With or without U(nions)? Understanding the diversity of gig workers’ organizing practices in Italy and the UK

Since 2016, mobilizations of gig workers across European countries have become increasingly common within location-based services, such as food delivery. Despite remarkable similarities in workers’ mobilization dynamics, their organizational forms have varied considerably, ranging from self-organization, to work councils, to unionization through rank-and-file or longstanding unions. To start making sense of this diversity in organizing practices, we compare two cases of mobilization in the food delivery sector: in Italy, where workers have initially opted for self-organization, and in the UK, where they have organized through rank-and-file unions. Drawing on interview and observational data gathered between 2016 and 2018, we find that the diversity of organizational forms across the two cases derives from the interaction between agential and contextual factors, namely: the capabilities of rank-and-file unions and the political tradition of militant organizing of the environment within which gig workers are embedded. These findings contribute to the emerging debate on labour relations in the gig economy by showing the central role that factors external to the labour process and to the institutional context play in shaping the structuring of labour antagonism in a still lowly institutionalized sector characterized by transnationally homogenous challenges.


Introduction
The last decade has seen the rise of the so-called 'gig economy'. This refers to the emergence of companies whose business model is based on contracting out tasks or jobs through the use of platforms connecting demand and supply of labour. Gig work can be either tied to a physical location, as in the case of transport or delivery services, or be fully performed online (De Stefano, 2016). Because gig workers usually have few employment rights and experience atomization and pervasive monitoring, the gig economy can be characterized as the last frontier of precariousness (Tassinari and Maccarrone, 2020). Since 2016, episodes of labour conflict in the gig economy, especially within locationbased services, have become increasingly frequent across the world, signalling an increase uneasiness of gig workers towards their working conditions (cf. Cant, 2019;Tassinari and Maccarrone, 2020).
A notable feature of the first instances of gig workers' mobilization in some of the largest European countries, such as France, Germany, Italy and the UK, is that they have mostly been driven not by established unions, but rather by other actors, such as grassroots unions and self-organized groups of workers (Trappmann et al., 2020). Only in recent years established unions have started to 'catch up' and be effectively present in this new sector. Yet, whereas there is a growing Industrial relations (IR) literature on the gig economy, gig workers' novel and heterogenous organizing practices have been thus far relatively under-researched.
In this paper, we try to address this gap by comparing processes of gig workers mobilizations in the food delivery sector in Italy and the UK, with the aim of making sense of their diversity in organizational forms. These are highly salient cases, as since 2016, these countries have been the hotspots of the first mobilizations of food delivery gig workers in Europe-often known in the public debate as 'riders'. Despite similar dynamics of mobilization, different organizing practices have emerged in the two cases. In the UK, grassroots unions, such as the Independent Workers Union of Great Britain (IWGB) and the Industrial Workers of the World (IWW), have played an important role in sustaining the riders' struggles, starting successful processes of unionization. In Italy, conversely, workers' self-organization has been the prevalent mode of organizing in the first years of mobilization.
Throughout the paper, we argue that, while the inherent contradictions of the gig economy labour process help to understand why in both cases workers decided to mobilize (Tassinari and Maccarrone, 2020), the diversity of organizational forms can only be understood by looking at factors external to the labour process (Atzeni, 2016;Cini and Goldmann, 2020;Nowak, 2021;Rizzo and Atzeni, 2020). In particular, to explain the different organizing choices of gig workers in the UK and Italy, we focus on the interaction between agential and contextual factors namely: the capabilities of rank-and-file unions and the political tradition of militant organizing of the environment within which gig workers are embedded.
Through this analysis, we make a twofold contribution. First, we advance the emerging debate on IR in the gig economy by providing a context-specific explanation for the diversity in organizational forms which has characterized labour conflict in the gig economy thus far. Second, we contribute to the IR debate on the collective action of precarious workers with insights drawn from scholarship on social movements, highlighting the important role that specific features of the socio-political context not usually considered in the IR literature play in shaping dynamics of emerging labour militancy in sectors characterized by low degrees of institutionalization of IR.

Understanding variation in precarious workers' organizing practices: Insights from the extant literature
The rise of the gig economy has attracted scholarly attention, both for its implications for the future of work and as a site of labour conflict. Industrial relations and labour sociology scholars have thus far focused mostly on the characteristics of the gig economy labour process (Gandini, 2019;Veen et al., 2019) and their potential for igniting or thwarting worker mobilization (Cini and Goldmann, 2020;Tassinari and Maccarrone, 2020;Wood et al., 2019). Although some scholars have noted the heterogeneity of the organizational practices chosen by gig workers (Vandaele, 2018), there has been however relatively little reflection on the drivers behind gig workers' organizing practices.
In many cases of gig workers' mobilization across Europe, including our two cases of interest, the UK and Italy, longstanding unions have remained on the sidelines in the first years of riders' organizing efforts, regardless of their extant power resources and institutional context (Tassinari and Maccarrone, 2020). Rather, many mobilization processes have seen the intervention of different actors moving outside the framework of the 'traditional' organized labour movement. In contexts such as the UK and Germany, these have mostly comprised radical and often new rank-and-file unions (Vandaele, 2018). In contexts such as Italy and France, radical activists and self-organized solidarity groups have instead been most prominent (Borghi et al., 2021;Cini and Goldman, 2020).
While the IR literature on the gig economy has only recently started to explore the issue of gig workers' organizing practices and the role played by 'new' IR actors (cf. Borghi et al., 2021;Vandaele 2018), over the last decade, a growing strand of literature has contributed to the study of precarious workers' mobilizations practices in other sectors, adopting a bottom-up perspective (Atzeni, 2021) that positions workers and their relationship with the surrounding environment at the centre of the analysis, rather than taking characteristics of the institutional context as the starting point (Atzeni, 2016;Alberti and Però, 2018;Cini and Goldmann, 2020). This literature highlights both the role of nontraditional labour actors in supporting precarious workers' mobilization and the role of social and contextual factors which shape mobilization processes, such as the presence of strong community ties, the intervention of informal leaders and groups or the resilience of a radical political culture (Atzeni, 2021;Martin and Quick, 2020;Nowak, 2021).
'New' IR actors with a prominent role in precarious workers' mobilization include grassroots members-led unions, like 'indie' unions in the UK (Però, 2019), which have in recent years managed to organize precarious migrant workers in sectors such as cleaning and security; as well as precarious workers' self-organized groups, which have been active in many countries since the mid-2000s to offer representation to precarious workers outside of the framework of traditional unionismsuch as for example the Italian Camere del Lavoro Autonomo e Precario or the San Precario movement (Hyman and Gumbrell-McCormick, 2017). Similar groups have recently started to get involved also in the representation of platform workersas in the case of the French Collectif des Livreurs Autonomes (Borghi et al., 2021).
The literature highlights that these 'new' IR actors often possess capabilities and enact repertoires of action and discourses enabling them to offer more effective representation to precarious workers and gain their trust than is the case for large traditional unions (Borghi et al., 2021;Però, 2019). These include the adoption of oppositional discourses emphasizing precarious workers' identity as unrepresented constituencies (Borghi et al., 2021: 11) and placing attention on their specificities (Però, 2019: 907); the building of horizontal networks of collaboration with self-organized groups of workers and other activist and community groups, avoiding a logic of top-down incorporation in preexisting union structures (Borghi et al., 2021: 14); the adoption of participatory practices and the encouragement of members active involvement in negotiations, following a 'logic of membership' (Offe and Wiesenthal, 1985) in contrast with 'traditional' unions practices of bureaucratic representation; and the adoption of agile and speedy forms of industrial action (Però, 2019). The successes of 'new' IR actors in mobilizing precarious workers have also served, over time, as an example to established unions to innovate their practices vis-à-vis these constituencies, often fostering positive complementarities between 'new' and 'old' labour actors (Bondy, 2021;Hyman and Gumbrell-McCormick, 2017;Smith, 2021).
This body of literature also highlights that the connections of precarious workers with the social and political environment beyond the workplace are equally central in shaping their dynamics of mobilization. For instance, in his research on the UK 'indie' unions, Però (2019) stresses that the organizational linkage between these actors and the precarious workforce was made possible and reinforced thanks to pre-existing ties and previous engagement between the Latin American community, of which many of the interested workers were part, and the grassroots unions (Alberti and Però, 2018). The preexistence of such social and cultural linkages favoured the development of mutual trust between workers and unions over time, enabling the const ruction of a cohesive 'community of struggle' (Però, 2019). In Italy, the growth of rank-and-file unions in sectors traditionally characterized by low union density, such as services, logistics and meat processing, has been characterized by similar interaction dynamics, with the intervention of small groups of radical unionists able to gain the trust of a highly exploited north-African workforce through the long-term consolidation of community ties (Cini and Goldmann, 2020). Research on precarious workers' mobilizations in the global South, such as for example Rizzo and Atzeni (2020)'s study of precarious workers' bottom-up organizing in the transport and delivery sector of Buenos Aires and Dar es Salaam, also highlights how their mobilization practices were significantly influenced by local sociopolitical factors and distinct political legacies. Likewise, Ford and Honan (2019) show how the presence of community organizations operating at the urban level has been key in enabling the collective action of app-based transport drivers in Indonesia, facilitating workers' grassroots mobilization at the expense of traditional union organizing.
In parallel to this strand of IR literature, the literature on social movements is also helpful to further conceptualize the role of contextual factors external to the labour process and to the institutional context of IR in shaping the organizing forms adopted by precarious workers (Grote and Wagemann, 2018). Indeed, social movement scholarship has for long paid attention to features of the socio-political context, besides institutions, to make sense of social movement formation processesfocussing on factors such as the protest culture of a given country, the informal networks of activists of a given community, the presence of social movement organizations and the organizational traditions of mobilization (Della Porta and Diani, 2020).
For instance, extant research in the field of social movement studies on precarious workers' organizing in Italy from the early 2000sin particular on the movement of 'San Precario', a fictitious patron saint invented by activists and assumed as the political symbol of new 'non-standard' workershas highlighted how processes of collective identity formation among precarious workers evolved outside of the framework of traditional unions, and were rooted in pre-existing configurations of informal organizations and social movement actors, drawing on local collective repertoires of contention and political practices (Mattoni, 2012;Mattoni and Vogiatzoglou, 2014;Politi et al., 2021). These studies also highlighted how the specificities of the frames of collective identity formation adopted by the 'San Precario' movementrooted in the articulation of a new identity as a 'precarious class', historically neglected by traditional unionssignificantly shaped the organizational strategies that precarious workers adopted in their mobilization, which were geared towards spontaneous self-organization. These experiences, in turn, facilitated the emergence of other experiences of self-organisation within different sectors, such as call centres, publishing houses and airport services, which drew inspiration from the 'San Precario' repertoires of contention (Murgia and Selmi, 2012). Hence, these studies highlight that resources and opportunities for workers' organizingnot necessarily related to the features of work organization or to the specific institutional contextcan be developed and exploited by various types of informal actors in their mobilization efforts, also outside the framework of traditional unions; and that specific features of the local socio-political context and of the forms of collective identity that this gives rise to can in turn impact upon the organizational forms that precarious workers choose to adopt, often forging legacies in the local 'cultures of solidarity' (Fantasia, 1988) with lasting effects over time.
To sum up, integrating IR scholarship on precarious workers' organizing with insights from the social movement literature seems particularly helpful to make sense of the organizational practices of the Italian and British riders and of their variety. This integrated approach offers two advantages for the analysis of the mobilizations under investigation: first, it provides a more agential, worker-based and not union-centred, interpretation and second, it emphasizes the importance of the socio-cultural context of mobilization, beyond the workplace and the institutional landscape, in which the relation between workers and their surrounding socio-political environment is central. Building on these insights, we now provide a theoretical argument to account for the observed diversity in organizational forms among gig economy workers in Italy and the UK.
Understanding the variety of organizing practices in the gig economy Our argument starts from the observation that whilst the wide array of labour processcentred analyses of workplace dynamics in the gig economy (cf. Tassinari and Maccarrone, 2020;Veen et al., 2019) offer suitable explanations for why gig economy riders have been able to mobilize, they do not yet offer a satisfying interpretation of the different ways in which these workers have organized. In short, if the labour process in the gig economy is 'constant' across cases, it cannot account for the observed variations. By contrast, we argue that the attention that both the strands of IR scholarship reviewed above and the social movements literature devote to workers' agency and to their relationship with other non-formalized actors and the surrounding socio-political context of mobilization helps to better identify the social processes that lie behind riders' organizing practices and that account for the main differences between the Italian and British case.
More specifically, we contend that the different ways of organizing adopted by riders in the early phase of their mobilizations depends neither on features of the labour process, nor on the institutional characteristics of the employment relations context; but rather on factors external to the workplace and pertaining to the social and political environment in which the workers and their mobilization were embedded, and which shaped the resources and opportunities available to them, with an impact on their organizing practices (Cini and Goldmann, 2020;Nowak, 2021). In particular, we postulate that the combination of two factors, connected to the broader context of riders' mobilization, explains the differences in organizing. The first factor refers to specific capabilities of the grassroots unions active in the respective localities. The second refers to the political traditions of militant organizing present in the contexts of mobilization. Let us analyse these in turn.
The first factor which we identify as relevant are the 'capabilities' (Levesque and Murray, 2010) of the local rank-and-file unions, and specifically what we call their relevant 'sectoral expertise'. By capabilities, we mean 'sets of aptitudes, competencies, abilities, social skills or know-how that can be developed, transmitted and learned' (Levesque and Murray, 2010: 336), that allow actors to exercise power and act effectively in specific contexts. As Gumbrell-McCormick and Hyman (2019, p. 101) emphasize, to be effective unions must be able to mobilize the collective action of their targeted constituency, building on workers' 'willingness to act' (Offe and Wiesenthal, 1985), and channel this through appropriate strategies. This presupposes the construction of specific expertise able to involve in action the targeted constituency. Whilst in the cases of emerging riders' mobilization under analysis the workers' 'willingness to act' was present across the board, rooted in social processes of solidarity formation with their origins in the labour process (cf. Tassinari and Maccarrone, 2020), an important difference were the capabilities that enabled rank-and-file unions to intervene effectively vis-à-vis these constituencies and to operate in the specific gig economy context. Indeed, whilst rank-and-file unions generally share many features in their ideological and organizational outlook, they often differ in their sectoral origin and scope of action. We argue that this difference in the sectoral origin and scope of action shapes rank-and-file unions' capabilities for effective intervention in the gig economy context. In particular, we contend that grassroots unions with a recent history of organizing in sectoral contexts similar to that of 'location based' gig economy platformsthat is, the urban service sector economysuch as for example the British ones are better placed to establish organizing relationships with gig workers than unions which lack similar expertise, in virtue of the relevant action repertoires that they are able to deploy. First, prior experience of operating in the urban service sector shapes unions' capacity to interpret the specificities of the urban environment where location-specific gig economy activities are embedded, and thus to adapt their 'repertoires of contention' (Della Porta and Diani, 2020) so as to achieve greater effectiveness in their organizing and mobilizing tactics. Second, prior experiences of organizing in the urban service sector furnishes unions with a nuanced understanding of the composition and attitudes of the urban service workforceusually characterized by a complex mosaic of young people and students with fragmented work lives and marginal attachments to the labour market, and of migrant workers in positions of greater vulnerability and dependency towards precarious work (Tassinari and Maccarrone, 2020). These established understandings increase unions' capabilities to intermediate between the demands of different constituencies and to adapt their operating modes to the inclinations of the workforce. Third, prior sectoral expertise and activity in similar segments of the urban service economy equips unions with a track record of past successes that can increase the responsiveness, trust and confidence towards these actors from the workers themselves.
The second explanatory factor we identify refers to the 'political tradition of militant organizing' present in the context of mobilization. Here we draw on the concept of 'popular and political traditions of mobilization' developed by Nowak, 2021. In his view, these traditions go beyond a focus on the workplace, and while they are at times connected to certain industries, they rather form part of the larger community, at both the local and national level. For us, such organizational traditions encompass both the political activists and their backgrounds, their experiences of activism and their social networks, as well as the 'social movement infrastructures' (e.g. collectives, social spaces, associations, etc.) embedded in the respective context of mobilization (Mudu, 2004). Legacies and past experiences of militancy may provide workers with ideological frames and established scripts for action to rely on, shaping their attitudes towards the different organizational forms that they adopt.
We retain that the interaction between the gig workers and these features of their surrounding environment shapes their ways of organizing. Specifically, we argue that the presence in given localities of networks of activists with past collective experiences of workers' self-organization and of social movement infrastructures supporting mobilization efforts can equip militant workers with political skills and organizational capacities that makes them more inclined and able to pursue forms of autonomous organizing. This was, for example, the case in the first Italian cities where mobilization took place: Bologna, Milan and Turin. Conversely, the presence of political legacies of successful rank-and-file organizing of precarious workers through the framework of grassroots unions, such as in the cities of London and Brighton, can equip core militant workers with political skills, ideological inclinations and social networks geared towards the establishment of relations with grassroots unions to support the emerging mobilizations in the gig economy. In short, we claim that the different political experiences and organizational capacities of activists, arising from the different legacies of the militant milieu, contribute to the different organizational forms through which the couriers' mobilization became structured.

Case selection, methodology and data
To develop our argument, we adopt a comparative qualitative case study research design (Yin, 2013), focussing on salient cases of mobilization of gig workers: the case of the riders of UK-based food delivery platform Deliveroo in two British cities (London and Brighton), and the case of riders working for the platforms of Foodora, Deliveroo and Glovo in three Italian cities (Turin, Milan and Bologna). The logic of case selection was exploratory: we focused on salient cases of workers' collective action in a new segment of the gig economy that featured similar dynamics of mobilization, but displayed differences in the organizing forms adopted. Due to this exploratory intention, we did not strictly follow a 'most similar' or 'most different' comparative research design. We rather considered how the two cases share important commonalities in some dimensions, whilst differing in others; and looked at the roles of such similarities and differences in shaping the case histories.
With regard to similarities, first, the features of the labour process in the platformmediated food delivery are remarkably alike (Tassinari and Maccarrone, 2020): across all locations, couriers operate in a spatially dispersed workplace (the city itself) and their primary interaction is with the app through which they receive the assigned deliveries; to maximize work flexibility, workers are hired as independent contractors and finally, the work process is monitored through various form of algorithmic management (Gandini, 2019). Second, the cases share a similar workforce compositionwith the majority of workers being male, young, and segmented along national-ethnicity lines. Third, the cases show some similarities in the 'voluntarist' framework of IR governance, characterized by comparatively high levels of labour conflict, union pluralism and, most importantly, coexistence (or fragmentation) between mainstream 'confederal' unions and grassroots unions. In both countries, the growth of precarious work over the last two decades has incentivized the growth of new labour actorsboth grassroots unions or self-organized labour collectives. Fourth, in terms of socio-political context, all the cases of mobilization that we focus on -London, Brighton, Milan, Turin and Bolognaare examples of urban areas with a rich tradition of social movements, labour activism and political infrastructures. Holding constant these similarities enabled us to hone in, in the analysis, on the varying factors which might explain the observed differences in organizational forms.
With regard to noteworthy differences, first, the legacies of mobilization of precarious workers in the two countries are diverse. In Italy, these have longer-term roots, and have been characterized by a co-existence between rank-and-file unions and workers' self-organized collectives (Mattoni and Vogiatzoglou, 2014). In the UK, conversely, independent unions have more recent origins, dating back at most to the mid-late 2000s, and traditions of workers' self-organization are less developed. Second, the sectoral anchoring of rank-and-file unions differs. Whilst in the UK, rank-and-file unions were mostly born in the private services sector, such as the outsourced cleaning services, with some recent penetration also in the delivery industry (Alberti and Però, 2018); in Italy, they were instead historically established and geared to the public sector, with a recent but very significant intervention in the logistics sector (Cini and Goldmann, 2020).
To build our theoretical argument, we followed an abductive process (Glynos and Howarth, 2007) which entails an iterative movement between data and theorystarting from the empirics to develop theoretical constructs that best explain the relationships identified in the data, building upon existing theories, then verifying how these constructs perform in making sense of the phenomena and revising them iteratively. We draw on indepth interviews, participant observation and document analysis deployed in fieldwork conducted by the authors in Italy and the UK between October 2016 and July 2018. We conducted thematic analysis of over 300 press releases and other public communications collected from the Facebook and Twitter pages of the riders' groups; we participated as observers in various riders' meetings and protest events and carried out thirty-one semistructured interviews with couriers, activists and trade unionists involved in the mobilizations (see Appendix). These data were triangulated and coded thematically to reconstruct detailed case histories of how the mobilizations unfolded, and to situate these in the local histories of workers mobilization; to identify the specificities of the repertoires of contention and organizing strategies adopted by riders' groups and grassroots unions in the different cities and to analyse the attitudes of the mobilizing workers towards different organizing forms. All data have been anonymized to preserve the identity of the interviewees.

UK
Since 2014, food-delivery multinational Deliveroo has expanded its London fleet to comprise several hundreds of riders, hired as independent contractors. The first protest against Deliveroo took place in August 2016, when dozens of workers gathered outside the company's main office in Central London to protest the decision of the company to move from an hourly pay system to piecework payments. This turned into the first 'unofficial' strike of Deliveroo workers in the UK: the workers logged out from the app, effectively causing a work stoppage, and set up a kind of 'picket line' outside of the head offices (Cant, 2019).
The first contact of Deliveroo workers with trade unions took place at this point. The first union to offer its support was the rank-and-file union IWGB. Created in 2012, IWGB had already been active for 2 years in organizing workers in the courier industry in London, as well as other low wage service sector workers (e.g. in cleaning and security). When protests at Deliveroo started, IWGB organizers went down to offer workers their help in coordinating the formulation of demands and negotiating with management, to avoid victimization of individual workers. According to one of the IWGB organizers interviewed, UK2, the union managed to gain the trust of the workers at this point as they could present their track record of successfully organizing couriers in London and winning significant pay concessions from other companies.
The workers waged a 6-days long wildcat strike action. The support offered by the IWGB was crucial at this stage: as well as helping with the formulation of demands, the union organized a crowdfunding to support the striking workers which, in a few days, raised £20,000 and helped to facilitate the sustainability of the protest. The level of public attention and support received by the mobilization was very high. Eventually, Deliveroo agreed not to impose unilaterally the new piecework payment system on all the riders, but to start an opt-in trial in some areas. This partial victory demonstrated that collective action could bear fruits, giving the impetus for a unionization drive.
Following this first episode, a sizable group of riders joined the already active couriers' branch of the IWGB. Since then, the union has started a longer-term campaign of unionization of Deliveroo riders, active across several cities in the UK. This has entailed ongoing recruitment efforts and two guiding demands: a guaranteed hourly living wage for all and the re-classification of all riders as workers rather than self-employed. Further campaigns have also been launched around health and safety, insurance coverage, sick pay and, most recently, the provision of income support and adequate protections during the Covid-19 pandemic. The organizing methods pursued by the IWGB effectively combine three strands of complementary tactics: grassroots organizing of riders through recruitment stalls in the city, flyering and public mass meetings; mobilization of public opinion through social media and the pursuit of high-profile media coverage to maintain attention on the issue; and legal action to pursue formal union recognition and contest the riders' self-employed status through legal avenues.
The success of the IWGB's tactics in raising the profile of the first Deliveroo dispute and sticking by workers at risk of victimization encouraged over time other riders to join the union and the protests (interview, UK8). The protests of the Deliveroo riders in London have had significant ripple effects. First, Deliveroo riders' mobilization and selforganization have since sprung up in other UK cities, such as Bristol, Brighton and Leeds, where workers have held protests and raised similar demands towards the company. In all these cities, the riders have been supported either by the IWGB Couriers' branch (as in Brighton) or by another anarcho-syndicalist union, the IWW (as in Leeds and Bristol), which operates with similar tactics.

Italy
In the Italian context, the first riders' mobilizations had their epicentres in the cities of Turin, Milan and Bologna. The first company targeted was the German food delivery multinational Foodora, which in October 2016 became the object of the protest staged by a group of approximately 50 couriers in Turin. Like in London, the explosion of workers' protests was triggered by a change from an hourly rate to a payment-by-delivery system. The terms of the employment contract (categorization as parasubordinate contractors rather than employees) and the intrusive punitive practices adopted by the firm against perceived 'troublemakers' were also central to the riders' demands.
As in the UK, the first mobilization was organized as a 'proto-strike', combining 'unlogging' from the app with mass pickets and an online campaign on social media. These protest methods fell mostly outside the traditional repertoire of trade unions' mobilization tactics, resembling instead some of the practices adopted by the Italian precarious workers' movements in the early 2000s. The riders' mobilization soon spread also to other cities, such as Milan and Bologna, extending also to other food delivery platforms such as Deliveroo, Glovo, JustEat, as well as local platforms.
In the first years of mobilization, the workers involved in the protests consistently chose to organize themselves through self-organized collectives or what they called 'informal unions'for example, Deliverance in Milan, Deliverance Project in Turin, and Riders Union in Bologna (cf. Caruso et al., 2019;Marrone and Finotto, 2019). Interviewed riders involved in the Turin, Bologna and Milan mobilizations highlighted that established trade unions played a very limited role in organizing the riders, and that they had largely relied on their experiences of prior mobilization acquired through involvement in other social movements (Cini and Goldmann, 2020). To the extent that relationships between the riders' collectives and established unions have taken place in these cities, they have been either of a purely informal character or of external support. Established (such as UILTuCS in Milan or FILT-CGIL in Bologna) or rank-and-file unions (e.g. ADL-COBAS in Bologna or SI-COBAS in Turin) offered practical help and amplified the riders' demands through their own public and institutional channels, without however initiating their own processes of unionization. Only over the last 2 years unions have managed to organize workers successfully in some Italian cities, although selfestablished collectives have continued to operate and expand.

Exploring differences in organizing forms
As the two case summaries have shown, the difference in the organizing models adopted by the riders in the UK and in Italy in the first years of mobilization is significant. In the British case, workers accepted the organizational support and expertise of rank-and-file unions, whilst in the Italian case they preferred to organize in informal collectives. We attribute this difference to two factors, which we identified from our abductive analysis as especially relevant, one relating to the features of the actors involved, and one to the features of the socio-political context: the capabilities of intervening unions arising from their relevant sectoral expertise, and the political tradition of militant organizing. We now show how these two factors interacted in the cases.

The relevant sectoral expertise of the intervening unions
Whereas in the UK, grassroots unions played a pivotal role in the couriers' mobilization, in Italy, similar unions did not. We argue that the capabilities of grassroots unions arising from their 'original' sectoral scope of intervention shaped their diverse involvement in the couriers' mobilizations and the relation established with the workers.
In the UK, grassroots unions, such as the IWGB or IWW, had their 'traditional' scope of intervention and extensive prior experiences of organizing in sectors with features similar to the gig economythat is, other private service industries in the urban context. Rank-and-file unions, such as the IWGB, were indeed born in the urban private services sector, more specifically in the outsourced cleaning and security services, and had more recently expanded in the courier and delivery/bike messenger sector even before the arrival of the gig economy (Alberti and Però, 2018). This experience endowed them with specific capabilities (Levesque and Murray, 2010) and skills that enabled them to intervene effectively also in the location-based gig economy. In Italy, conversely, grassroots unions were instead historically established and geared to the public sector, with a recent but very significant intervention in the peri-urban logistics sector (e.g. warehouses and distribution centers) (Cini and Goldmann, 2020).
The IWGB's prior experiences in orchestrating disputes in the urban service sector helped in various ways. First, the experience of organizing in sectorssuch as bike deliverycharacterized by physical dispersal of the workforce and 'atypical' employment arrangements akin to the location-based gig economy gave the IWGB a good understanding of the approaches and tactics that would be most effective in a sector characterized by high informality, low institutionalization of voice channels and physical dispersal through multiple urban locations. Indeed, the IWGB developed a distinctive capacity of combining grassroots mobilization, often moving outside the confines of the law through unofficial strike action and a combination of physical and digital pickets that increased workers visibility in the urban space, with high-profile legal challenges waged through the employment tribunal system, so as to use creatively the channels of labour law and collective bargaining to its favour. Second, the IWGB managed to adapt its repertoire of action to the features of the gig workforce, made up mostly of workerseither young students or migrantwithout prior unionization experiences. For example, a Deliveroo rider involved in the London dispute (UK8) recalled how one of his colleagues leading the unionization efforts was successful in establishing a relationship between the couriers and the union, as he was able to frame this to his colleagues in an accessible and relatable way: "I had never been in a union and I was leftist in like, broad strokes, but I really hadn't done any of this stuff before [...]. He [the IWGB organizer] was like, slightly older but still kind of my age, but much more clued up, and he didn't talk in unionist language at all. […] he was just like -'hey this really sucks, we're meeting in this park in a week's time, you know, to hang out…" Through an effective use of social media, the union was also able to promote a successful campaign to support Deliveroo's workers mobilization efforts and managed to frame the couriers' struggle as part of a broader agenda of mobilization against the gig economy, connecting the riders' mobilization with analogous disputes involving other firms belonging to the gig economy and to the transport sector in general, such as Uber, Addison Lee and City Sprint.
Third, a credible track record of victories helped to increase its credibility and legitimacy vis-à-vis the riders. The union's successes in getting pay increases for couriers in other London-based firms was, according to interviewees, an important factor that increased its credibility to the eyes of the couriers in the initial stages of the dispute (interview, UK5).
These capabilities were less present in the Italian case, where rank-and-file unions, such as SI-Cobas or USB, in recent years had instead been successfully organizing workers in the peri-urban logistic sector, within companies such as Amazon, TNT or DHL. While gig economy food delivery platforms also belong to the logistic sector at large, there are considerable differences with respect to companies such as Amazon, including the absence of common warehouses and the dispersed urban workplace location. Italian rank-and-file unions, endowed with their experience in 'traditional' logistics, could not readily 'translate' the repertoires of struggle employed theresuch as warehouse occupations and transport blockadesto food delivery platforms. This led to some mismatch between the tactics and repertoires of action that grassroots unions were used to employ, and those that gig workers found consonant with their own vision and interpretation of the dispute.
For example, in September 2016, as the dispute in Turin escalated, the riders sought the support of the SI-COBAS, already active in the mobilization of workers in the traditional logistics sector. Through the union, they presented their demands formally to the firm. Yet, soon after the first strike of October 2016, disagreements over the next steps of the struggle emerged. The views of the union and the workers diverged over the union proposal to block some restaurants in the Foodora circuit, mutated from other struggles in the logistic sector (Interview, IT1 More generally, no grassroots union in the Italian case was readily able to adapt its repertoire of action to a context where, since the beginning of labour unrest, 'digital' ways of struggle acquired importance. Commenting on the difficulties encountered in the relation with grass-root unions in Turin, one rider (IT2) commented: 'We speak about apps, facebook, shitstorming and mailbombing, and then they say "yes let's go there and block the restaurants" (…) strategically it seems a bit anachronistic to me'. In this sense, and unlike the British grassroots unions, in the first years of mobilization no Italian counterparts were able to adapt their intervention strategies in a way that was appropriate for the gig economy context. The differing capabilities of grassroots unions also impacted on the confidence that the couriers themselves displayed towards these actors in the nascent phase of their mobilizationhigh in the UK (Cant, 2019), ambivalent in Italy, where gig workers came to perceive these unions as not fully able to understand their needs and action tactics (Caruso et al., 2019).

The political tradition of militant organizing
The second relevant factor accounting for the different organizational modalities across the cases is the political tradition of militant organizing. In Italy, the presence of a protest culture based on self-organization favoured the emergence of informal political collectives among gig workers. Such tradition of self-organization was absent in the UK, where rank-and-file unions (Cant, 2019;Però, 2019) played a much more prominent role.
In Italy, self-organization and direct action had been strategies of precarious workers' mobilization since the early 2000s (Murgia and Selmi, 2012). While self-organization did not necessarily exclude cooperation with confederal and/or grassroots unions, these groups maintained their own autonomous identity and organizational practices. Experiences of precarious workers' self-organization were often interconnected with the tradition of mutualism, well embedded within Italian social movements, receiving infrastructural and material support from social centers and occupied spaces (Bosi and Zamponi, 2019). Concretely, legacies of self-organization in various Italian cities were evident in the continued presence of local activist networks with past involvement in precarious workers organization efforts and of self-help infrastructures such as social centers, squatted houses and militant associations (cf. Mudu, 2004).
Practices of self-organization supported by Italian social movements resurfaced within and contributed to shaping the struggle of food couriers in Italy (Cini and Goldmann, 2020). In Turin, Milan and Bologna, gig workers set up self-organized collectives, which benefited from the material support of social centers and occupied spaces to hold meetings, promote fundraising and organize mutualistic activities like bike repair clinics (Caruso et al., 2019;Marrone and Finotto, 2019).
The connections between the riders and the militant context of the respective cities were instrumental in bringing about these self-organizing practices. The gig workforce mobilized in Turin, Bologna and Milan comprised, at least at the beginning, a significant component of students of native origins (Cini and Goldmann, 2020). Many of these workers had a relatively high political expertise and previous experiences in various social movements, ranging from student protests to anti-austerity mobilizations and selforganized collectives and social centres (cf. Interviews with IT3, IT11, IT16, IT17 and IT20).
This tradition of self-organized activism shaped the mobilization practices of Italian gig workers in two main ways. First, past and current experiences of self-organization contributed to shaping the preferences, ideological inclinations and 'organizing cultures' of the key activists and workers involved in the early stages of the mobilization. Concrete involvement in autonomous social movement spaces made them more inclined to reproduce similar models and repertoires of action in their emerging labour mobilization. For example, although none of the interviewed activists in the various Italian cities expressed a diffidence towards unionization per se, their prior direct involvement in, or appreciation of positive past legacies of, experiences of self-organization made them value highly the preservation of political and organizational autonomy in their mobilization practices.
The interviews highlighted this as an important guiding principle. Commenting on the decision to not engage with established unions, one rider from Turin (IT3) noted that '…otherwise, the struggle would have ended in issuing union membership cards and dealing with bureaucrats'. Key organizers in Milan (cf. IT11, Deliverance Milano) and in  expressed similar opinions about the importance of self-organization as a lesson learnt from past experiences. These past legacies also shaped workers' repertoires of action, as they repurposed scripts that had proven successful in the past. For example, some of the mobilization practices adopted by the Italian riders' collectives, such as holding 'critical mass' demonstrations on their bikes, had traditionally been part of the toolbox of Italian social movements (Bosi and Zamponi, 2019).
Second, these experiences of militancy provided committed workers with organizational skills and resources which helped their collectives to take shape, as they were a sort of 'training ground' where they learnet how to organize protests, assemblies, lead demonstrations and rallies. Various riders (IT21 and IT22) pointed out how such experiences provided them with self-confidence in organizing collective action. Put otherwise, their prior experiences helped these workers take up the role of informal leaders in the riders' mobilization.
Differently from Italy, in the UK all major prior experiences of precarious workers organizing in the urban service sector had previously been articulated through grassroots unions, whilst no substantive experiences of precarious workers self-organization of the Italian kind had taken place. In particular, since the mid-2000s both mainstream unions such as Unite and the GMB on the one hand, and rank-and-file union, such as the IWW, on the other, had engaged in deliberate strategies of organizing precarious and migrant workers, drawing on tactics coming from the tradition of social movement unionism and seeking to replicate the success of similar US-based initiatives (Cant, 2019). These experiencessuch as, for example, the long-running organizing campaigns of cleaning and security migrant workers in universities, hospitals and large service sector firms across Londonenabled the formation of a milieu of activists and organizers, many of whom former student movement activists, that had accumulated in previous years expertise or familiarity with practices of organizing precarious workers through the framework of rank-and-file grassroots unions, such as the IWGB or the IWW.
When the more politicized share of the Deliveroo workforceseveral of whom were or had previously been university studentsstarted trying to coordinate organizing tactics after the initial outbursts of 'spontaneous' discontent, they therefore drew on their networks and prior experiences in the universe of London grassroots unionism. This facilitated a fruitful encounter between gig workers and these segments of the emerging gig workers' movement. For example, UK4, a key organizer in the Brighton Deliveroo dispute, had had substantive prior involvement in the UK student movement and in supporting the organizing campaigns undertaken by the IWGB/IWW among university cleaners in London. These networks and experiences proved crucial in forging relationships with the IWGB. In his words.
"[at the beginning of the dispute] I raised the idea that we should have a union meeting and everyone was like 'Yeah, that sounds like a good idea', so I set it up and got in touch with the IWGB, asked them if they could come down, can we join, can you help us unionise." Prior experiences of organizing labour disputes were also instrumental in the London context, where UK8, a Deliveroo rider, recalled the role played by skilled activists in forging links with the union and organizing one of the first strikes in South London: "There were some guys who had done their homework and they knew how you carry out an action. and they got in touch with lots of different branches, including the IWGB [...] The people who led it (…) they (…) had interacted with these ideas a little bit more and had a sensibility of what an action was and why, how it worked." Overall, the specific tradition of militancy of the two contexts impacted on the different ways in which Italian and British riders organized the mobilization. While in Italy practices of self-organization were pursued by drawing on past legacies of the autonomous precarious workers movement, in the UK organizing through grassroots unions was privileged as activists drew on their prior experiences of setting up 'indie unions' in the urban service sector.

Discussion and conclusions
In this paper, we aimed to shed light on a topic which is still relatively unexplored in the IR literature: the diversity of workers' organizing practices within the so-called 'gig economy'.
To do so, we analysed the struggles of food delivery couriers in the UK and Italytwo of the first instances of mobilization in the European gig economy. Across both countries, these mobilizations were driven by 'new' IR actors, namely grassroots trade unions and self-organized workers' collectives. Our goal was to understand why in the British context gig workers have established solid relationships with rank-and-file unions, whilst in the Italian context they opted for self-organization. While the similar labour process dynamics explain workers' 'willingness to act' and capacity to mobilize in both cases, they do not explain their organizational differences. Through the comparative abductive analysis of these two cases, we identified two factors pertaining to the socio-political context where the mobilizations unfolded, and of the network of actors embedded therein, that, we argue, accounted for workers' choices of different organizational forms.
First, we have shown that unions' capabilitiesand especially the prior sectoral expertise of grassroots unions active in the context of mobilizationmatter in shaping the relationships that they are able to establish with location-based gig workers. The fact that the gig economy is still a relatively new and unregulated segment of the economy may favour union organizations, such as grassroots unions, which are used to operating outside established channels of workers' representation. However, our findings underscore the importance of considering, rather than features of rank-and-file unions in general, their specific capabilities arising from their prior relevant sector of expertise as determinants of their ability to intervene effectively in new sectors and contexts. Second, we found that the tradition of militancy where the mobilizations unfolded influenced riders' organizing practices, by equipping the workers with specific ideological orientations, resources and skills to support their mobilization efforts. Our study highlights the importance of the political features of the social environment where workers are embedded as a crucial determinant of their organizing practices (Nowak, 2021).
By accounting for the diversity in the organizational forms chosen by gig workers in Italy and the UK, our findings advance the emerging debate on IR in the gig economy, so far predominantly focused on the processes underpinning the emergence of gig workers mobilization at the workplace level. In this respect, our study shows that these mobilizations are heavily influenced by the socio-political context in which they are located. In other words, if the challenges arising from food delivery platforms are sector-specific and homogenous at the labour process level, they are mediated by specific socio-political dynamics of interaction, resulting in a variegated picture in terms of organizational practices of mobilization across contexts. By placing workers and the political context in which they are embedded at the centre of analysis, we showed how these factors shaped the dynamics of labour conflict in a sector characterized by low institutionalization of IR.
The main caveat of our research concerns the generalizability of our theoretical insights to other cases of gig workers mobilizations. Across Europe, the organizing practices and representative forms adopted by gig workers in the first phase of their mobilization have been characterized by great heterogeneity, both across and within countries. Self-organization has been predominant in Italy and France; rank-and-file unions have been the most common form of organizing in the UK, Germany and Spain, whilst in Denmark, Netherlands and Sweden longstanding confederations have been playing a more central role. Moreover, the organizing dynamics are by no means static. In Italy, since 2020 both established and grassroots unions have undergone a process of learning, adapting and updating of their strategies and have accordingly managed to start recruiting and organizing gig workers. Self-organized collectives have also continued to operate, and some forms of cooperation between these actors have emerged, with the creation of a national network comprising established unions and workers' collectives, although tensions have remained. This point towards a possible conflictual complementarity (Bondy, 2021) within the gig economy, still to be explored.
Moreover, whereas institutional factors did not play a prominent role in shaping the dynamics of organizing in the cases analysed here, they might nonetheless be relevant in accounting for broader patterns of variation across a larger number of countries. Institutional differences in IR settings may also become more relevant over time, as the gig economy proceeds from being a new sector relatively disembedded from the IR framework to a more regulated setting, as recent developments in Italy might suggest. Questions of internal union democracy and structures of decision-making might also play a role in accounting for how the relationships between gig workers and unions of different kinds evolve over time. Further research across a larger number of cases will be needed to develop these insights. At the same time, the observed heterogeneity of organizing practices confirms the value of adopting a perspective on precarious workers' mobilization which is attentive to the role of new IR actors and non-union forms of organization.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
interviewing workers and activists who took part in the mobilizations. To contact and involve these workers in our investigation, we adopted the technique of 'snowballing sampling', that is, to contact and interview workers involved in the mobilizations whose contact was given to us by their colleagues (usually key 'gatekeepers') whom we had previously contacted and interviewed.
For the riders in Italy, we first contacted through specific mailing lists, websites and Facebook pages various activists of the social centres, where we knew the riders regularly met to organize collective actions. These activists played for us the role of 'gatekeepers'. They put us in contact with various riders who were part of the mobilizations. Once we interviewed these riders, the latter helped us to involve other colleagues in our investigation. As for the riders in the UK, we used informal contacts with riders and activists in London and Brighton, some of whom were active in the IWGB, who acted as gatekeepers and in turn put us in touch with other riders and organizers.
Although snowballing sampling exhibits some limitations, among which the inclination of reproducing a self-selecting dynamic (interviewing only the social and amical network of the first respondents), we believe that this technique was most appropriate for conducting interviews in our study, as our main research purpose was precisely to explore the motivations of those who were involved in the mobilizations.
This choice has important implications on the socio-demographic aspects of both of our samples, which show the following peculiar characteristics. As for the riders' samples in Italy and the UK, all respondents among the workers were young (aged between 21 and 36) and, with the only exception of one rider in Turin and one rider in London, male and of respectively Italian or British origin. Most riders in Italy were either university students or young people at their first work experience, whilst in the UK, most interviewees had had other casual jobs in the past. The specific nature of the riders' jobs (low pay, flexible contract, part-time and physically demanding activity) may incentivize the entry of a significant component of young people or students, who are normally the segment of population with the most politically militant attitudes.
In Table A1 and Table A2 we list the respondents interviewed for this research and the location and time of the interview, and the details of the events where participant observation was conducted by the authors.