies. She observed how warehouses provided a “ready-made physical space for workers to cultivate these relationships by `hanging out' between deliveries” (Orth, 2022:165). In contrast, the meal delivery couriers (e.g. Lieferando) in the sample tended to have a more distant relationship to the hub, as they spend significantly less time there. Interactions with others, while at times quite meaningful, were less frequent and more transitory in nature.

The interviews suggest that these spaces acquired a shared history. Workers remembered conversations, conflicts, jokes, relationships and changes in the composition of the team. In this sense, the hub became a social territory layered onto its logistical function over time. The existence of these spaces does not mean that workers automatically formed a cohesive community. Rather, repeated encounters created opportunities for relationships to develop.

4.1.2 Learning, advice and informal mentorship

New riders were initially introduced to the workplace by supervisors or rider captains. Over time, however, they also learned from more experienced colleagues. Veterans provided informal advice about how to manage work pressure and how to navigate the platform's systems. In some cases, these relationships involved co-nationals who adopted newcomers and instructed them in their own language. This was not surprising, as many interviewees had first entered the food delivery sector through contacts in their own religious and/or national communities (Heiland, 2025; Popan and Jesnes, 2025). This advice was not necessarily political or organised. It could concern how to manage waiting times, how to distribute effort over a shift, how to deal with supervisors or how to avoid unnecessary pressure. Riders also described practical ways of working around monitoring and algorithmic controls. Examples included informing newcomers about specific labour laws prohibiting supervisors from formally pressuring workers to drive faster based on their performance data (“because doing so implies that the person could cause an accident” – Felix) and switching off phones shortly before the end of a shift to avoid unpaid overtime (Iazzolino and Varesio, 2023).

4.1.3 Corporate teambuilding, gamification and competition in times of COVID-19

Several interviewees had started to work in the sector at the onset of the COVID-19 pandemic when food delivery apps were relatively new. They often expressed a certain fondness about the “early days” at the respective platforms, which were characterised by an atmosphere of optimism. This was especially apparent among former Gorillas riders, who were encouraged to identify with the company's future-oriented rhetoric. Owing to the unique historical circumstances, the warehouse became one of the only spaces where people could socialise at a time when much of the city was closed.

Rafael recalled:

Then, of course, the CEO made everyone believe that this would be the future of everything, that the different groceries/supermarkets wouldn't be there, and everyone was buying into that idea. And then, the social part (of Gorillas) was really nice because it basically, it wasn't like Lieferando where you have certain rides, and you go and deliver them. You had a location with a kitchen and people (…) So, I made a lot of friends there. Even my ex-girlfriend, I met her in Gorillas. And I would say not only for me, but for a lot of people, it gave them everything, it gave them friends, relationships.

During the pandemic, the Gorillas warehouses emerged as a home territory, where workers could continue to socialise while the rest of the city was in lockdown. This contributed to an atmosphere in which shared meaning resources were mutually reaffirmed: being part of something special and “cool” while “doing an important service” (Ernesto). The latter refers to food delivery labour as temporarily being elevated to the status of “essential work” due to social distancing measures.

During the COVID-19 lockdowns, many restaurants, clubs and cultural institutions had to close or severely restrict their services due to social distancing rules. The unique historical circumstances at the time meant that a significant number of skilled professionals with high cultural capital (Bourdieu, 1986) were put out of work. Given the lack of alternatives, many people opted for food delivery jobs to stay economically active. When describing the special atmosphere during the “early days” of Gorillas, Manuel remembered:

It was a crazy time, because of the curfew and everything. We got a special permit, and at some point, we were the only ones outside in this big city. That was really crazy. (…) I met incredibly interesting people, especially during the early days. I met Michelin star chefs, sommeliers … and there was this DJ, this woman, who is a DJ and she has really skyrocketed internationally. (…) At night, we used to stay at the hub and have a beer and it felt like – when everything else was closed down, all the clubs – it felt like a party really.

The data suggest that food delivery hubs emerged as international – and sometimes unexpectedly cosmopolitan – social spaces. A new kind of work milieu emerged which participants enjoyed at the time, but it was nevertheless very fragile and highly dependent on external conditions (i.e. the easing of social distancing rules).

The same period was also characterised by the ambivalence of gamification (Woodcock and Johnson, 2018). Christian, who was 27 years old at the time of the interview, remembered a competition around delivery speed at the start of his platform career:

There was the KPI [key performance indicator], which was super important, the “average delivery time”. I remember these early times in the beginning in Ottensen, when we were temporarily the fastest warehouse in Germany in terms of the “average delivery time”. We had times when the “average delivery time” was just 6 $\frac{\mathrm{1}}{\mathrm{2}}$ minutes. This was, of course, really amazing. We had so much fun (…). We had a really friendly battle between the riders. Who can do more orders? Who is faster? That was really motivating.

The practice created a shared “we” through competition. Yet this was not workers' solidarity in a straightforward sense. It did not involve “gaming the system” from below as a form of collective “everyday resistance” (Iazzolino and Varesio, 2023:878) but rather voluntarily participating in a quasi-athletic competition that was imposed from above to increase profitability. This type of gamification blurred work and play while simultaneously intensifying performance expectations. The excitement of competing with colleagues could make productivity feel like a collective game, even as the underlying activity remained physically demanding and remunerated at low levels (Woodcock and Johnson, 2018). The game was generally experienced as enjoyable by riders, who were male, young and athletic. It was a different story for those who deviated from this norm, including non-athletic, inexperienced, disabled, older and female riders. Ben, who was 46 years old, struggled to keep up with the pace of his colleagues, who tended to be at least 20 years his junior:

On the phone you get told to hurry because you get shown your performance on the app. It would say, “Well done” or “Hurry up” and things like that. It's a bit degrading, yeah [laughs]. Being patronised by an app on the phone. At the beginning of working there, you'd experience, get, “Oh, I'm a snail, okay. Great.”

Receiving the image of a “snail” as feedback for one's performance can be interpreted as a stigmatising mechanism (see Pulignano and Domecka, 2025). It goes to show how the same “game” that some workers perceive as a unifying activity can be experienced as discriminatory by others. This ambivalence is consistent with research showing that gamification can encourage workers to internalise performance targets while also generating tactical forms of engagement with platform systems (Woodcock and Johnson, 2018; Iazzolino and Varesio, 2023; Santabarbara et al., 2025). The point is therefore not that gamification simply creates solidarity. Rather, it can generate social bonds and exclusionary mechanisms through practices that simultaneously individualise and intensify labour.

4.1.4 Recognition beyond downward occupational mobility

The biographical narratives – especially those of recent migrants – commonly revealed processes of downward occupational and social mobility. Sometimes the narratives revealed situations in which colleagues listened and gave advice concerning this sensitive matter. Layla, a 40-year-old Moroccan woman, remembered how a colleague confided in her:

One day, I was with a colleague, and he met with a girl, and I think he was interested in her. And he told me, “Now, look, I'm driving, I'm wearing the Lieferando uniform, she will know that I'm a rider”. I said, “Nobody cares about the uniform”. I didn't want to push upon the fact that, yes, this is like a small job and it's making a bad impression on girls.

The interviewees sometimes described interactions in which colleagues' previous biographical achievements became visible beyond their status as platform workers. Emma, a 21-year-old German student, remembered an older Ghanaian man who worked in cleaning and had previously been a teacher:

There is a man, he's there, he's always there to clean and tidy up the place. (…) And when you talk to him, you notice that he was a teacher. He is an incredibly smart man, and in my opinion overqualified for the job that he is doing. You realise that he knows a lot about many things, and he really enjoys talking to young people. He always asks me, for example, how my exams went and when I have them. He always remembers and asks afterwards: “How did it go? You have to finish your degree, it's very important!” You can tell that he used to be a teacher, and I am very sure that he was a good teacher.

The significance of this interaction lies less in its supposed “cosmopolitanism” than in the temporary suspension of the apparent occupational downgrade. The encounter allowed Emma to recognise a colleague through aspects of his biography that were invisible in the occupational hierarchy of the hub. These moments were not necessarily transformative or lasting, but participants remembered them as meaningful when reflecting on their experience. Social interactions can therefore provide moments of biographical continuity in which identities suppressed by current occupational hierarchies become visible (Riemann, 2022:142–143).

4.2 Drifting apart

4.2.1 Competition and ethnic boundary making

The same spaces that enabled social bonding could also produce conflict. At Gorillas and Flink, riders were initially given greater autonomy in distributing orders among themselves. This was a short experimental stage after the platforms' first launch before they moved towards full automatisation. The interviewees remembered how they could decide informally who took which order. This created opportunities for negotiation but also for competition. Pedro recalled:

In the early days, we had this chaotic … in fact non-system, which meant that people took orders referring to their own experience. If I get in the earliest, that means I get the order. Many Africans stuck together. They supported each other when it came to the question “but it's his turn”, “no, no, no”, and the Africans always stuck together. (…) In the beginning, you just didn't know each other and the atmosphere was respectful. Some people also held back, and lost out to avoid conflicts. But yeah, with the Africans … I always had the impression that it escalated a bit. There were quite a few who would get very loud.

He also described instances in which workers allegedly changed names on a paper list to obtain orders more quickly. These accounts reveal the emergence of ethnic boundary making, but they cannot simply be treated as objective descriptions of African workers' behaviour. They are accounts produced from a particular position within an unequal workplace. To avoid reproducing these boundaries analytically, it is important to place them alongside African workers' own accounts. Alan, a 21-year-old Ghanaian worker, described the presence of many co-nationals at Gorillas very differently. For him, the concentration of Ghanaian workers could provide a sense of familiarity and mutual support:

I felt at home, because most of the other riders were from Ghana. You'd hear a lot of Twi, a language in Ghana but also English. (…) I know that Ghanaians, basically Ghanaians are people who support each other. That's why I love Ghanaians in Hamburg. Because if one Ghanaian comes in, he's going to bring his friend, his relatives. […] You have the Ghanaian community that works for each other. They help each other grow and everything like that.

At the same time, Alan did not describe this community as homogeneous or unambiguously positive. He spoke about generational differences and about the need to decide which aspects of the community's expectations fitted his own goals. His father, who had lived in Germany for over 20 years, deeply distrusted the Ghanaian community in Hamburg and had even prohibited his children contact when they first joined him as teenagers. Alan's account therefore complicates both the negative perception of alleged ethnic favouritism and a romanticised image of “African solidarity” (see Baril, 2023). The contrast between these accounts demonstrates that the same practice can be interpreted differently depending on one's position. Recruitment through friends and relatives can appear to outsiders as preferential treatment while functioning internally as a resource for migrants navigating an unfamiliar labour market. Bonding and exclusion can therefore occur simultaneously.

4.2.2 Different stakes: migration, class and transnational obligations

The intergroup tensions among riders, for example, complaining about advantages that “the others” have obtained illegitimately, need to be understood in relation to substantial differences in workers' economic and biographical circumstances. James, a 28-year-old worker from Zimbabwe, described the transnational financial obligations faced by many African migrants. He used the term “black tax” to describe the expectation that migrants send money to relatives:

Most of my African brothers and sisters have to do which we know defined a term; we call it black tax. It's basically where you send money back home because you have to support your families back home. So, most people here, like, I think 90 %, have to send a certain amount back home every month. You find that it becomes difficult, or it doesn't, even if it's not even every month, sometimes it can be, after every 2 weeks you hear something like, “Oh, your brother did this or this. Oh, mama, she's not feeling well; this happened, you need to send money.”

For James, these obligations were incorporated into everyday budgeting. He explained that relatives might request money for medical treatment, children or education, meaning that financial support was not an occasional act but something workers had to anticipate. He also described his own efforts to support his sister while she was studying:

Right now, my sister, she's studying. So sometimes she'd be like, “I need money for this, I need money for that.” Then, I know she's a girl. I don't want her to live in a situation where she is now compromised to do other things so that she can get money. So, I always try with all means, if I can, to assist her, so that she has everything she needs. So, she doesn't have any, what you call this, an excuse to say, okay, I ended up sleeping with this old man, because I wanted to get money for this. You did that because you wanted to do it – not based on that I didn't have money to support myself.

This account demonstrates how paid work in Germany can be connected to social reproduction across borders. James's work as a rider was not only about sustaining himself. It was also connected to his responsibilities towards family members elsewhere, who he considered to be in danger. These obligations help to explain why workers can have different stakes in the same workplace, may engage in practices others perceive as ruthless (“the Africans always stuck together”) and why participation in collective struggles may carry unequal risks (Mendonca and Kougiannou, 2025).

Other participants expressed their differences through class boundaries. Recep, who had qualified as a lawyer in Türkiye and came from an affluent family, described discomfort at being around colleagues whose behaviour he associated with a lower social status: “I was questioning myself why I am among these people”. His account illustrates what can be understood as a biographical breach: migration and occupational displacement had placed him in a workplace with people whose social backgrounds differed sharply from his own. These differences illustrate how workers can share an occupational setting while occupying very different positions within wider relations of class, migration and family responsibility.

4.2.3 Gendered vulnerabilities

The food delivery workforce in the sample was overwhelmingly male. Only two of the 22 interviewees were women. This numerical imbalance is not merely descriptive. It shapes the social environment in which work takes place. Female participants described a lack of female restrooms, unwanted attention and certain interactions that made them feel uncomfortable (Kwan, 2022). As a visibly Muslim woman, Layla recalled how her presence at the hub caused some irritation, primarily among fellow Arab riders, who expressed both judgement and paternalistic concern:

Arab people are not used to see women driving on a bike. They say, “Doesn't she have a father or a husband who would take care of her? How can you let a woman do that kind of job?” I said, okay, I'm thinking [with] my brain, my family doesn't know, they will not allow me to do this kind of job if they [did] know. If they [did] know, they would say, “Come back to Morocco, we didn't [send you] to do this kind of job.”

Prior to moving to Europe, Layla had lived in a very sheltered environment in Morocco, where men and women socialise in segregated spaces (Chafai, 2021). At the Lieferando hub in Hamburg, she suddenly found herself alone among dozens of men. She recalled multiple interactions she considered disrespectful, including jokes about her marital status. Layla was particularly disappointed that these comments came from Muslim riders, who she believed “know the culture. It's not by mistake”.

Felix, who had held a managerial position at Flink, similarly described the workplace as highly male dominated. He observed that female shift leaders could be treated as less legitimate and that women sometimes experienced sexualised attention.

It is very male dominated. If I was a woman, I would rather do another job. The reason is that there are so many men, and often men, who have different cultural backgrounds. They didn't grow up with the concept of gender equality. A good example: my girlfriend at the time helped out at a hub for a few days, but she said she would never do that again. She felt that they undressed her with their eyes. That's how she felt. (…) Oftentimes we also have this problem that a female shift lead is respected less, and they tell her: “No, I won't do that. I'm not going to let a woman tell me anything.”

These experiences demonstrate that the social bonds formed at hubs are gendered. The informal atmosphere that some male riders experienced as friendly and communal could simultaneously be experienced by women as uncomfortable or exclusionary (Kwan, 2022). This matters for a feminist analysis of platformised care and social reproduction. Food delivery is part of a wider reorganisation of everyday reproductive labour: activities such as shopping and obtaining meals can be delegated to platform workers. Yet the workers performing this labour are themselves embedded in unequal gendered and racialised relations. The social infrastructure of the platform therefore redistributes reproductive labour without necessarily challenging the hierarchies through which that labour is organised (Strüver, 2024).

4.2.4 Processes of collective disenchantment and decline

The social bonds that emerged in the early period were not stable. Many interviewees described increasing managerial intervention, surveillance and pressure from company headquarters. Alan recalled that supervisors increasingly contacted riders during deliveries to ask where they were and what they were doing. The qualities that made the work meaningful to Alan – helping customers, interacting with people and preserving some autonomy in decision-making – were increasingly experienced as incompatible with managerial expectations of speed and monitoring. He later described the deterioration of relationships among workers: “At some point, there were a lot of fights. […] A lot of backstabbing. The chemistry went away, just literally didn't exist anymore.”

The “chemistry” that had previously characterised the hub was therefore not simply undermined by interpersonal conflict. It was affected by changes in how the work itself was organised. Several participants associated the deterioration of social life with changes in the composition of the workforce (Popan and Jesnes, 2025). During the early pandemic period, food delivery attracted people from cultural industries, gastronomy and other sectors that had temporarily closed. Some of these workers later left when their original occupations became available again or when better alternatives emerged. Eric described the change at Flink:

Everything became more. More and more riders, more orders, more exhaustion of colleagues, and more deliveries […] It all started out well […] and then it gradually changed and Berlin started to intervene more, trying to manage everything centrally.

He also remembered the transformation of the warehouse itself: “We used to have a kitchen, where people could heat up some food, which was then packed with bicycles.” The material deterioration of the workplace also coincided with workforce re-composition (Popan and Jesnes, 2025). Workers who had experienced the early period as socially stimulating increasingly left, while precarious migrants and others with fewer labour-market alternatives remained. This process complicates the interpretation of “collective disenchantment”. Workers may remember the past positively because working conditions objectively deteriorated, because their own biographical circumstances evolved, because the composition of the workforce changed, or because memories of the exceptional pandemic period are quite unique. These explanations are not mutually exclusive. The literature on workforce re-composition is useful here because the composition of platform labour can change rapidly as workers enter and exit the sector (Popan and Jesnes, 2025). Such changes matter not only for the labour supply but also for the social bonds that workers create. A workplace community depends on continuity: when experienced workers leave and new workers enter, relationships, shared histories and informal norms must be rebuilt.

The interviews suggest that “disenchantment” was experienced collectively. Workers did not merely become individually dissatisfied. They noticed changes in how colleagues treated one another, how managers interacted with workers and how the physical workplace changed. The deterioration of the kitchen at Flink provides a useful symbol. A space initially used for eating and socialising became storage space for bicycles as the company expanded. At the same time, workers described more orders, more pressure and less autonomy. This does not mean that platforms simply destroyed social bonds through a deliberate strategy of isolation. The process was more gradual and indirect. Changes in business models, managerial control, workforce composition and material infrastructures altered the conditions under which relationships could be sustained. The analysis revealed that the hub as a place for building a sense of community was dependent on organisational arrangements that workers themselves did not control – and that workers shared the memory of this process and mutually reinforced one another through their collective remembrance.

The most dramatic “organisational change” that several interviewees experienced first-hand was Gorillas' gradual decline and eventual bankruptcy. This period was marked by shared doubts about the platform's future, deep disappointments, conflicts with superiors, sudden mass layoffs, a sense of desperation and a perceived lack of transparency (see Altenried, 2024). Some participants became quite emotional when speaking about this period, as visible in Emre's account:

Gorillas closing in Hamburg has a lot of emotional value for us. One of my friends out of the excitement built a baby bike with the Gorillas design individually, and we used to keep it in the warehouse. Now, he took it back. He called me (…) I know it's not a great example, but it's like losing a social army. I don't know, it was like a big student camp, that we were in together for a couple of years. We had a lot of connections, now going apart, and for some people we are really worried.

In the aftermath of Gorillas' bankruptcy, Emre tried to reach out to former colleagues to support those who he considered to be especially vulnerable. This included workers who had depended on this job for their visa applications and were now at risk of deportation. His attempts illustrate how some solidarity structures can survive even an organisational breakdown of a platform.

5 Discussion and conclusion

This article has examined a relational dimension of food delivery work that is often overshadowed by accounts of algorithmic control (Heiland, 2025) and collective resistance (Tassinari and Maccarrone, 2020). Drawing on 22 autobiographical narrative interviews with (former) riders in Hamburg, it has shown how platform workers develop social bonds in shared logistical spaces and how these relationships can subsequently weaken or disintegrate.

First, the article sheds light on the relationship between fragmentation and the emergence of we-groups. Platform work is indeed fragmented in legal, technological, organisational and spatial terms (Della Porta et al., 2022). Yet fragmentation does not eliminate social relations. Workers create social infrastructures through repeated encounters, informal advice, digital communication and shared spaces. The relevant question is therefore not whether platform workers are fragmented but how social bonds emerge within and across fragmented arrangements.

Second, the article demonstrates the ambivalence of social bonds. Cooperation, friendship and mutual assistance coexist with competition and boundary making. The same ethnic network can provide newcomers with access to employment while appearing exclusionary to other workers. The same gamified competition can create excitement and belonging while intensifying individual performance and stigmatising those who deviate from the norm (Woodcock and Johnson, 2018). The same informal workplace culture can feel friendly to some workers and unsafe or exclusionary to others.

Third, the analysis shows why these relationships need to be understood biographically. Workers who occupy the same occupational position may have very different reasons for doing the job and different levels of dependence on it. James's account of transnational family obligations demonstrates how platform labour in Germany can support social reproduction elsewhere. Other workers entered the sector temporarily during the pandemic or while studying. These differences affect how workers interpret workplace conflicts and what they have at stake.

Fourth, the article contributes a feminist perspective by showing that the platformisation of food delivery should be understood within the broader transformations of social reproduction (Strüver, 2024). The convenience provided to consumers depends on labour performed by a highly unequal workforce. Gendered relations within hubs shape who feels comfortable, respected and legitimate at work, while migration and class shape workers' positions within the labour process. Social bonds can provide resources for recognition and coping but can also reproduce gendered and racialised hierarchies (Kwan, 2022).

Finally, the analysis highlights the importance of organisational change and workforce re-composition (Popan and Jesnes, 2025). The deterioration of social relations cannot be reduced to workers becoming less sociable or to platforms intentionally preventing bonding. Rather, changes in managerial control, spatial infrastructure, business models and workforce composition alter the conditions under which a sense of community can persist. Collective disenchantment is therefore a relational and organisational process.

Methodologically, the study also demonstrates both the strengths and limitations of autobiographical narrative interviews. Their open structure made it possible to reconstruct long-term changes such as the emergence and gradual decline of warehouse communities. The interview format also allowed participants to reflect openly on their work experiences during and after the COVID-19 pandemic. These temporal changes were important to how workers evaluated their rider occupation in retrospect. At the same time, interactions were not observed directly. Accounts of other workers, particularly members of groups to which the narrator did not belong, therefore need to be understood in the context of a particular speaker's perspective and biographical circumstances. Ethnographic research could complement this approach by observing everyday interactions in hubs and warehouses directly (Timko and van Melik, 2021).

The article does not seek to replace conflict-centred accounts of platform workers' solidarity (Tassinari and Maccarrone, 2020). Rather, it suggests that organised resistance is embedded in wider social settings whose everyday dynamics deserve attention in their own right. Understanding how platform workers grow together and drift apart helps to illuminate the conditions under which solidarity becomes possible, fragile or exclusionary. It is therefore important to examine the significance of their biographical circumstances and experiences in this regard. The article also makes visible forms of agency and recognition that may remain hidden when solidarity is considered only when it culminates in collective action.