Skip to navigation – Site map

HomeThemed issues21IssueMeal delivery logistics and the a...


Meal delivery logistics and the agencies of distribution in urban economies of food provision in the UK

Lizzie Richardson


Newcastle-upon-Tyne has recently seen an expansion in takeaway prepared food thanks to the meal delivery service provided through companies such as Deliveroo. These companies have what they call a “technology platform” that enables customers in certain urban areas to order and purchase a meal which is delivered to their location, usually within thirty minutes. Although a seemingly mundane moment of consumption, the processes underpinning this act compose a distinctly urban manifestation of the broader logistical logics of contemporary capitalism. Delivering a prepared meal from a restaurant to its nearby site of ingestion in the city is symptomatic of the role of logistics in valuation, through achieving the right place at the right time for a given commodity. Such logistics require the emergence of new urban infrastructure and the reconfiguration of existing components through the combination of the technology platform, the restaurant collection point and rider-deliverybox-vehicle assemblage. Building from this infrastructure, Deliveroo are able to extend their operations into food production and supply, whilst also re-organising socio-spatial divisions of labour in food provisioning. Distribution is therefore no neutral mediator between production and consumption, but rather has far-reaching effects, altering existing and generating new spatial practices composing urban economies of food provision.

Top of page

Full text

Lizzie would like to thank Elke Beyer and Anke Hagemann for their work editing this special issue and organising the AAG 2018 session from which it emerged. She is also grateful to the reviewers and editor for their helpful comments and would like to acknowledge that small elements of this article appeared in “Platforms, markets and contingent calculation: the flexible arrangement of the delivered meal” published in Antipode. The research underpinning this article was supported by a Leverhulme Early Career Fellowship.

1The city of Newcastle-upon-Tyne, like other UK cities, has recently seen an expansion in takeaway prepared food choice thanks to the meal delivery service provided through companies such as Deliveroo and UberEats. These companies have what they call a “technology platform” that enables customers in certain urban areas to order and purchase a meal which is then delivered to their location, usually within thirty minutes. Although a seemingly mundane moment of consumption, the processes underpinning this act compose a distinctly urban manifestation of the broader logistical logics of contemporary capitalism (Lecavalier 2016). Delivering a prepared meal from a restaurant to its nearby site of ingestion in the city is symptomatic of the role of logistics in valuation, through achieving the right place at the right time for a given commodity. It is such contingent valuation of commodities through their qualities and quantities of movement that has been the focus of interest at sites such as ports and warehouses where the myth of often automated and seamless logistical flows can be unpacked through examination of site-specific processes of friction (Gregson et al 2017; Stennmans 2019). In this “blending of production with elements of circulation and exchange” (Neilson 2014, 87), cities are not so much seen as an end-point for commodities, but rather raise their own questions concerning how movement and thus valuation can be achieved through processes of assembly, storage and (re)distribution in sometimes densely populated urban areas (Cidell 2011; 2012). Meal delivery creates a particular logistical challenge in this regard where the perishability of food requires rapid and accurate charting of urban space.

2Deliveroo began its meal delivery operations in Newcastle in 2015, shortly after its establishment in 2013 in London. While the delivery of prepared meals in Newcastle certainly pre-existed Deliveroo, the company’s operation in the city has nonetheless enabled “on-demand” convenience by offering customers and restaurants not only an online ordering and payment mechanism but also a delivery service to the customer via a fleet of delivery workers. Such logistics of meal delivery indicate the significance of mechanisms of distribution for urban economies of food provision, processes that have historically been underplayed in discussions of urban food governance in the UK (Davies and Evans 2019). As in other renderings of economy (Gibson-Graham 2006; Ferguson 2015), distribution and associated socially reproductive activities like food sharing and recycling have for the most part played second fiddle to questions of food production (or supply) in these policies (Sonnino 2019). Whilst interest in distribution is growing, for example through sustainable urban food initiatives including “community” orientated food sharing practices via the use of ICTs (Davies et al 2017), the powerful role of logistics actors in the valuation of food as a commodity requires further attention. Far from any alternative to commodity exchange, this article suggests that it is precisely the commodification processes involved in distribution, occurring here through the logistics of pairing a meal with its delivery, that are important for understanding how urban food economies in the UK are being (re)shaped.

3To examine this, the first section sets out how the commodity form of the delivered meal is co-produced with urban space, involving the emergence of new urban infrastructure together with the reconfiguration of existing ones. Considering three infrastructural components: the technology platform; the rider-deliverybox-vehicle assemblage and the restaurant collection point; the section draws upon auto-ethnographic research in Newcastle in 2017-18 involving work as a delivery rider through Deliveroo in order to better understand the urban operations of meal delivery. The second substantive section of the article analyses how the incorporation of urban space into the commodity form through this infrastructure results in changes to urban economies of food provision. It suggests that the placing of Deliveroo’s operations in the sphere of distribution for the realisation of personal convenience has left the company outside the purview of urban food governance strategies. At the same time, Deliveroo’s value extraction from the coordination of such distribution is enabling the company to extend their operations into other economic areas such as food production and supply, whilst also re-organising socio-spatial divisions of labour in food provisioning, most notably through the delivery rider. Distribution therefore has far-reaching effects, resulting in capacities to alter existing and generate new spatial functions and arrangements of urban economies of food provision.

Meal delivery: the co-production of urban space and commodity form

4Rather than the pairing of the meal with its delivery per se, the key innovation associated with Deliveroo in Newcastle, and in other cities, is the possibility to coordinate at very short notice, the restaurant, the means of delivery and the customer. Meal delivery is thus achieved as a particular rendering of urban space and its performance, through which the pairing of an object with its movement creates a commodity form that has a distinctive value precisely because it is mobile. The perishable nature of the meal means that this value is short-lived and so requires exact but often momentary spatial arrangements to ensure the commodity situation. This occurs through the technology platform, which “piggy-backs” on existing infrastructure such as restaurants, as well as generating the possibility for new infrastructural components, most notably the rider-deliverybox-vehicle assemblage. The performance of urban space that is coordinated through the technology platform then is also incorporated into the commodity form through the “last-mile” of post-purchase delivery (Boyer et al 2009), which puts different and residually independent food production and distribution networks into relation. The commodity of meal delivery thus reconfigures socio-economic definitions of urban space, which by combining production with elements of distribution and consumption do “not possess the traditional spatial properties that divide inside from outside in any meaningful sense” (Martin 2003, 7). Before examining the implications of this commodity form for urban economies of food provision, the remainder of this section will outline in more detail three base infrastructural components necessary for Deliveroo’s meal delivery in Newcastle: the technology platform; the rider-deliverybox-vehicle assemblage; and the restaurant collection point.

The technology platform

5The purpose of the technology platform is to coordinate the other elements necessary to achieve the delivered meal. The platform operates by collecting data produced by external actors in order to ensure coordination of delivery. When customers, riders and restaurants engage with the platform through the interface of an app, they each generate geolocational data, which is then recentralised when it is fed into the platform’s algorithm that plans the coordination of rider with restaurant and customer. Depending on the size of the urban area, this data and thus delivery itself, is divided into operational zones. In Newcastle, there were three such zones: the city centre, Gateshead (officially a separate city south of the River Tyne) and Gosforth (north of Newcastle city centre). These zones extended roughly 2-3kms, encompassing participating restaurants, and also marking the boundary within which riders must enter in order to receive an order (see below). The implication of zoning is that certain restaurants that are not situated within a zone are excluded from Deliveroo’s service, although in Newcastle this was not strictly the case. Several restaurants located approximately 1km outside of the city centre were included for delivery in this zone, indicating that the decision whether a given restaurant can participate is made on a case-by-case basis in a given city. Thus whilst the basic process of zoning urban space to delimit the collection of geolocation data is repeated in each city that Deliveroo operates, this is accompanied by local adaptation according to existing geographies of restaurant consumption.

6The physical presence though of representatives from Deliveroo in Newcastle to make these decisions was limited. Further, the majority of activities related to the establishment of a delivery fleet, take place online. This includes applying to work and various “on-boarding” activities that includes safety and hygiene training for riders. However, at the time of this research (2017-18), it remained the case that a local organising rep managed the distribution of rider equipment, and also checked IDs to ensure legal right to work. This rep was located, with the rider equipment (including delivery boxes, clothing, lights and so on), in a temporary Regus office in the city centre. For restaurants whose product appears on the app, interaction with a Deliveroo representative in person may never occur, although a variety of services are provided that enable a smooth transition to the Deliveroo system, most notably Point of Sale Technology. Given Deliveroo’s model of “partnering” (ie. outsourcing or subcontracting) with restaurants and riders, the actions of or occurring through the technology platform should be seen as distinct from the company itself. Whilst the platform software is the property of the company, essential to the definition of the platform as a technology is the possibility for others to build upon it in order to alter or adapt a particular operation or activity:

“a “platform” is a system that can be reprogrammed and therefore customized by outside developers – users – and in that way, adapted to countless needs and niches that that platform’s original developers could not have possibly contemplated, much less had time to accommodate”

(Andreessen cited in Helmond 2015, 3)

7Thus what occurs through the platform is necessarily shaped by but nonetheless distinct from the actions that Deliveroo intended in their design of the technology, as is clear in the problems that may arise through other components of the delivery infrastructure.

  • 1 The route is charted by the Deliveroo app but opens in whatever mapping app is the default on the p (...)

8To achieve the primary intended outcome - the coordination of delivery - the platform must calculate the journey of the meal from the restaurant to the customer. The solution to this problem of delivery involves not simply mapping the route1, but ensuring that there is an available delivery rider to undertake the journey for each order. Given the perishability of a meal, this allocation of a rider to an order must occur as quickly as possible, which requires utilising a delivery rider who is close to the restaurant. This in turn means ensuring that there are sufficient numbers of riders on the road to meet the demand for meals. Calculation to resolve these problems of delivery occurs through the algorithm designed by Deliveroo called “Frank”. Frank has two roles, one of which is selecting the appropriately located rider to pick up a meal, and the other is:

“Matching the number of orders … [Deliveroo] expect customers will make (order volume) with the number of riders needed to deliver them … As well as order volume, the tool takes a number of others factors into account, such as seasonal trends and holidays, e.g. Valentines Day to help [us] determine the number of sessions to make available in each zone, for every hour, in all [our] cities and countries around the world!”

(Deliveroo Tech Round-Up, June 2018, from author’s field notes)

9Frank is therefore a calculative agent that formulates a problem and its solution (what is to be done), and combines this with the strategy and instructions for carrying it out (how it should be done) (Kitchin 2017: 16). Such calculative agency is greatest when it can establish a long but finite list of entities, among which rich and varied relations are allowed, and through which formal procedures for the classification of these entities are set out (Callon and Muniesa 2005: 1238). This means that, in order to identify and classify these entities, calculative agency is distributed across different human and non-human actors and their equipment (Hutchins 1995).

10Thus, to calculate the delivered meal requires the collation and manipulation of data, which concerns a variety of different actors and concerning which the actors have different degrees of agency in producing. For example, in the allocation of orders to different riders, this data concerns:

“The specific dish that is being prepared; the location of the restaurant; the time of day and the day of the week; the number of riders on the road; how many live customer orders there are; the distance from the restaurant to the customer.”

(Deliveroo Tech Round-Up, May 2018)

  • 2 When a low ration of riders is predicted against a high volume of orders, incentives are given to e (...)

11Further, during the course of each individual order, there are particular spatial and temporal points marking key moments to coordinate the meal, the rider and the customer. These are distributed switches – or exchanges – when two (or more) parties (including people and things) interact, such as the rider recording via a swipe on their phone that they have collected the meal from the restaurant (Ruppert et al. 2013: 35). More broadly, to ensure that there are sufficient numbers of riders on the road, data concerning other factors are accounted for such as the weather, events like large football games and so on2. The possible implications of some of these factors (e.g. for volume of orders or availability of riders) can be predicted further in advance than others. So the closer to the present that different factors can be accounted for (i.e. the more immediate the production and communication of data), the more accurate is the calculation for each individual order and the matching of order volumes with another infrastructural component: riders.

The rider-deliverybox-vehicle assemblage

12Riders are one part of an assemblage for moving the meal through the city. Most typically in the case of Deliveroo, this assemblage comprises a rider (or driver), a vehicle (often but not exclusively a pedal bike) and a container for carrying the meal. This rider-deliverybox-vehicle assemblage constitutes a “vehicular agency”, a “hybrid and moving entity comprising a person and her things” (Calvignac and Cochoy 2016, 134). Calvignac and Cochoy (ibid.) argue that this type of assemblage as a whole directs the course of action, rather than any of the individual elements that constitute it. For example, the vehicle makes a difference to the speed and dexterity of movement, with bicycles generally slower than motorcycles and cars, but potentially more versatile in terms of the range of urban ground that they can cover. The rider or driver is nominally in charge of the vehicle, but the mobile device containing the app and directions, which tends to be attached to the vehicle rather than person, has some degree of agency in routing. And of course, the vehicle may suffer a mechanical failure, such as a puncture or a breakdown, so that the rider can no longer be on the road. Meanwhile, the delivery box is a piece of equipment, a “mobility-thing” (Hansson 2015), that enables the vehicle and rider to carry the meal from restaurant to customer. The delivery box is the key piece of equipment that new riders with Deliveroo have to pick up when they begin riding. It can be attached to the vehicle, go inside the vehicle, or most commonly carried on the back of the rider as a rucksack.

13Although the outside of the delivery box can be a hard (when attached to a vehicle) or soft (when a rucksack) plastic, all drivers must put food into a cushioned thermal bag that aims to both preserve the composition and the temperature of the meal whilst on its journey. Although the delivery box does not itself fail in terms of its capacities to carry food (except after a long duration), by the time of its delivery a meal is often not at the right temperature nor kept still and stable, which can result in spillages. The appearance of the meal when it reaches the customer can therefore be quite different from the “Instagram-able” images of food on the Deliveroo customer app. The temporariness of the collection point in Newcastle for this equipment (noted above) was intended to service particular moments of expansion of Deliveroo’s operation in the city. The company ran a number of promotional sessions in a separate temporary office space in the city at the beginning of the academic year in October 2017, to encourage new university students to sign up to ride. The aim was to have a limited “on the ground” presence in Newcastle to establish and expand operations for the winter months, including through provision of the necessary equipment. Since 2018, Deliveroo have changed their standard rider bag, away from a fixed box to a roll-top bag. The roll top bag has the advantage for the rider of being lighter, easier to wear and zip-free, so in theory longer lasting. For Deliveroo, the new possibility to compress the bag also meant that they could be sent via courier directly to the rider (as had been standard practice for their competitor UberEats), rather than employing local reps in each city to stock the equipment.

  • 3 Riders reject orders for a variety of reasons: two common reasons are because they want to go “offl (...)

14Despite the significance of the vehicle and the delivery box in the challenges of moving the meal through the city, there is greatest structural contingency to the mobile assemblage in the presence of the delivery rider, without whom there would be neither delivery box nor vehicle. This contingency derives from the operation of the platform, as the rider is the element that must be the most flexible to respond to order volumes in general, and specifically for each particular order, they must be able to reach the location of the restaurant and the customer. The requirement that prepared food is collected almost as soon as the order is placed means that a rider must be mobilized immediately to go to the restaurant to collect the food. Such a system relies on riders being “free” (i.e. from other orders through the platform) to pick up a meal, but also therefore creates a situation where if riders are “free” they could potentially be engaged in other activities (e.g. resting, delivering orders through other platforms). There is thus no guarantee that an order will be accepted by the rider to which it is allocated by the algorithm3, meaning that a new calculation of time and route for delivery has to be made when the order is assigned to a different rider. The action of the rider-deliverybox-vehicle assemblage as a piece of infrastructure is therefore necessarily dependent upon a set of contingent conditions that enable a rider to be available for a given order.

The restaurant collection point

My first order is at 7:15pm at a place called Smash Burger that I didn’t know. This is located inside Eldon Leisure [the 2016 redevelopment of a part of a larger “mall” in central Newcastle, now housing primarily restaurants]. I lock up my bike on the other side of the road and end up going into the wrong entrance of Eldon Leisure. There is no map inside, so I use my phone to navigate to the restaurant. By the time I get there, the order is ready. The waitress is not very busy and jokes that we should swap jobs when the man putting together the delivery orders mistakes her, out of the corner of his eye, for a delivery rider.

(Author’s field notes, 26/05/2018)

15The delivery rider interrupts existing infrastructures of food provision, as this episode indicates, with each order undertaken to collect a meal involving an interaction between them and the staff at a specific restaurant. If the rider-deliverybox-vehicle assemblage is a new piece of commodity infrastructure, or at least one that has increased in volume, then the restaurant as a collection point involves the transformation of an existing urban site intended for production and consumption. As indicated earlier, the major offer for restaurants from “partnering” with Deliveroo was the potential increase in sales, which the company claim can be up to 30%, especially for many high street restaurants that had previously not offered takeaway and delivery. Such restaurants thus become infrastructures not simply for production or on-site consumption but rather for the collection of a prepared meal, a change in function that places different requirements on existing restaurant space.

16These potential changes to restaurant space are most apparent when considering the positioning of the delivery rider when (s)he is waiting in the restaurant. Although for each individual rider this period is usually only a few minutes, at busy times there can be several riders waiting at once. The rider is neither restaurant staff nor customer and so does not have a clearly defined routine for their presence. The exact spatial practices of the waiting delivery rider therefore depend on adaptations of the specific service performances of each restaurant (Crang 1994). This means that delivery riders may wait with customers picking up orders, which is typical in fast food restaurants such as KFC and McDonalds in Newcastle. Here, the presence of the delivery riders made little difference to the performance of counter service staff whose role is to package different elements of an order and, once all items are assembled, to call out the number to waiting customers. In this case, the delivery rider is a proxy for another customer whose order must be assembled for collection. Elsewhere, riders may be directed to wait separately from customers who are (waiting to be) seated, a practice that occurred particularly in restaurants with a “maire d” at the entrance. At these sites, there would be a direct conversation between a member of staff and the rider about their order when they entered, and in one Chinese restaurant in Newcastle, an invitation to the rider to take a seat at a table with some water. In a minority of cases, riders could be instructed to go to a back entrance of the restaurant to collect the meal directly from the kitchen staff. This option was particularly favoured in restaurants located in more residential areas where space inside the restaurant was often limited and back lanes (vehicle accessible roads between the back yards of terraced housing) were present.

17So whilst all participating restaurants serve in some way as collection points, many do not have a clear location for either collection or waiting to occur. Thus the restaurant collection point is an idiosyncratic infrastructural element whose functioning currently varies greatly between different sites. Each time they enter a restaurant, the waiting rider (and usually their deliverybox) is inserted into an existing contained system of food production and consumption that they have to navigate. This can result in forms of miscommunication, for example, instances of being instructed to wait incorrectly:

At around 8:25pm I receive the fourth order of the night which is … [restaurant name] in China Town, just for some rice. The restaurant seems to be empty and looks a little bit like a warehouse; there are a lot of random objects around on the surfaces near the door, including a Hungry Panda leaflet. The two women at the door tell me it will be a moment, as I see a small bagged box with an order receipt on it near them. I stand around waiting for somewhere in the region of five minutes before two men emerge from the back of the restaurant, presumably the kitchen, and ask me which order I am. I tell them the number and they point to bag I had spotted earlier and say it’s already there.

(Author’s field notes, 10/06/2018)

18This and other nuances concerning the presence of the delivery rider in the restaurant indicate the difficulties in articulating the collection point as an infrastructural boundary between the directions of the Deliveroo app and the specific modus operandi of a given restaurant.

Distribution and the changing socio-spatial definition of urban economies of food provision

19In combination these infrastructural components mean that the commodity of meal delivery in Newcastle requires attention to the shifting socio-spatial definition of economies of urban food provision. The distribution of the meal – its movement through urban space – is vital to its situation as a commodity, creating new functions of production and consumption at its origin and destination sites. So Deliveroo’s operations demonstrate the importance of distribution not simply as a means of moving from sites of food production (or supply) to food consumption (or access), but rather for the processes of commodification and value extraction that can occur through distribution itself; processes that necessarily alter the socio-economic purposes of urban space. Such a shifting geography associated with meal delivery in UK cities is emerging at a time when “urban food governance” is a growing agenda. Over the last decade, there has been a “proliferation of urban food strategies, charters and plans” in the UK aimed at addressing diverse issues related to food provision in cities (Sonnino 2019, 14), but thus far Deliveroo’s operations have largely been understood outside of the remit of such governance.

20As the next subsection outlines, this is because the technological order of reality created through Deliveroo’s operations is one that foregrounds individual “on-demand” delivery, setting out to use distribution to respond to problems of personal convenience, and thus situating the company neither clearly in the realm of food production nor consumption. The second subsection outlines how city authorities meanwhile have been focused on issues of sustainable food production and consumption, and so have made limited attempts to target the operations of Deliveroo, particularly from a food governance perspective, despite evidence that the company is extending its reach to intervene in geographies of food supply. Therefore, although currently subject to minimal city governance, the final subsection suggests that Deliveroo, and associated agencies of distribution that can coordinate meal delivery at a city and neighbourhood scale, are reorganising urban economic functions through changing socio-spatial definitions of “productive” and “socially reproductive” activity. Social roles and interdependencies are emerging that do not always fit easily into existing socio-spatial categories for production or consumption but rather point to the forms of complexity and diversity required to maintain an urban infrastructure of food distribution (Tsing 2009).

Realising on-demand urban delivery

21The core business offer behind Deliveroo’s service is that of an “application” that “brings you food whenever and wherever you want it”, in what is termed “on-demand” delivery that foregrounds the importance of individual customer convenience and choice. By providing delivery, Deliveroo plays the role of distributor, sitting at the site of production nor the site of consumption but rather mediating them. Thus, on-demand consumption intensifies pre-existing notions of convenience, building on two historical tendencies in the discursive realities associated with technologies that shape the customer-service provider relationship in food retail in the UK. The first is “speed”, epitomised by fast food, which has historically supported a culture of individual mobility, often but not exclusively realised through the car (Jakle and Sculle 1999; Urry 2004). Rather than speed itself (which is of course relative), it is the explanatory power of speed that is significant as a means of validating a given technology involved in food provision and consumption (Sharma 2017, 138). The second is “self-service”, where customers select items (in a store) themselves, therefore providing an illusion of the removal of dependence on (waiting for) service by others. This is justified by the discourse of speed, which brings a logic to the goal of rapid movement from one thing to the next that is supposedly enabled by self-service. The supermarket is the archetypal example of a self-service site, which initially gave shoppers an experience of “liberation” from “class-based status and ascription” associated with previous personal service (Du Gay 2004, 158). The relative ease with which consumers have subscribed to the on-demand convenience provided through Deliveroo’s food delivery infrastructure in Newcastle thus in part can be explained by its extension of these existing discourses of speed and self-service associated with contemporary technologies of food provision.

22Deliveroo’s framing of its on-demand operations is also orientated around individual choice, with the company promising the customer with not simply any food, but rather “good food”, often meaning food of “restaurant quality”. Prior to the launch of Deliveroo in the UK in 2013, aggregator platforms such as Just Eat and Hungry House provided a portal for online selection, ordering and payment from existing take-away food establishments, with the delivery of the order to the place of consumption to be carried out either by the restaurant or by the customer. Deliveroo similarly aggregated restaurants in one virtual place, but with a key difference that enabled the realisation of greater consumer choice through the infrastructure. The company enrolled (or “partnered” with) – for a commissioning fee per order – restaurants that had previously not offered delivery. Deliveroo could thereby provide options that diverged from the pizza, Indian and Chinese (and of course “fish and chips”) meals that had previously dominated UK takeaway food (Warde and Martins 1999). The restaurants that were newly offering delivery in Newcastle, as elsewhere, included a mixture of independents, global fast food names (such as KFC and Burger King) and British “casual dining” restaurant chains. The latter are restaurants with table service and a license for alcohol (so distinct from fast food restaurants), and therefore play an important part in Newcastle’s night-time economy (Shaw 2015). The investment group backed expansion of casual dining has contributed to the rise in the frequency of “eating out” in Britain; from roughly 80% eating out less than once a month or not at all in 1989, to nearly 70% eating out once a month or more by 2015. However, from 2017 onwards, a “crunch” on such casual dining restaurants occurred, with closures of sites by several well-known restaurant brands.

23In the on-demand reality that Deliveroo seeks to create though, the existing urban geography of physical restaurants, and their success or failure, can easily fade from view. The company aims through its customer app to provide a “virtual marketplace”, a one stop shop for many different food options in which the origins of the foods’ preparation (ie. the physical restaurant space) are not always the most important means of indexing the choices available. Although searching for a specific restaurant on the app is possible, Deliveroo determines the display of food choices to include other ways of categorising the available products, such as types of cuisine, “deals” or previous customer orders. The simplicity of the on-demand order therefore can hide the geography of production and supply required to furnish it. This is a complex geography that is constituted by multiple different layers of relational space-time information concerning data on distance to be travelled, food preparation time and rider availability. It is these highly geographically specific actions of certain actors that somewhat paradoxically enable a degree of deterritorialisation of customer meal choices from their restaurant location in the city. This deterritorialisation associated with the customer app combines with other elements of Deliveroo’s technology platform, which despite having highly territorialised implications for food supply and distribution in Newcastle, can simultaneously create a reality in “which actions are performed outward into a world that is itself not depicted” (Guyer 2016, 114). Thus the customer app is both the site of individual choice and simultaneously the means for disappearing the collective urban infrastructures upon which the realisation of this choice depends.

Urban food governance and the neglect of delivery

  • 4 Such policy includes a broad range of issues relating to food supply (and production) and distribut (...)
  • 5 Produced by The Food Newcastle Partnership in collaboration with Newcastle City Council.

24By casting itself as such an individual customer-orientated delivery service, and so neither quite in the realm of food production nor consumption, Deliveroo has thus far managed to separate its infrastructural operations from the purview of city authorities targeting issues of urban food governance. In the past decade in the UK, the conviction that city, rather than national governments, can play a significant role in issues of food policy4 has become increasingly prevalent for two reasons (Sonnino 2016). Firstly, a motivation to “harness the power of civil society groups and food movements” that often operate at the urban scale and align with the agenda of “sustainable food security” (Sonnino 2019, 14). Secondly, the attempt to fill the policy vacuum left by national policies that have had a limited (and even negative) impact upon individual household ability to secure food (ibid.). Both of these elements are present in Newcastle’s 2018 “Good Food Plan”5, in which sustainable food consumption and food poverty are for the most part addressed through existing civil society organisations rather than the national or local state. At issue in Newcastle, as elsewhere, are questions of the governance of food distribution, meaning the mechanisms for access to food, rather than the governance of overall quantity of food supply (i.e. whether there is enough food). This politics of distribution therefore spans mechanisms of price-setting to the material means of provisioning food in a given urban area. Given the latter, it is somewhat surprising that Deliveroo’s infrastructure for meal delivery has thus far largely not been considered within urban food governance debates.

25This is not to say of course that city authorities have been oblivious to Deliveroo’s meal delivery infrastructure. Newcastle City Council for example have cautioned Deliveroo riders against the improper use of roads in the city centre to access restaurants for order pick up (The Chronicle 2019). As this indicates, however, the terrain for government intervention relates to forms of “disorder” of urban space associated with this infrastructure rather than any direct concern with what the infrastructure is providing (i.e. food). Indeed, it is not only Deliveroo’s marketing of meal delivery as an individualised service that leaves the infrastructure outside the purview of government, but also a wider UK political context in which infrastructure is largely “private”, meaning beyond the scope of public control with any state intervention usually limited to security grounds (Graham and Marvin 2001). The recent events requiring social distancing have exposed the problems of this retrenchment of the state but also illustrated the insufficiency of any easy separation between public and private infrastructure (although not ownership), as Deliveroo’s capacities for urban food distribution were able to meet some of the new household provisioning needs (The Independent 2020). The absence then of Deliveroo’s meal delivery from urban food governance agendas can be understood as a dual function of firstly its positioning as an individual customer-orientated service rather than a collective mode of provisioning or urban infrastructure, the latter of which have secondly through privatisation anyway fallen increasingly beyond the remit of city authorities.

26Whilst the neglect in urban food governance strategies of these meal delivery infrastructures is therefore understandable within this context, it may become increasingly untenable. Deliveroo has a more and more visible presence in, and firm grip on, elements of urban food production, as well as distribution, through extensions to the basic method by which the company has built the meal delivery infrastructure. The original opportunity for customers to order online was pitched by Deliveroo to restaurants as a means of “virtually” expanding their site, with Deliveroo providing the technology for meal selection, ordering, payment and delivery. Without adding to the number of tables or service staff, the company claimed restaurants could up their revenues by furnishing orders received through the Deliveroo restaurant app, a claim supported by data in a study carried out by Deloitte (2019) that found an increase in the number of meals sold (up 4% in London), translating to a rise in revenues of 1.4% across London’s restaurant sector. Deliveroo’s initial step using such existing restaurant sites for urban food preparation and processing has subsequently been followed by the establishment of ‘dark kitchens’ in London, that were renamed by 2017 as “Deliveroo Editions” sites. These are purpose built kitchens designed solely to service orders via the Deliveroo app or website, hence the branding of “Editions”. In order to establish any such Editions site, “Deliveroo uses its data insights to identify cuisines that customers want but do not have access to locally, and then invests in kitchen infrastructure to allow restaurants who excel in those cuisine types to set-up in those local areas” (Deliveroo, n.d.). Deliveroo Editions therefore enacts a further delinking of “restaurant” food from the physical space of the restaurant, constructing an altered geography that is quite deliberately intervening in urban food supply.

27As it establishes these physical sites, the company is making decisions about what type of food is accessible and where, based on data from customer consumption patterns rather than any other measures of local need. The result is that for their Editions kitchens, Deliveroo target both restaurants with a physical “high-street” presence and also micro businesses (termed “street-food entrepreneurs” by Deliveroo) who are untested in fixed urban retail premises but have a menu that is perceived to fulfil a particular market niche. The Editions label is then also used on the customer app to categorise so-called “virtual brands” that comprise menus created by existing physical restaurants solely for order through Deliveroo, to provide these businesses with the possibility of developing an additional revenue stream. Through the Editions sites then, Deliveroo are thus seeking to increase the choice of food can be distributed to customers, but further, the company are also reaching into the management of kitchen food production and supply. Since 2017, Deliveroo have been offering “The Deliveroo Toolkit”, a workflow technology aimed at more efficiently managing production in the kitchen. The technology, which was acquired through Deliveroo’s purchase of company Maple in 2017, was initially used at Deliveroo Editions sites before being offered to other restaurants in 2018. The company claim that the technology can improve kitchen output by up to 100% through various techniques such as providing preparation time on an item level basis; displaying a real-time kitchen overview dashboard so that adjustments can be quickly identified to ensure “uninterrupted flow”; and managing purchasing operations by predicting weekly order quantity (Deliveroo, 2017). So although initially claiming a position as an apparently neutral mediator between restaurant and customer, Deliveroo’s most recent moves to control elements of food supply indicate that the ensuing mixture of hard and soft infrastructure can significantly shape the production and distribution of food in cities.

Boundary work, spatial definition and the social maintenance of delivery infrastructure

  • 6 Preparation that can take place in restaurants but also includes the manufacture of processed food (...)

28Regardless of its model of ownership then, the infrastructures supporting and emerging from meal delivery are changing the socio-spatial definition of urban food provision in the UK. Whilst Deliveroo emphasise the new possibilities for personal convenience, realising such convenience has wider effects on the spatial division of labour for food preparation and thus the organization of (paid and unpaid) roles and interdependencies constituting “social reproduction” (Dalla Costa and James 1972; Hochschild 1997). By increasing the possibilities for domestic meal delivery, Deliveroo extend the existing trend towards food preparation outside the home6 that depends on highly gendered and racialised labour for its execution (Glenn 1992). As well as food service workers, this socially reproductive labour of urban food provisioning in UK cities such as Newcastle also now includes increasing numbers of delivery riders, whose apparently simple task of getting the meal from A to B belies their wider social role. So on-demand consumption by no means removes social dependencies but rather relies on “other temporalities that labour in order to maintain this time” of convenience for others, through complex syncopation to deliver according to the governing temporal architecture (Sharma 2017, 141). As Sharma (2010) has noted in relation to taxi drivers, such transport labour as “low skill” and low pay is frequently male and non-White, partly as a result of wider structural racism in the labour market but also because of the qualities of the job. Rhythms of food delivery riding involve moments of rapid movement punctuated with potentially long, lonely and unpredictable periods of waiting. The rider’s stamina thus acts as a buffer that compensates for lags and inconsistencies in the spatio-temporal coordination of different networks required for meal delivery: waiting (in the cold and wet) to receive an order to go to a restaurant; waiting for a meal to be prepared and finally waiting for the customer to come to door.

29Through these movements and stoppages then, the delivery rider translates between different socio-technical systems, together with their associated spatial and temporal patterns of organisation, in ways that both retain and alter existing socio-economic functions of space. In particular, the rider is incorporated into a separate temporality of food service in each restaurant. The meeting of restaurant staff with the delivery rider is therefore a point of interaction – or “boundary work” (Downey 2001) – between separate networks, which by enabling the movements of goods and information across these spaces simultaneously alters their function. Boundary work is necessary because:

“No matter what automated protocols are in place in any given moment, they will be imperfect and incomplete; disparate information networks can only work together through the efforts of specific workers who maintain the links, transform the content, and police the boundaries between those networks.”

(Downey 2001: 225)

30New occupations and new spatialities of activity therefore emerge as a result of boundary work, as Downey illustrates through his example of the telegraph messenger boy. These roles “become involved in smoothing the transition from one network to another”, with changes in these occupations helping to define the spaces of the network themselves, “especially those spaces where the networks present themselves to their users” (Downey 2001: 225). The presence of the delivery rider with their box in the restaurant space then becomes such a sight of transition between networks and redefinition of urban space, where delivery processes are visible to restaurant customers and problems of translation, such as the wrong order number or composition, can arise, be found and generally resolved.

  • 7 Being online does not guarantee though that any orders will be received, and therefore that any mon (...)

31As per the history of socially reproductive labour, the essential function of such boundary work in the food delivery infrastructure is not reflected in the remuneration or working conditions of delivery riders. The poor pay and self-employed status (meaning few social securities) of riders is exacerbated by Deliveroo’s deliberate deprofessionalisation of such work that justifies worker disposability. The company consistently label delivery work “flexible”, and through this seek to imply that riders can choose when they want to work, but such rhetoric also enables Deliveroo to brush aside any necessity to guarantee delivery work. Instead, riders must be willing “to conform to the calculative rationalities that companies project onto them” (Shapiro 2018, 2968), which are often based on resolving immediate and predicted problems of market supply and demand. This uncertainty associated with flexible working is encapsulated by the operation of the Deliveroo rider app, via which riders have to go “online” so that they can receive orders and thus get paid7. However, riders can only go online when there is space available in an individual hourly slot on any given day. The purpose of the slots is to ensure both that there are the right number of riders on the road to meet demand for orders to be delivered, and that the riders that are on the road will receive orders to deliver, to avoid oversupply of delivery capacity. The occupation of the rider therefore not only exists to service the temporalities of others, but further, to perform this role unpaid work must always be undertaken to secure the possibility of being paid.

32In practice then, by rendering work contingent, flexibility undermines attempted professional commitment by the delivery worker to their work and further, is frequently accompanied by suggestions from Deliveroo that delivery work is a lifestyle choice or a means of supplementing income; the subtext being that this is not a “proper job”. Yet despite this, workers are subject to performance management through the use of data on their working patterns that shapes their possibilities to work in the future. These statistics are three measures calculated on a weekly basis: the percentage of attended sessions; the percentage of late cancellations; and the percentage of super-peak sessions attended. Riders who have high overall and super-peak attendance, together with a low percentage of late cancellations, relative to other riders in the delivery zone, are given priority booking of slots to work. Therefore, contrary to the rhetoric of flexibility, Deliveroo rewards riders who are more reliable by giving them the opportunity for greater choice over when they want to work. The opportunity for certain “priority” riders to choose when to be available to work therefore demands greater flexibility from those who are unable to book as they scrabble to gain access to slots. The maintenance of the on-demand meal delivery infrastructure depends on the possibility to leverage flexibility of labour in general, which has long been required to service just-in-time factory production (Pollert 1988). However, on-demand meal delivery further individualises flexibility, as riders exist without employment or workplace but rather work contingently at the boundaries and redefinitions of the socio-economic functions of urban space.


33The distribution of prepared meals in cities such as Newcastle extends some of the logics of logistics underpinning global supply chains for all sorts of commodities. The technologies of spatial and temporal coordination that are necessary for the movement, tracking and tracing of goods in transnational circulation are operationalised in cities to furnish the post-purchase delivery of perishable prepared meals. The technology platform that calculates coordination operates in combination with an existing physical urban infrastructure of restaurants, and facilitates the proliferation of another infrastructural component, the rider-deliverybox-vehicle assemblage. By delivering the meal to the right place at the right time, the performance of urban space through distribution is incorporated into the commodity form, as the meal’s valuation is dependent on its qualities and quantities of movement. The post-purchase delivery of food, and indeed other goods in cities, is thus important not simply for the ways that it is altering acts of consumption, but equally as significantly because of the extraction of value by different actors in distribution has agency to change existing and generate new urban socio-spatial functions and arrangements. This is apparent with Deliveroo’s current operations that are not only reorganising urban geographies of eating out through “on-demand” consumption, but also building from these operations, the company is beginning to alter processes of food production and preparation. As such, there is an emerging reorganisation of socio-spatial divisions of labour in urban food economies in the UK.

34The significant yet somewhat fragile power of logistics noted in the global economy is thus gaining form in urban space through agencies of distribution (Gregson et al 2017). As with global supply chains, such agencies creating distribution are dispersed, not simply concentrated in technologies that calculate coordination but rather realised through the contingent arrangement of independent actors, the combination of whom calculation sets out to resolve. As well as the traffic of delivery, the agencies of distribution are apparent in changes to the retail geography of the city, as shops and restaurants take on the function of stores and assembly spaces rather than sites of purchase. Distribution therefore also has important implications for questions of urban food governance (Sonnino 2019). Delivery can potentially remove some barriers of access to food that have been associated with geographical proximity to shops and restaurants, to some extent building on exist technologies of food sharing coordination (Davies et al 2017). However, distribution can also create new problems concerning supply and pricing, and perhaps most importantly an insecure and potentially unreliable urban infrastructure largely serviced by those working in precarious conditions.

Top of page


Boyer, K.K., Prud’homme, A.M., Chung, W., 2009. The last mile challenge: evaluating the effects of customer density and delivery window patterns. Journal of Business Logistics 30, 185–201.

Callon, M., Muniesa, F., 2005. Peripheral Vision: Economic Markets as Calculative Collective Devices. Organization Studies 26, 1229–1250.

Calvignac, C., Cochoy, F., 2016. From “market agencements” to “vehicular agencies”: insights from the quantitative observation of consumer logistics. Consumption Markets & Culture 19, 133–147.

Cidell, J., 2012. Flows and Pauses in the Urban Logistics Landscape: The Municipal Regulation of Shipping Container Mobilities. Mobilities 7, 233–245.

Cidell, J., 2011. Distribution Centers among the Rooftops: The Global Logistics Network Meets the Suburban Spatial Imaginary. International Journal of Urban and Regional Research 35, 832–851.

The Chronicle. 2019. Deliveroo riders in Newcastle putting pedestrians at risk, city transport bosses warn. Available at:

Crang, P., 1994. It’s showtime: on the workplace geographies of display in a restaurant in southeast England. Environment and Planning D: Society and Space 12, 675–704.

Dalla Costa, M., James, S. (Eds.), 1972. The Power of women and the subversion of the community. Falling Wall Press Ltd, Bristol.

Davies, A.R., Edwards, F., Marovelli, B., Morrow, O., Rut, M., Weymes, M., 2017. Making visible: Interrogating the performance of food sharing across 100 urban areas. Geoforum 86, 136–149.

Davies, A., Evans, D., 2019. Urban food sharing: Emerging geographies of production, consumption and exchange. Geoforum 99, 154–159.

Deliveroo. 2017. New technology will help UK restaurants produce double the number of meals. Available at:

Deliveroo. no date. New London Plan: Deliveroo Response. Available at:

Deloitte. 2019. Delivering Growth: the impact of third party platform ordering on restaurants. Available at:

den Hartog, A., 2003. Technological innovations and eating out as a mass phenomenon in Europe, in: Jacobs, M., Scholliers, P. (Eds.), Eating out in Europe: Picnics, Gourmet Dining and Snacks since the Late Eighteenth Century. Berg, New York, pp. 263–280.

Downey, G., 2001. Virtual Webs, Physical Technologies, and Hidden Workers: The Spaces of Labor in Information Internetworks. Technology and Culture 42, 209–235.

du Gay, P., 2004. Self‐Service: Retail, Shopping and Personhood. Consumption Markets & Culture 7, 149–163.

Ferguson, J., 2015. Give a man a fish: reflections on the new politics of distribution. Duke University Press, Durham.

Gibson-Graham, J.K., 2006. A postcapitalist politics. University of Minnesota Press, Minneapolis.

Glenn, E.N., 1992. From Servitude to Service Work: Historical Continuities in the Racial Division of Paid Reproductive Labor. Signs: Journal of Women in Culture and Society 18, 1–43.

Graham, S., Marvin, S., 2001. Splintering urbanism: networked infrastructures, technological mobilities and the urban condition. Routledge, London ; New York.

Gregson, N., Crang, M., Antonopoulos, C., 2017. Holding together logistical worlds: friction, seams and circulation in the emerging “global warehouse.” Environment and Planning D: Society and Space 35, 381–398.

Guyer, J., 2016. Legacies, logics, logistics: essays in the anthropology of the platform economy. The University of Chicago Press, Chicago.

Hansson, N., 2015. “Mobility-things” and consumption: conceptualizing differently mobile families on the move with recent purchases in urban space. Consumption Markets & Culture 18, 72–91.

Helmond, A., 2015. The Platformization of the Web: Making Web Data Platform Ready. Social Media + Society 1.

Hochschild, A., 1997. The Time Bind: When work becomes home and home becomes work. Henry Holt, New York.

Hutchins, E., 1995. Cognition in the Wild. MIT Press: Cambridge MA.

The Independent 2020. Which supermarkets let you order groceries through Deliveroo? Available at:

Jakle, J., Sculle, K., 1999. Fast food: roadside restaurants in the automobile age. Johns Hopkins Univ. Press, Baltimore.

Kitchin, R., 2017. Thinking critically about and researching algorithms. Information, Communication & Society 20, 14–29.

LeCavalier, J., 2016. The rule of logistics: Walmart and the architecture of fulfillment. University of Minnesota Press: Minneapolis.

Martin, R., 2003. The organizational complex: architecture, media, and corporate space. MIT Press, Cambridge, MA.

Neilson, B., 2014. Beyond Kulturkritik: Along the Supply Chain of Contemporary Capitalism. Culture Unbound: Journal of Current Cultural Research 6, 77–93.

Newcastle Food Partnership 2018. Newcastle Good Food Plan 2018. Available at:

Pollert, A., 1988. Dismantling flexibility. Capital & Class 12, 42–75.

Ruppert, E., Law, J., Savage, M., 2013. Reassembling Social Science Methods: The Challenge of Digital Devices. Theory, Culture & Society 30, 22–46.

Shapiro, A., 2018. Between autonomy and control: Strategies of arbitrage in the “on-demand” economy. New Media & Society 20, 2954–2971.

Sharma, S., 2017. Speed traps and the temporal: of taxis, truck stops and TaskRabbits, in: Wajcman, J., Dodd, N. (Eds.), The Sociology of Speed: Digital, Organizational, and Social Temporalities. Oxford University Press, Oxford, pp. 131–151.

Sharma, S., 2010. Taxicab publics and the production of brown space after 9/11. Cultural Studies 24, 183–199.

Shaw, R., 2015. ‘Alive after five’: Constructing the neoliberal night in Newcastle upon Tyne. Urban Studies 52, 456–470.

Sonnino, R., 2019. The cultural dynamics of urban food governance. City, Culture and Society 16, 12–17.

Sonnino, R., 2016. The new geography of food security: exploring the potential of urban food strategies. The Geographical Journal 182, 190–200.

Stenmanns, J., 2019. Logistics from the margins. Environment and Planning D: Society and Space 37, 5, 850-867.

Tsing, A., 2009. Supply Chains and the Human Condition. Rethinking Marxism 21, 148–176.

Urry, J., 2004. The ‘System’ of Automobility. Theory, Culture & Society 21, 25–39.

Warde, A., Martens, L., 1999. Eating out: reflections on the experience of consumers in England, in: Germov, J., Williams, L. (Eds.), A Sociology of Food and Nutrition. Oxford Univ. Press, Oxford, pp. 116–134.

Top of page


1 The route is charted by the Deliveroo app but opens in whatever mapping app is the default on the phone, such as Google Maps. The initial route is not always the best or fastest route, so riders often use their initiative to make the delivery.

2 When a low ration of riders is predicted against a high volume of orders, incentives are given to encourage riders to deliver, for example a £2 bonus per order.

3 Riders reject orders for a variety of reasons: two common reasons are because they want to go “offline” (finish working); or because they have an issue with delivering food from a particular restaurant. That issue could be the distance to the restaurant or the wait time it takes for orders to be prepared. In both cases, riders perceive that delivering from that restaurant is not an efficient way to make money. When a rider rejects an order, the app automatically asks the reason why the order is rejected (a menu of choices is provided from which the rider has to select).

4 Such policy includes a broad range of issues relating to food supply (and production) and distribution at both the macro and micro-scales; often referred to as “food security”, although this term tends not to be used explicitly in urban policy documentation (Sonnino 2016).

5 Produced by The Food Newcastle Partnership in collaboration with Newcastle City Council.

6 Preparation that can take place in restaurants but also includes the manufacture of processed food in factories for sale in supermarkets (Den Hartog 2003).

7 Being online does not guarantee though that any orders will be received, and therefore that any money will be made.

Top of page


Electronic reference

Lizzie Richardson, Meal delivery logistics and the agencies of distribution in urban economies of food provision in the UKArticulo - Journal of Urban Research [Online], 21 | 2020, Online since 15 December 2020, connection on 04 December 2023. URL:; DOI:

Top of page

About the author

Lizzie Richardson

Lizzie Richardson is a Lecturer in Digital Media and Society in the Department of Sociological Studies at the University of Sheffield (UK). Much of her current research focuses on geographies of work in UK cities. Address: Department of Sociological Studies, University of Sheffield, Sheffield, UK.

Top of page



The text only may be used under licence CC BY-NC-ND 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search