This is a post-peer-review, pre-copyedit version of an article published in International Journal of

Logistics Management. The final authenticated version is available online at:

https://doi.org/10.1108/IJLM-03-2020-0150

To cite this article:

Seghezzi, A., Winkenbach, M. and Mangiaracina, R. (2021), "On-demand food delivery: a systematic
literature review", The International Journal of Logistics Management, Vol. 32 No. 4, pp. 1334-1355.
https://doi.org/10.1108/IJLM-03-2020-0150

On-Demand Food Delivery: a systematic literature review

Abstract

Purpose – The purpose of this paper is twofold: first, to review the extant academic literature on

On-Demand Food Delivery (ODFD) services, i.e., the delivery of freshly prepared meals from

restaurants to customers enabled by online platforms; second, to propose directions for future

research in this field.

Design/methodology/approach - Our work reviews 59 papers published between 2016 and

2020. They are classified with respect to both their research methodology and the addressed

themes, namely the actors involved and the activities creating value for the ODFD ecosystem.

The latter are analysed applying a framework derived from the integration of traditional models

and literature about platforms/business ecosystems. Results were validated through interviews

with practitioners.

Findings – Due to its huge success and the significant complexities behind it, the ODFD

business has been gaining the interest of academics. Our review highlights that (i) the

perspectives of the various actors involved should be integrated, moving towards an ecosystem-

based view; (ii) future research should focus more on restaurant operations and their role in on-

demand ODFD systems; and (iii) despite they have been investigated by several academic

contributions, human resources management and logistics of ODFD systems still present room
for further extensions, in the areas of intervention/regulation and distribution network/batching

respectively.

Originality/value – This review offers insights to both academics and practitioners. On the

academic side, it analyses the literature on ODFD systems, outlining directions for future

research. On the managerial side, it provides a comprehensive view on the most critical value-

creating activities for ODFD businesses.

Keywords – Food Logistics; Urban logistics; Logistics services; Omni-channel.

Paper type – Literature review

Introduction

The so called On-Demand Food Delivery (ODFD), i.e., the purchase and delivery of freshly

prepared meals from restaurants to the customers’ home enabled by the use of online platforms,

has been steadily growing all over the world in the last few years. This novel and disruptive

business – based on an ecosystem made by the platform, restaurants, riders and customers – is

both very significant (as the related market is growing) and critical (due to very peculiar

complexities that characterise it). Accordingly, it has gained the attention of practitioners, policy-

makers and academics.

Considering practitioners operating in the food and grocery industry, ODFD is a very attractive

market to new and technology-driven players, while being a threat to the incumbents in the more

traditional non-delivery food, grocery, and hospitality industry (He et al., 2019). What once was

a limited niche sector populated by start-ups, is now a flourishing business ruled by big players,

which is changing consumers’ habits related to a fundamental part of their daily lives: eating

(Wang and Somogyi, 2018). The significance of this business – documented by steady growth

rates in the US, Europe and Asia – has become manifest to practitioners all over the world

(Furunes and Mkono, 2019). As ODFD companies rapidly gain market share, they are

increasingly seen as rivals of two well-established, profitable businesses: grocery – since ready-

to-eat meals promptly delivered are a true alternative of buying and prepare food – and offline-

only restaurants – since customers may conveniently opt for ordering food from the Internet

rather than going out to eat (Ray et al., 2019).

Despite their positive economic outlook, ODFD businesses have been subject to controversy in

both the public opinion and the press, especially with reference to two key aspects. First, the

working conditions of the riders, who perform the delivery of the meals from the restaurants to
the customers. They typically are not employed, but work according to a crowdsourcing

paradigm (Altenried, 2019). Second, the impact that the diffuse presence of delivery agents

driving in an urban environment may have on traffic and road safety conditions (Christie and

Ward, 2019).

In the academic discourse, the features of ODFD make it a fruitful area of research for many

different fields and disciplines. First, the associated pickup and delivery problem is very

attractive for scholars in the logistics and operations research domains, due to its inherent

complexities: highly dynamic demand characteristics, very stringent time requirements, small

order dimensions, and highly perishable products (Allen et al., 2018). Second, it is very

interesting from a finance, strategy and micro-economics perspective. Several major mergers and

acquisitions completed in recent years all over the world have marked the onset of a significant

consolidation of the ODFD industry. This trend is driven by the need to aggregate the demand, as

the number and density of served customers is a key driver of cost and thus long-term viability of

ODFD business models (Zambetti et al., 2017). Third – aligned to the importance of serving a

high number of customers – ODFD has been capturing the interest of authors from the marketing

field, who are determined to investigate initiatives aimed to enlarge the customers’ base and

increase their intention to buy (Yeo et al., 2017). In addition, academics coping with labour

policies and legal issues of ODFD services may find a fertile context, with a particular

implementation of the disruptive gig-economy phenomenon that deserves an in-depth study

(Goods et al., 2019).

In light of these premises, there is growing interest of the academic community and literature in

ODFD. However, the research field is very broad and includes diverse domains. Thus, extant

knowledge is very dispersed and fragmented, and a clear and organised overview of the current
state of the literature is missing. While some attempts have been made in this direction

considering just food logistics (e.g., Cerchione et al., 2018) or crowdsourcing logistics (e.g.,

Allen et al., 2018), similar works specifically targeting ODFD (and therefore the intersection of

food logistics and crowdsourcing) are still unavailable.

Objectives and Methodology

The increasing interest of academics, the significance of the topic for practitioners and

policymakers and the lack of a clear and organised view of extant knowledge are adequate

reasons to conduct a systematic review of the literature (Huscroft et al., 2012). There is a

particular need for such a review in light of novel phenomena such as ODFD services (see for

instance the paper by Lim and Winkenbach, 2019). In this context, this work aims to (i) provide

a systematic analysis and classification of extant the body of research on ODFD, (ii) outline the

major gaps in the extant literature, and (iii) accordingly propose directions for future research,

which may be of relevance for both the academic and the managerial communities.

In line with the set objectives, and with the aim of developing the framework to classify the

extant body of research, this work addresses the following two questions:

RQ1. What are the main actors in ODFD ecosystems and what is their role? – to identify the

main actors involved in ODFD ecosystems and to understand their role and interactions.

RQ2. What are the main value-creating activities and processes performed in ODFD

ecosystems? – to identify and discuss the most significant value-creating activities and processes

performed by the actors in ODFD ecosystems.

To answer these questions, a systematic review of the literature about ODFD was conducted.

Based on the recommendations in the methodological paper by Seuring and Gold (2012), our
review followed four main phases: phase 1 (material collection) to retrieve and select papers;

phase 2 (descriptive analysis) to analyse the contributions based on their main “descriptive”

characteristics, i.e., year of publication, region/country, title of the journal/name of the

conference; phase 3 (category selection) to identify the dimensions/categories for the content

analysis; phase 4 (material evaluation) to review and classify the works based on the defined

dimensions.

At the end of the review process, in addition to the defined steps, five interviews with ODFD

practitioners were performed to validate the results (phase 5).

Phase 1: Material collection

The first phase – i.e., collecting and selecting the papers – can be seen as composed of three

main sub-steps, which coincide with steps 2 to 4 of those proposed in the methodological paper

by Durach et al. (2017).

(i) Determine required characteristics of primary studies – The requirements for the papers were

defined, and they were subsequently expressed as inclusion and inclusion criteria. The focus of

the analysis is the delivery (a) of freshly prepared meals (b) from restaurants to customers’

houses (c) enabled by the use of online platforms. Accordingly, as shown in Table 1, we

excluded both (a) the delivery of non-fresh – i.e., enogastronomic or grocery – products (such as

oil or canned food) and (b) the delivery of fresh prepared meals from supermarkets. In addition,

we rejected (c) works investigating the delivery of meals directly managed from single

restaurants (e.g., delivery of a pizza ordered from a local “pizzeria” by the phone, accomplished

by the deliveryman of the restaurant). As a matter of fact, their case is significantly different with

respect to online platforms such as Glovo or Deliveroo that have the peculiarity of combining the
offer of different restaurants, which accordingly share not only the app, but also logistics

resources (and they benefit from a centralised management).

Inclusion criteria Exclusion criteria

(a) Delivery of fresh prepared meals Delivery of enogastronomic or grocery
products
(b) Delivery of fresh prepared meals from Delivery of fresh prepared meals from
restaurants supermarkets
(c) On-demand delivery of fresh prepared meals Delivery of fresh prepared meals managed
from restaurants enabled by online platforms by single restaurants themselves
Table 1: Inclusion and exclusion criteria

Considering the unit of analysis, it was selected as a single scientific paper, not only from

journals, but also from conference proceedings. The choice to include not only black, but also

grey literature is due to the novelty of the theme. As a matter of fact, the considerable time

needed for a paper to be published on international journals could otherwise result in missing

relevant contributions (Mangiaracina et al., 2019).

(ii) Retrieve sample of potentially relevant literature – first, a search by keywords was performed

in two library databases (i.e., Scopus, ISI Web of knowledge). The selected keywords included

both alternative names for ODFD found in literature (which are shown in Table 2, together with

the related source) as well as the names of some popular ODFD platforms (“Deliveroo”,

“Zomato”, “Uber Eats”, “Glovo”, “Swiggy”, “FoodPanda”, “Doordash”, “Grabhub”).

Keyword Source
Food Delivery Apps (FDA) Ray et al., 2019
Food Online Order and Delivery (FOOD) Preetha and Iswarya, 2019
Online Food Aggregators (OFA) Kapoor and Vij, 2018
Online Food Delivery (OFD) Yeo et al., 2017
Online Meal Order and Delivery (OMOD) Pinto et al., 2019
Table 2: Keywords

As suggested by Durach et al., (2017), we searched the titles, abstracts, and keywords of papers

for the above-mentioned terms. In addition, similarly to Nguyen et al. (2016), a further
“snowballing” step was carried out to broaden our search: articles listed in the references of the

works we initially retrieved were also included in our search.

(iii) Select pertinent literature – the outcome of the previous activities, without considering

duplicates, was a set of 183 eligible papers, which were then filtered according to the inclusion

and exclusion criteria shown in Table 1. All the inclusion criteria had to be jointly true to select a

paper, while one exclusion criterion was sufficient to reject it. The selection process followed

three steps. First, works were filtered based on their title (resulting in 122 eligible papers).

Second, a further refinement was made reading the abstract (coming to 81 eligible papers).

Third, the remaining papers – whose abstract was not sufficient to understand the alignment with

the scope of the analysis – were read in their entirety. Finally, 59 papers were selected for in-

depth examination.

Phase 2: Descriptive analysis

The 59 publications we reviewed are from a number of different sources: 86% are articles

published in scientific journals, while the remaining 14% come from conference proceedings. Two

main considerations may be derived from this preliminary analysis. First, despite the novelty of

the theme – for which knowledge is still at early stages and not consolidated in literature – the

majority of the papers come from journals. This is representative of how the academic community

– considering both authors and editors – agrees on the high significance of the topic. Second, both

the journals and the conferences are characterised by a high heterogeneity in terms of the core

theme they address, including industrial engineering, hospitality management, marketing, logistics

and human resources management. This is coherent with the variety of both topics and disciplines

linked to ODFD. Considering the journals, the sources are diverse. Except from the Journal of
Retailing and Consumer Services, which has 5 publications, there are 10 journals with 2 papers

each and the remaining 26 journals count 1 paper only. Among the journals with 2 works, the

majority belongs to the technology and engineering field (Journal of Advanced Research in

Dynamical and Control Systems, Journal of Theoretical and Applied Information Technology,

International Journal of Advanced Science and Technology, International Journal of Innovative

Technology and Exploring Engineering, International Journal of Recent Technology and

Engineering), followed by food and hospitality management (British Food Journal, International

Journal of Contemporary Hospitality Management, Journal of Foodservice Business Research),

logistics (International Journal of Logistics Research and Applications) and social sciences (Work

Organisation, Labour and Globalisation). An aligned by-topic distribution (majority in the

technology and engineering field, followed by food and hospitality management, then logistics

and social sciences) is maintained also for those journals counting one paper each. ODFD is a

broad research theme that, due to the great variety of activities it requires, encompasses numerous

disciplines, very different from each other. Accordingly, extant research on the topic is fragmented

and dispersed, and results are presented in conferences and journals from fields that barely benefit

from mutual contributions

Figure 1 illustrates the uptake of ODFD-related works in the academic discussion based on their

year of publication. Two insights clearly emerge. First, the novelty of the theme: there are no works

prior to 2015. This is well aligned with the newness of this business, for which the majority of the

leading companies were founded after 2013. Second, the rapidly growing interest of the academic

community towards ODFD. Specifically, nearly 90% of the works we identified in our review

were published since 2018. Despite being quite limited, the emerged time frame (6 years) may be

considered a suitable period for a literature review in the SCM field. First, the studied phenomenon
is very recent. Second, the number of analysed contributions is significant and aligned to that of

other literature reviews in the SCM field covering wider periods (e.g., Perego et al., 2011). Third,

the methodological paper by Seuring and Gold (2012), which analyses the characteristics of 22

SCM literature reviews, shows that only 6 papers (27%) targets a period higher than 7 years.

Figure 1: Timeline of the papers

Considering the geographical dimension of the reviewed works, our analysis shows how ODFD is

considered an interesting research theme globally and, in addition, many publications are the result

of international collaborations. However, particular attention towards the topic has been devoted

to use cases in Asia (e.g., China, Thailand, and India). Especially Indian use cases are covered by

31% of the contributions. The high diffusion of ODFD services in India and the associated

concentration of academic work is partially driven by a deep-rooted tradition of so-called

“Dabbawala” systems in the country, i.e., the delivery of prepared meals to people working away

from home (see, e.g., Ganapathy et al., 2016). This well-established habit may be considered as

the pre-digital ancestor of the current ODFD, which has accordingly been more easily adopted in

this context.
Phase 3: Category selection

The content-based review was performed– consistently with the stated research questions –

based on a twofold perspective.

(i) First, we distinguished works based on which of the actors involved in the ODFD

ecosystem were considered, i.e., platforms, restaurants, riders and customers (RQ1)

(ii) Second, we distinguished the works according to the value-creating activities and

processes that were investigated (RQ2)

The classification categories – detailed in the following – were selected upfront, thus relying on a

(mainly) deductive approach. The choice to define categories “a priori” is supported by different

methodological papers about literature reviews for SCM (i.e., Seuring and Gold, 2003; Tranfield,

2003). This deductive mode is beneficial when attempting to classify such dispersed and diverse

contributions, in a broad and novel research field where no previous attempts had been made in

this direction. An inductive approach has been adopted as a second step to revise the previously

identified categories, highlighting the peculiarities of the ODFD. With regard to the two main

views identified with the deductive approach - i.e. the actors involved and the value chain

activities – we can state that mapping the actors involved in the business is a common practice in

SCM literature, as it allows to better comprehend the relationships among them (Kaihara, 2003),

whereas the value chain – similarly to previous works addressing business ecosystems – served

“as the core mental model in order to explore the value creation process within companies and

across their network of relations” (Nucciarelli et al., 2017).

The two views were derived combining and elaborating literature addressing “traditional”

business strategy and literature in the platforms/business ecosystems fields.

Considering perspective (i), the ODFD actors were identified taking as a baseline the scheme of

business ecosystems proposed in most of the seminal works in the field, which is represented in

Figure 2 (Mikkola and Skjøtt‐Larsen, 2006).

Figure 2: Actors of platform ecosystems. Source: Mikkola and Skjøtt‐Larsen (2006)

The main feature of platforms, which differentiates them from traditional business models, is that

they create a two-sided market, in which the value does not flow unidirectionally (left to right),

but moves on both sides (Iansiti and Levien, 2004). In fact, the value creation entails the

participation of multiple users (Eisenmann et al., 2006), and more specifically platforms,

suppliers and customers. Applying this model to ODFD systems, two main differences emerge

compared to the reference one. First, suppliers were identified as the restaurants. Second, a

fourth significant actor was added: the riders (i.e., the people in charge of delivering the meal

from the restaurants to the customers).

Considering perspective (ii) – i.e., value creating activities and processes – the classification axes

were defined relying on academic works aimed to revise a “traditional” business strategy model -

like Porter’s value chain - in the light of platforms literature. According to Porter (1985), the

activities that allow firms to create value can be grouped in two main categories: “Primary

activities” (Inbound and Outbound Logistics, Operations, Marketing and Sales, Service) and

“Support activities” (Firm Infrastructure, Human Resource Management, Technology

Development, Procurement). Despite Porter’s value chain was originally developed to fit one
single firm, literature shows later that this model, if conveniently revised, can be applied to wider

contexts. Among others, Di Martinelly et al. (2009) switch from a company-centric view to a

supply chain one, implementing Porter’s model to the whole supply chain. A subsequent work by

Walsh (2011) – focused on developing nations – makes a further extension applying the model to

society. According to different academics, the value chain is still a reliable tool to investigate

also platforms/business ecosystems, but it needs to be appropriately adapted based on the new

way value is created in ecosystems: as anticipated, the value does not move unidirectionally from

the left side – supplier – to the right one – customers, but it flows on both sides (Eisenmann et

al., 2006), as represented by the arrows in Figure 1. As a result, when addressing platforms,

Porter’s value chain needs to be adapted. More specifically, the main change to be made

concerns the list of value-creating activities to be considered, which are: Marketing and Sales,

Human Resources Management, Technology Development, Logistics and Operations (Van

Alstyne et al., 2016).

Based on these premises it is possible to state that, despite the deductive nature of the category

selection process, the first version of the category list was revised and modified following an

inductive sub-step (aimed to catch the peculiar features of the ODFD business). Considering

actors, the main change relied in the introduction of a fourth player in the ecosystem, namely the

riders. Considering activities, the traditional set of activities proposed by Porter (1985) was

modified in the light of ODFD ecosystems and only the relevant ones have been selected. This

combined process, where the deductive category definition has been followed by an inductive

category revision, is supported by the methodological paper by Seuring and Gold (2012), who

recommend a first deductive category building step, followed by iterative cycles of inductive

category refinement.
Phase 4: Material evaluation

After the descriptive analysis, the articles were classified using a two-pronged approach. On the

one hand, they were categorised based on the research method(s) adopted by the author(s). On

the other hand, they were classified and analysed based on their content, relying on the

categories defined in Phase 3, to highlight significant themes and trends and to identify research

gaps.

The analysis was performed according to the following process: first, 10 papers were jointly

classified by all the three authors to get to an agreement on both the two dimensions (i.e., actors

and activities) and the way in which the classification should be performed. The remaining

papers were subsequently independently analysed by the authors, whose percentage agreement

was 1 for the methodology, 0.95 for the actors and 0.94 for the activities (the values are good

according to Wowak and Boone (2015)). Agreement was considered achieved in case all the

three authors had obtained the same result. Those papers for which there was not initial

agreement were jointly discussed and classified again by the three authors together, until a

consensus was reached.

Research methods used in the reviewed works

After the descriptive analysis, articles were reviewed based on the research method(s) adopted by

the authors. According to Meixell and Norbis (2008), seven main research methods are generally

applied in academic research: analytical models, conceptual models or framework, case studies,

interviews, surveys, simulations, and others. In the light of this high-level classification, three

main clusters of papers can be identified in our review: quantitative models (analytical models
and simulations), empirical analyses (surveys, interviews, and case studies) and conceptual

models or frameworks (which also include literature reviews). Table 3 shows the classification of

the found papers based on these clusters. In case multiple methods were applied, the work was

accounted for in all the related categories. Most of the papers rely on just one method (e.g.,

Wang and Somogyi, 2018), while some of them are multi-method (e.g., Jacob et al., 2019).

Research method Number of Contributions

Quantitative model 14
Empirical analysis 35
Conceptual model 10
Other 5
Table 3: Research methods

Considering empirical analyses, the majority is represented by surveys. In general, they take the

customers’ perspective and aim to identify the main sources of value in the use of ODFD

platforms. Vinaik (2019) surveyed 300 people to understand their attitude toward ODFD apps,

focusing on their awareness level of the services, their expectations and the dimensions

according to which they compare different players. The work by Gilitwala and Nag (2019)

presents the results of another survey on 400 clients of a ODFD platform operating in Bangkok,

to understand the factors influencing their willingness to repeat the purchases with the same app.

Beside surveys, other authors also present case studies (e.g., Furunes and Mkono, (2019)

investigate the peculiarities related to the application of the sharing economy model in ODFD).

Quantitative models are fewer than empirical ones (26%), and a great part of them propose and

solve an optimisation problem to efficiently and effectively manage the delivery of meals from

restaurants to customers. One of the main insights from our analysis is the diversity in the

considered objective functions. Some authors take the perspective of platforms and aim to

minimise their costs and/or to maximise the covered demand (Zambetti et al., 2017). Other
scholars focus instead on the riders, and try to maximise their personal income stemming from

the completion of delivery tasks (Li et al., 2018). Some works set instead objectives from the

customer’ perspective, e.g., prioritising the reduction of delivery delays (Fikar et al., 2018).

Finally, there are some authors proposing multi-objective functions, aimed at concurrently

considering the perspectives of different actors, due to the need to align their interests, which are

frequently conflicting (e.g., He et al., 2019). In general, the objectives of the models may be very

different; nonetheless, a recurring element (either as a part of the objective function or as one of

the constraints) is the service level, since customer satisfaction is key in determining the success

of a ODFD initiative (Furunes and Mkono, 2019).

Concerning the conceptual models, the majority is represented by frameworks aimed at

identifying the main factors influencing the ODFD market, and at investigating their mutual

relationships (see, e.g., the work by Thamaraiselvan et al. (2019), focussed on India).

Among the multi-method papers considered, there are two main clusters. Works belonging to the

first one combine frameworks and surveys. They typically develop frameworks based on the

literature, and then test the found hypotheses through questionnaires (see, e.g., Jacob et al.,

2019). Among the second – and quantitative – subset of multi-method works, the most frequent

association is between analytical models and simulations. Authors develop a model, that is

subsequently tested on simulated data (see, e.g., Steever et al. 2019).

Themes arising from the review

After the descriptive and the method-based analyses, the content-based review was performed –

as anticipated – based on the twofold actors-activities perspective. One of the outcomes of this
analysis is a framework (Figure 3), which represents value-creating activities next to the links

that connect the various actors involved for those activities.

Figure 3: Framework – ODFD actors and activities

The development of the framework as well as the association (and thus the position) of the

activities to the links were driven by a quantitative analysis, whose baseline is shown in Table 4.

This table reports, at the intersection of the actors and activities, the number of papers addressing

a specific activity with reference to a specific actor.

Total
Customer Rider Restaurant
(activity)
Marketing and Sales 27 0 1 27
Human Resources Management 1 14 1 15
Technology Development 9 7 4 14
Logistics 9 11 5 12
Operations 0 0 1 1
Total (actor) 36 26 11

Table 4: Classification based on themes

Considering for instance the first row (i.e., Marketing and Sales), 27 works investigate this

activity from the perspective of customers, 1 treat it relating to restaurants, while there are no

papers analysing Marketing and Sales for riders. As anticipated, this quantitative breakdown was

used to correctly associate the value-creating activities to the links among the actors. Marketing

and Sales was placed at the link between Platforms and Customers, Human Resources

Management between Platforms and Riders, and Operations between Platforms and Restaurants.

Technology Development and Logistics could instead not be uniquely associated to a specific

dyad of actors, but they lie at the intersection among all of them. Details concerning both the

actors and activities perspectives are presented in the following sections.

Two clarifications should be made about Table 4.

First, it must be noted that only three actors (i.e., Customer, Restaurant and Rider) are listed,

while the presence of Platforms is not explicitly stated: the reason behind this choice is that

Platforms were addressed in all the papers (as it could be expected being this a literature review

about ODFD).

Second, the total number of contributions addressing one activity, which can be read in the last

column (e.g., 27 papers for Marketing and Sales), may be lower than the sum of the numbers in

the row (e.g., 28 papers for Marketing and Sales, found as 27 for Customer/Marketing and Sales

+ 1 for Restaurant/Marketing and Sales). As a matter of fact, in the part of the table combining

the actors and activities dimensions, the same paper could be accounted for twice (or three times)

in case it includes the perspective of two (or three) actors concurrently (e.g., one paper tackling

Marketing and Sales was accounted for considering both the Customer and the Restaurant

perspectives). The same reasoning is valid for the number of works addressing the single actors:

the total number reported in the last row is lower than the sum of the different numbers in the
column, as the same paper could concurrently treat different activities from the perspective of the

same actor.

The next sections take the twofold actors-activities perspective and discuss the main findings of

the review, based on the results of the quantitative analysis reported in Table 4.

(i) Main findings: actors perspective

The last row of table 4, displaying the distribution of the papers based on the addressed actors,

shows how the highest number of papers (49%) is associated to the customers, followed by riders

(36%), while less than 20% consider the perspective of restaurants.

Platforms

ODFD platforms act as intermediaries between restaurants and customers. In the context of

ODFD services, it is possible to identify two main types of platforms, depending on the type of

services provided. First, the so-called aggregators, that arose at the beginning of the 21st century.

They provide customers with a technological solution by which a wide variety of restaurants can

be accessed, compared, and selected, and orders can be issued directly to the restaurants. They

do not provide logistics services, since the restaurants are directly in charge of arranging and

providing the delivery on their own (Zambetti et al., 2017). An example of these companies is

represented by JustEat. Beside aggregators, the second – and currently most diffused – type of

platform also offers the required logistics services, relying on their own logistics network

(typically based on freelance couriers) and provide the delivery in addition to the order

arrangement, issuing, and management (Fikar et al., 2018). Some examples are Deliveroo,

UberEats and Swiggy. Due to the wider range of activities they manage and the higher
complexities they have to tackle with respect to those faced by the first type of ODFD platforms,

this second type has gained the interest of academics.

Customers

They order the meal using the app and receive it at home. Despite being the final recipient of the

service, customers are typically treated as a part of the food delivery ecosystem in literature (see

for instance the work by Furunes and Mkono, 2019). As a matter of fact, their judgement about

the performances of the other actors in the ecosystem determines the success of an ODFD

initiative, thus reflecting on both the strategic choices and the management of daily operations

for platforms. This is manifest in most of the reviewed works that propose analytical models, in

which customers’ perspective is typically included either in the objective function – e.g.,

minimisation of their waiting time (He et al., 2019) – or as constraints to be met – e.g., target

service quality (Yildiz and Saverlsberg, 2019). Customers’ utility typically depends on three

dimensions that are by some means in a trade-off among each other: delivery effectiveness – in

terms of speed, timeliness and punctuality – (Zambetti et al., 2017), food quality (Suhartanto et

al., 2019) and price (Chandrasekha et al., 2019).

Riders

Riders are in charge of the delivery process, i.e., they collect the meals at the restaurant and

deliver them to the customers. They are typically freelance workers, operating in a crowdsourced

context: the app of the ODFD platform proposes a delivery task, and they may decide whether to

accept it or not (Goods et al., 2019). As in every crowdsourcing environment, it is fundamental

to consider the workers’ perspective in order to build a successful business. As a matter of fact,

on the one side, their unavailability – or the missing acceptance – to perform a task may
undermine the accomplishment of an on-time delivery to a customer (Dahle et al., 2017). On the

other side, a rider is “an individual who only cares about his own profit, instead of the overall

performance of the system” (Li et al., 2018, p. 2). Ignoring the way in which they operate could

thus result in a myopic view, potentially causing significant losses for the other actors in the

ecosystem.

Restaurants

Restaurants prepare the food to be delivered. Their role is key in different ways. First, they are in

charge of preparing the meal and therefore, the quality of the food – which is one of the main

factors affecting customers’ satisfaction and thus the intention to buy their meal online – depends

on them (He et al., 2019). Second, restaurants are responsible for the production planning. As a

result, they determine different constraints to the food delivery problem; among them there are

the production capacity, i.e., the number of orders that may be processed (and thus proposed to

customers), and the preparation time (Yildiz and Savelsbergh, 2019). Third, their position limits

the available options in defining the structure of the distribution network, as their location – i.e.,

the location of the points of origin for the last-mile delivery – has to be considered as fixed when

designing the network (Zambetti et al., 2017).

(ii) Main findings: activities and processes perspective

The last column of table 4 shows the distribution of the papers based on the addressed activities.

The activity associated to the highest number of papers is Marketing and Sales (27), followed by

Human Resources Management (15), Technology Development (14) and Logistics (12). Only one

paper was instead found addressing Operations.

 An in-depth activity-based analysis allowed to classify the papers based on the main addressed

topics. Results (in terms of topics, number of papers and main references) are shown in Table 5.

Activity Topic Number of papers Main references

Promotion of ODFD vs. Gunden et al. (2020);
Traditional food 14 Jacob et al. (2019);
Marketing and businesses Lee et al. (2017)
Sales Yeo et al. (2017);
Promotion among
13 Suhartanto et al. (2019);
different ODFD players
Ray et al. (2019)
Altenried (2019);
Positive/negative aspects 12 Christie and Ward (2019);
Human
Leonardi et al. (2019)
Resources
Goods et al. (2019);
Management
Intervention/regulation 3 Viossat (2019);
Richardson (2020)
Ilham et al. (2018);
App 8 Kapoor and Vij (2018);
Technology Fauzi (2019)
Development Correa et al. (2019);
Algorithm 6 Cosmi et al. (2019);
Hu et al. (2019)
Yildiz and Savelsbergh (2019);
Last-mile delivery 6 O'Neil and Hoffman (2019);
Li et al. (2019)
Pinto et al. (2019);
Logistics Distribution Network
4 Yildiz and Savelsbergh (2019);
Design
He et al. (2019)
Steever et al. (2019);
Batching 2
Fikar et al. (2018)
Operations Food preparation 1 Das and Ghose (2019)
Table 5: Value-creating activities

Marketing and Sales

Marketing and sales refer to all the actions taken to inform buyers about products and services,

and induce them to make the purchase (Porter, 1985). When considering ODFD, it includes all

the activities and initiatives aimed to push consumers to order their meals online, through the

app. This theme has been capturing the attention of academics, as shown by the significant

number of works aimed to identify the main factors affecting the customers’ intention to order
their meals online. These factors may be classified in two groups, depending on the perspective

taken by the respective authors in their research.

Promotion of ODFD vs. Traditional food businesses. The first perspective is broader and

investigates the drivers of convenience persuading consumers to choose ODFD over other –

more traditional – businesses, i.e., restaurants or grocery shopping. When the alternative is

represented by eating at the restaurant, the main benefits stemming from the use of on-demand

ODFD services are the access to a wider choice of restaurants and types of food without having

to move from home (Kapoor and Vij, 2018) and the possibility to avoid waiting time at the

restaurants before being seated and ordering food (Jacob et al., 2019). If the benchmark is instead

grocery shopping and subsequently cooking a home-made meal, the main benefit is ascribable to

the lack of time and effort needed to prepare the food (Bagla and Khan, 2017).

Promotion among different ODFD players. The second perspective is more narrowly defined

and focusses on the ODFD sector only: provided that customers have already decided to order

their meals online, authors investigate the factors driving the choice of which ODFD service they

will be using. When comparing different vendors, customers consider the variety of restaurants

among which they may choose (Bagla and Khan, 2017), the price (Chandrasekha et al., 2019),

the ease of payment methods (Vinaik et al. 2019), the speed of the service (Yeo et al., 2017), the

firm’s reputation (Gilitwala and Nag, 2019), and the absence of a minimum order value

(Thamaraiselvan et al., 2019). In addition, recent studies are showing that a successful initiative

in increasing customers’ loyalty to ODFD services (i.e., the intention to purchase again from the

same service) is the introduction of cashback and other reward programs (Vinaik et al., 2019).

Human Resource Management.

When referring to workers, ODFD platforms – as well as many other phenomena belonging to

the gig-economy – have been massively capturing the interest of academics, practitioners and

policy-makers. In the ODFD business, the management of workers – often referred to as ‘riders’

– follows the logic of ‘crowdsourcing logistics’: riders receive delivery proposals they may

decide whether to accept or not (Fikar et al., 2018). They “are non-salaried but nevertheless

highly dependent on digital platforms that pay them by the job” (Viossat, 2019, p. 63), and “as

independent contractors (they) engage in digitally-enabled and controlled work that is

remunerated on a piece rate basis” (Goods et al., 2019, p.502). This emerging working paradigm

has been rising numerous questions and driving the development of very different – and

sometimes conflicting – theories and views. Two main streams of works may be identified.

Positive/negative aspects. The first cluster of works discusses the positive/negative aspects of

crowdsourcing logistics, often comparing it to more traditional working paradigms. On the one

hand, considering the drawbacks of employment with ODFD services, it generates various health

and safety risks both for the workers and the other road users, due to the pressure on riders to

avoid delays and to the use of the phone while riding (Christie and Ward, 2019). On the other

side, this type of flexible work allows riders to benefit from autonomous organisation

(Briziarelli, 2019) and short-term sources of extra-incomes (Goods et al., 2019). Therefore, an

agreement about the theme has not been reached. Accordingly, some authors propose a dual

view, and look at logistics connectivity through an ambivalent lens: the digital connectivity

embedded in ODFD may be seen as a tool to supervise labour exploitation, as well as an

opportunity to establish new types of relations and to organise working activities (Leonardi et al.,

2019). Similarly, what may be asserted as labour precariousness can actually also be seen as

flexibilization of the jobs (Altenried, 2019).

Intervention/regulation. The second cluster of works proposes instead potential interventions that

should be pursued to correctly regulate and manage crowdsourcing logistics. Whatever their

opinion about the positive and negative effects of ODFD, academics agree in recognising that it

is now reconfiguring both urban spaces and labour conditions. Accordingly, there is a need for

new measures and legal frameworks, which adapt to current and future developments of this

emerging paradigm (Viossat, 2019), “the consequences of which need to be taken seriously by

regulators, scholars, workers and other relevant stakeholders” (Goods et al., 2019, p.502). The

main topics in this direction are labour regulations, urban freight policies an urban planning

choices.

Technology Development

Technology is a key theme when dealing with ODFD, due to the significance of the digital

component in this business: the technological element, and more in detail the rise of the Internet,

has been the enabler for the birth of ODFD. The literature exhibits two main angles from which

technology may be seen with reference to ODFD platforms.

App. The first is the app (i.e., mobile or web application) of the respective ODFD service. It is

the tool linking all the different actors of the ecosystem, and the mean of communication for

customers, restaurants and riders with the platforms. The literature shows both that it constitutes

the interface experienced by consumers when they decide to order their meal online, and how its

functionalities/features have a great impact on customer satisfaction. More specifically, the two

most impacting elements for the customers, which are addressed by the majority of the papers,

are the ease of use (Gilitwala and Nag, 2019; Ray et al., 2019) and the processing speed

(Pattnaik, 2019; Vinaik et al., 2019). In addition to them, two other characteristics influencing

customers’ intentions to use ODFD apps are the aesthetic design (Kapoor and Vij, 2018) and the
different payment options supported (Vinaik et al., 2019). Fewer papers focus instead on the app

features that are important to restaurants, and also in this case the ease of use and the processing

speed are the two key elements (Fauzi, 2019).

Algorithm. The second crucial technological element of platforms relies in the algorithm that

allows to manage the ODFD activities. It has three main purposes, which are strictly interrelated.

First, it is in charge of receiving customer orders: based on the delivery location, it allows

customers to issue orders only to those restaurants close enough to grant an acceptable delivery

time (considering both the preparation and the delivery time) (Yildiz and Savelsbergh, 2019).

Second, it assigns orders to the riders, who are then in charge of reaching the restaurant,

collecting the parcel and delivering it to the final customer. This is a very critical task, which has

to take into consideration the departing point of the rider, the position of the restaurant and that

of the customer (Huq et al., 2019). Third, it defines the routing, i.e., the route the rider has to

follow to reach the restaurant(s) and the customer(s) – one or more depending on the number of

orders that can be consolidated in one tour (Fikar et al., 2018). Most of the papers propose new

and improved versions of the traditional assignment-and-routing logic, and they recommend

enhancing the objective function to combine the perspective of platforms with that of the other

actors, i.e., restaurants (Yildiz and Savelsbergh, 2019), customers (Li et al., 2018) and riders (He

et al., 2019).

Logistics

As manifest in the scheme in Figure 3, logistics plays a fundamental role in the ODFD business,

as it constitutes the link among all the involved actors, including the platform itself. Papers

addressing logistics treat it from different angles, but three key domains may be identified: last-

mile delivery, distribution network design and batching policies.

Last-mile delivery. The majority of papers in the logistics field address the management of the

last-mile delivery, i.e., the delivery process through which riders collect meals at the restaurants

and deliver them to the customers’ doorstep. There is agreement among academics in

recognising how the last-mile delivery for ODFD has much higher complexities if compared to

more traditional vehicle routing problems with pick-ups/deliveries and time windows

(VRPPDTW by He et al. (2009)). Among the main pitfalls of the ODFD last-mile distribution

with respect to generic e-commerce parcels, there is the presence of strict requirements

pertaining to the time dimension. First, delivery lead times are very short, as the meal must

typically be delivered very quickly from the moment it is ready – often within 15 minutes (Allen

et al., 2018). Second, punctuality is key, since customers expect their order to arrive in the

selected time-window (Vinaik et al., 2019). In addition, the type of (produced, managed and

delivered) products is characterised by criticalities, as freshly prepared meals have a very short

shelf-life (just a few minutes, in many cases), and are often characterised by temperature

maintenance requirements (He et al., 2019). The delivery operations of ODFD services are thus

much more complex than those for non-food e-commerce parcels, where a rather conventional

Vehicle Routing Problem (VRP) has to be solved (Allen et al., 2018). In the ODFD case, the

underlying planning and optimisation problem is a variant of the dynamic Pickup and Delivery

Problem (PDP), in which (i) orders can often not be consolidated before starting the delivery tour

(Zambetti et al., 2017), and (ii) where there are rigid precedence constraints to be respected for

pickups and deliveries (i.e., the order pickup at the restaurant must happen before the delivery to

the customer, so that the restaurant needs to be reached before getting to the associated customer)

(Yildiz and Saverlsberg, 2019).

Distribution network design. The second – and more strategic – logistics domain addressed in

literature refers to the design of the distribution network, i.e., the choice of the number and

location of facilities supporting the distribution of goods. Also in this case, ODFD introduces

novel challenges compared to traditional e-commerce distribution. The points of origin of

individual deliveries are a large number of restaurants that – unlike traditional e-commerce hubs

– are dispersed in space and highly specialised (i.e., only one specific restaurant can be the point

of origin for a customer, due to its unique offer of freshly prepared meals and/or to the time

constraint for the delivery to be performed) (Bagla and Khan, 2017). On the demand side, one of

the peculiar characteristics of ODFD orders is the high intra-day seasonality of order volumes,

namely lunch and dinner peak times (Kapoor and Vij, 2018). This demand pattern requires a

tremendous flexibility that can be guaranteed only by crowdsourced work, as it allows engaging

riders for shorter time slots (Dahle et al., 2017). Given the position of restaurants (i.e., the pick-

up points), the most important choice to be made about the network is defining the demand area

to be served, and the subsequent planning of both the production and the delivery capacity

(Yildiz and Savelsbergh, 2019). Based on this first decision, two following choices are: first, the

location and the number of riders’ departing points (Zambetti et al., 2017). Second, the

possibility to adopt – and locate – transhipment points to transfer orders from one vehicle to

another, and to consolidate multiple shipments (Fikar et al., 2018).

Batching. The third, and least discussed, logistics topic concerns the batching policy, i.e., the

possibility to combine food coming from different restaurants in the same customer order

(Steever et al., 2019). Grouping different orders and assigning them to the same rider allows to

increase the number of deliveries in the same tour (as two or more customers belong to the same

delivery tour), thus having a positive effect on efficiency performances. Despite the potential
significant benefits that batching may entail, it is still not widely considered in literature, as it is

not suitable for all the contexts (Seghezzi et al., 2020). As a matter of fact, in order to have the

possibility to aggregate different orders, the number of orders need to be particularly significant

(i.e., the demand has to be high). It is otherwise very difficult to find two orders fulfilled by

restaurants near to each other and addressed to close customers in the same delivery time

window. This “reduces consolidation opportunities and imposes the need for more vehicles

operating simultaneously and executing shorter routes” (Reyes et al., 2018). Therefore, the

batching policy well fits contexts in which ODFD is already widespread, while there are several

countries – such as France or Italy (Pinto et al., 2019) – that are not mature and for which the

conditions for its introduction are still not met yet.

Operations

Operations refer to the activities transforming inputs into outputs (Porter, 1985). In the ODFD

context, this involves the preparation of the meal, of which the restaurants are in charge.

Differently from the other activities, literature counts few works addressing operations.

Accordingly, no clusters of papers were identified. Affecting both time performances and quality

of the delivered product, operations have a twofold role. First, the timeline presented by Yildiz

and Savelsberg (2019) and depicted in Figure 4 shows how operations are carried out between

the moment an order is placed by a customer and the moment the order is ready to be picked up

by the rider.
 Figure 4: Timeline of events. Source: Yildiz and Savelsberg, 2019

Preparation time has a huge impact on the overall delivery time seen by the customers, even if it

may not be immediately evident from this representation. As a matter of fact, this timeline

displays a situation in which the rider is not yet at the restaurant when the meal is ready. In this

case, there is a ready-to-pickup (RtP) interval that acts as a buffer: a longer preparation time does

not directly affect the overall delivery time, since the meal would not be able to leave the

restaurant anyway until the rider arrives. Nonetheless in reality, the RtP time is typically very

tight – usually no more that 5% of the overall click-to-door (CtD) time (Yildiz and Savelsberg,

2018). Often RtP time is null, as the riders may get to the restaurant before the meal is ready. In

this case, the preparation time is determinant in defining the delivery time seen by the customers.

Second, besides determining delivery times, restaurants are also responsible for the quality of

food, i.e., the output of the operations. Here, again, the impact of operations is huge, since

customers directly perceive and experience this performance dimension. Besides the quality of

the delivery service – which is crucial (Pattnaik, 2019) – food quality is one of the key factors

driving customers’ choice to rely on ODFD services (Suhartanto et al., 2019).

Phase 5: Interviews with practitioners

To complement our review of the literature, interviews were performed with 4 practitioners

working for 4 different multinational ODFD companies. Interviewees were selected based on a

twofold criterion. Similarly to Huscroft et al. (2012), first, potential participants were identified

among volunteers from previous research efforts and references from senior logistics

professionals. Second, practitioners had to belong to the senior management, and not to a

specific function (they are: 1 logistics manager, 1 marketing manager, 2 general managers); this
allowed to gain a wider perspective on the ODFD business, avoiding potentially misrepresented

or partial views.

The involvement of the managerial community in academic works is increasingly diffused in the

supply chain management field, especially in the logistics one, as it helps granting practical

significance of research for managers. The role of practitioners in literature review papers may

be very different: they may contribute in the definition of the keywords to be used for the search

(e.g., Huscroft et al., 2012); they may help identify elements for the emerging framework (e.g.,

Elbarkouky and Abdelazeem, 2013); according to an emerging trend for literature reviews –

which we adopted in this work – practitioners can also give a significant contribution in

validating the outcomes of the review (Mangiaracina et al., 2019) and interpreting them from a

managerial perspective (Bardauskaite, 2014).

Participants were first individually interviewed, deep diving both the framework and the sub-

topics related to each activity: the results of the literature review were shown and discussed.

These semi-structured interviews allowed to both validate the outcomes (i.e., showing that the

most significant actor and activity-related topics are those emerging from the literature) and

better read and interpret them (Harland et al., 2019). A group interview was subsequently

performed, in which all the practitioners discussed together under the guide of a moderator.

Group interviews benefitted from the participants' simultaneous interviewing, which allowed to

combine and stimulate their mutual contribution (Urciuoli and Hintsa, 2017). This was mainly

aimed to discuss the gaps from a managerial perspective, and to introduce potential sparks for

further research (e.g., analysing the introduction of dark kitchens).

Summary, gaps and future directions

ODFD is a broad research theme that, due to the great variety of activities it requires,

encompasses numerous disciplines, very different from each other. Accordingly, extant research

on the topic is fragmented and dispersed, and results are presented in conferences and journals

from fields that barely benefit from mutual contributions. In addition to identifying the key

themes, our review of the extant literature reveals the main shortcomings of extant academic

literature, and leads us to propose fruitful avenues for future research. Aligned to the literature

review by Mangiaracina et al. (2019), gaps can be identified on three different levels: (i) lack of

a comprehensive view integrating the different topics, (ii) presence of topics that – despite their

significance for both academics and practitioners – are under-investigated, and (iii) presence of

topics that – even if discussed by a higher number of papers – need to be studied in further depth.

Gaps (i) and (ii) may be derived from the analysis summarised in Table 4, gaps (iii) are instead

referred to that displayed in Table 5. We will comment on each of these in the following.

Lack of an integrative view. Even if some efforts are starting to be made to combine and include

the perspectives of the different actors involved, there is still great room for works aimed at

integrating them, trying to move towards an ecosystem-based view. This is manifest if

considering that only 3 papers out of 59 include all the three actors, and just 6 out of 59 combine

the perspective of two players. In this direction, literature could benefit from both qualitative

works (e.g., frameworks providing a complete picture of the ecosystem trying to explain the

relationships and the dependencies between the actors involved) and quantitative ones (e.g.,

analytical model with multi-objective functions that take into account the needs of all the players

involved and impacted by the decisions).

Under-investigated topics. Some themes are under-investigated, and this is true concerning the

actors and the value-creating activities. Focussing on Table 4, it is possible to decline this gap
from the perspective of both actors and activities. Considering actors (i.e., looking at the last row

of the table), the perspective of restaurants (treated by 19% of the works) results to be

significantly less studied if compared to both customers (61%) and riders (44%). Considering

activities (last column of the table), the analysis clearly shows how Operations have been barely

taken into consideration by academics (only one paper was found addressing them). Despite its

importance with respect to the core topic of the work, this activity – i.e., the preparation of fresh

meals – is typically just cited and treated as an ancillary. Evidences coming from our interviews

with practitioners confirm an urgent need for a deeper analysis of the perspective of restaurants,

and more specifically in the field of operations (food preparation above all), due to their crucial

impact on ODFD performance.

Topics requiring more in-depth investigation. Considering Table 5, three main fields were

identified as requiring higher attention: one concerning Human Resources Management

(Intervention/regulation) and two concerning Logistics (Distribution Network Design and

Batching). The reason behind the identification of these topics is twofold. On the one hand, the

high significance of the related value-creating activities; on the other hand, the low coverage of

the mentioned sub-topics.

Starting from Human Resources Management, the review of the literature showed how the

majority of the authors devoted their efforts to critically analyse the working conditions of the

riders, and to identify the positive and negative aspects of such a novel employment paradigm.

Nonetheless, only few of them made a step forward in proactively proposing and evaluating

potential solutions that could improve them, mitigating the criticalities of the gig-economy

consequences for riders. In this sense, the literature review suggests two potential directions

towards which new research could be aimed. On the one hand, legal matters about job and
employment regulations, which are also gaining the interest of both the public opinion and the

press. On the other hand, measures concerning urban planning choices, aimed to address and

manage the presence of riders in the cities, considering both the traffic and the safety conditions.

Switching to Logistics, the first domain relates to the strategic choices linked to the design of the

distribution network. As illustrated, most of the works in this field focus on the definition of the

points where the riders should wait before moving towards restaurants. Further research efforts

should instead study the location-allocation problem concerning restaurants, i.e., define the

optimal set of restaurants to target, and define the demand area to be covered by each of them

(i.e., combination of Logistics with Restaurants). In addition, new works should be steadily

developed to keep up with rapid industry innovation that characterises the dynamic ODFD

sector, which may also affect the design of the network. Among them, there are the so-called

dark kitchens, i.e., cooking areas shared by different restaurants and dedicated to the ODFD

business (Karamshetty et al., 2020). The introduction of dark kitchens changes the delivery

problem, since they concentrate different points of origins (i.e., restaurants) in the same location

with a dedicated and a bigger production capacity. As a result, both academics and practitioners

would benefit from works aimed to identify the optimal location-allocation of orders to dark

kitchens.

The second logistics topic to be addressed is instead related to the batching policy. As already

mentioned, the number of contributions in this field are few. As a matter of fact, being able to

consolidate different customer orders in the same delivery tour requires a significant demand

level (so that it is possible to combine orders for which the points of origin, the points of

destination and the delivery time window are compatible). While this is not the case of less

mature ODFD markets (in which the delivery density does not allow to consolidate orders), this
configuration gets increasingly interesting as the market is more diffused. Accordingly, these

contexts would deserve closer consideration from a go-to-market perspective, and the drivers of

profitability of ODFD services should be investigated in greater depth. In addition, further works

should investigate the effect that variations in key variables and parameters would have on the

found outcomes in a generic environment (e.g., analysing the impact of changes in the density of

customers and restaurants, the availability of riders) to derive principles of general validity.

Conclusions

In this work we reviewed 59 papers, including both works published in scientific journals and

presented at international conferences. After considering their general characteristics – i.e., year,

source, and country – the contributions were systematically analysed following a two-pronged

approach. On the one hand, the research method(s) adopted by the author(s) were identified. On

the other hand, the papers were analysed based on their content. Specifically, a classification was

built based on the integration of two dimensions. First, the involved actors, i.e., platforms,

restaurants, riders, and customers (thus answering RQ1). Second, the main activities and

processes, which allow to create value in the ODFD ecosystem: Marketing and Sales, Human

Resources Management, Technology Development, Logistics and Operations (thus answering

RQ2). Based on this classification, the extant literature was presented, some key themes were

derived, and the main shortcomings of the extant body of literature were outlined.

This work offers insights to both academics and practitioners. On the academic side, it analyses

and classifies the relevant literature on on-demand ODFD services, proposing directions for

future research activities based on the identified gaps. On the managerial side, it presents a
framework associating the main activities to be performed in a ODFD context to the

corresponding link between the relevant actors.

This work has two main limitations. First, it may not be considered as all-inclusive in terms of

analysed contributions, since some works could have inadvertently been missed. Nonetheless –

also thanks to comprehensive interviews with practitioners we conducted – we are confident that

the general insights and conclusions from our review are reliable, and that the presented results

are representative of the current state of research on the topic.

Second, activities and processes in the framework are reported just once, in correspondence to

the link between the two actors that each activity most significantly relates to. However, this

association is not always univocal. It could happen that an activity impacts the link between

more than just one pair of actors (e.g., marketing activities could be performed by platforms

towards new restaurants as well as towards new customers). The choice of proposing a slightly

simplified view on value-adding activities was motivated by two main considerations. First, this

simplified view makes the framework structuring our review clearer and more concise in

representing both the main logic behind ODFD businesses and the recent academic research. On

the other hand, it still provides an integral view of the main activities and links, without

neglecting any significant issue. Future works could be aimed at extending this generic

framework and tailoring it towards specific variants of ODFD ecosystems, highlighting the

additional roles activities may have for some specific actor pairs in these ecosystems.

References

Allen, J., Piecyk, M., Piotrowska, M., McLeod, F., Cherrett, T., Ghali, K., and Wise, S. (2018),
“Understanding the impact of e-commerce on last-mile light goods vehicle activity in
 urban areas: The case of London”, Transportation Research Part D: Transport and
Environment. Elsevier, 61(June 2017), pp. 325–338.
Altenried, M. (2019), “On the last mile: logistical urbanism and the transformation of labour”,
Work Organisation, Labour & Globalisation, Vol. 13 No. 1, pp. 114-129.
Bagla, R. K., and Khan, J. (2017). “Customers expectations and satisfaction with online food
ordering portals”, Prabandhan: Indian Journal of Management, Vol. 10 No. 11, pp. 31-
44.
Bardauskaite, I. (2014). “Loyalty in the business-to-business service context: A literature review
and proposed framework”. Journal of Relationship Marketing, Vol. 13 No. 1, pp. 28-69.
Briziarelli, M. (2019), “Spatial politics in the digital realm: the logistics/precarity dialectics and
Deliveroo’s tertiary space struggles”, Cultural Studies, Vol. 33 No. 5, pp. 823-840.
Cerchione, R., Singh, R., Centobelli, P., and Shabani, A. (2018), “Food cold chain management:
From a structured literature review to a conceptual framework and research agenda”,
International Journal of Logistics Management, Vol. 29 No. 3, pp. 792-821.
Chandrasekhar, N., Gupta, S., and Nanda, N. (2019), “Food Delivery Services and Customer
Preference: A Comparative Analysis”, Journal of Foodservice Business Research, Vol.
22 No. 4, pp. 375-386.
Christie, N., and Ward, H. (2019), “The health and safety risks for people who drive for work in
the gig economy”, Journal of Transport & Health, Vol. 13, pp. 115-127.
Dahle, L., H. Andersson, and M. Christiansen. 2017. “The Vehicle Routing Problem with
Dynamic Occasional Drivers.” In International Conference on Computational Logistics,
Southampton, October pp. 49-63.
Durach, C. F., Kembro, J., and Wieland, A. (2017), “A new paradigm for systematic literature
reviews in supply chain management”, Journal of Supply Chain Management, Vol. 53
No.4, pp. 67-85.
Eisenmann, T., G. Parker, M. Van Alstyne. 2006. “Strategies for two-sided markets”, Harvard
Business Review Vol. 84 No. 10, pp. 92–101.
Elbarkouky, M. M. G., and Abdelazeem, G. (2013). “A green supply chain assessment for
construction projects in developing countries.”, WIT Transactions on Ecology and the
Environment, Vol. 179, pp. 1331-1341.
Fikar, C., Hirsch, P., and Gronalt, M. (2018), “A decision support system to investigate dynamic
last-mile distribution facilitating cargo-bikes”, International Journal of Logistics
Research and Applications, Vol. 21 No. 3, pp. 300-317.
Furunes, T. and Mkono, M. (2019), “Service-delivery success and failure under the sharing
economy”, International Journal of Contemporary Hospitality Management, Vol. 31 No.
8, pp. 3352–3370
Ganapathy, V., Mahadevan, P., and Ravikeerthi, J. V. (2016), “An Empirical Study of the
Feasibility of Introducing the Mumbai Dabbawala Food Delivery System in Bangalore”,
SAMVAD, Vol. 12, pp. 9-22.
Gilitwala, B. and Nag, A. K. (2019), “A study of factors influencing repurchase intention of
online food delivery in Bangkok, Thailand”, International Journal on Emerging
Technologies, Vol. 10 No. 3, pp. 193–201.
Goods, C., Veen, A. and Barratt, T. (2019), ““Is your gig any good?” Analysing job quality in
the Australian platform-based food-delivery sector”, Journal of Industrial Relations, Vol.
61 No. 4, pp. 502–527
He, Z., Han, G., Cheng, T. C. E., Fan, B., and Dong, J. (2019), “Evolutionary food quality and
location strategies for restaurants in competitive online-to-offline food ordering and
delivery markets: An agent-based approach”, International Journal of Production
Economics, Vol. 215, pp. 61-72.
Huscroft, J. R., Hazen, B. T., Hall, D. J., Skipper, J. B., and Hanna, J. B. (2013), “Reverse
logistics: past research, current management issues, and future directions”, The
International Journal of Logistics Management, Vol. 24 No. 3, pp. 304-327.
Iansiti, M., and Levien, R. (2004), “Strategy as ecology”, Harvard business review, Vol. 82 No.
3, pp. 68-78.
Jacob, A. M., Sreedharan, N. V., and Sreena, K. (2019), “Consumer perception of online food
delivery apps in Kochi”, International Journal of Innovative Technology and Exploring
Engineering, Vol. 8 No. 752, pp. 302-305.
Kaihara, T. (2003). Multi-agent based supply chain modelling with dynamic environment.
International Journal of Production Economics, Vol. 85 No. 2, pp. 263-269.
Kapoor, A. P., and Vij, M. (2018), “Technology at the dinner table: Ordering food online
through mobile apps”, Journal of Retailing and Consumer Services, Vol. 43, pp. 342-351.
Karamshetty, V., Freeman, M., and Hasija, S. (2020), “An Unintended Consequence of Platform
Dependence: Empirical Evidence from Food-Delivery Platforms”.
Leonardi, D., Murgia, A., Briziarelli, M., and Armano, E. (2019), “The ambivalence of logistical
connectivity: a co-research with Foodora Riders”, Work Organisation, Labour &
Globalisation, Vol. 13 No. 1, pp. 155-171.
Li, M., Zhang, J. and Wang, W. (2018), “Task Selection and Scheduling for Food Delivery: A
Game-Theoretic Approach”, 2018 IEEE Global Communications Conference,
GLOBECOM 2018 - Proceedings. IEEE
Lim, S. F. W., Jin, X., and Srai, J. S. (2018), “Consumer-driven e-commerce: A literature review,
design framework, and research agenda on last-mile logistics models”, International
Journal of Physical Distribution and Logistics Management, Vol. 48 No. 3, pp. 308-332.
Lim, S. F. W., and Winkenbach, M. (2019), “Configuring the Last-Mile in Business-to-
Consumer E-Retailing”, California Management Review, Vol. 61 No. 2, pp. 132-154.
Mangiaracina, R., A. Perego, A. Seghezzi. and Tumino, A. 2019. "Innovative solutions to
increase last-mile delivery efficiency in B2C e-commerce: a literature review".
International Journal of Physical Distribution & Logistics Management, Vol. 49 No. 9,
pp. 901-920.
Martinelly, D. C., Riane, F., and Guinet, A. (2009), “A Porter-SCOR modeling approach for the
hospital supply chain”, International Journal of Logistics Systems and Management, Vol.
5 No. 3-4, pp. 436-456.
Meixell, M.J. and Norbis, M. (2008), “A review of the transportation mode choice and carrier
selection literature”, International Journal of Logistics Management, Vol. 19 No, 2, pp.
183-211.
Mikkola, J. H., and Skjøtt-Larsen, T. (2006), “Platform management: Implication for new
product development and supply chain management”, European Business Review, Vol.
18 No. 3, pp. 214-230.
Nguyen, D. H., de Leeuw, S., and Dullaert, W. E. (2018), “Consumer behaviour and order
fulfilment in online retailing: a systematic review”, International Journal of Management
Reviews, Vol. 20 No. 2, pp. 255-276.
Nucciarelli, A., Li, F., Fernandes, K., Goumagias, N. D., Cabras, I., Devlin, S., Cowling, P. and
Kudenko, D. (2017), “From value chains to technological platforms: The effects of
 crowdfunding in the digital game industry”, Journal of Business Research, Vol. 78, pp.
341-352.
Pattnaik, S. (2019), “Measuring service quality of food delivery apps-SITEQUAL”, Journal of
Advanced Research in Dynamical and Control Systems Vol. 11 No. 10, pp. 419-423
Perego, A., Perotti, S. and Mangiaracina, R. (2011),"ICT for logistics and freight transportation:
a literature review and research agenda", International Journal of Physical Distribution
& Logistics Management, Vol. 41, No. 5, pp. 457 - 483
Pinto, R., Zambetti, M., Lagorio, A., and Pirola, F. (2019), “A network design model for a meal
delivery service using drones”, International Journal of Logistics Research and
Applications, pp. 1-21.
Porter, M. E. (1985), “Creating and sustaining superior performance”, New York: Harvard
Business Review
Preetha, S., and Iswarya, S. (2019), “Factors Influencing the Intension to Use Food Online Order
and Delivery Appvia Platforms-Using Tam (Technology Acceptance Model)”,
International Journal of Recent Technology and Engineering
Ranieri, L., Digiesi, S., Silvestri, B., and Roccotelli, M. (2018), “A Review of Last Mile
Logistics Innovations in an Externalities Cost Reduction Vision”, Sustainability, Vol. 10
No. 3, pp. 782.
Ray, A., Dhir, A., Bala, P. K., and Kaur, P. (2019), “Why do people use food delivery apps
(FDA)? A uses and gratification theory perspective”, Journal of Retailing and Consumer
Services, Vol. 51, pp. 221-230.
Seuring, S., & Gold, S. (2012), “Conducting content-analysis based literature reviews in supply
chain management”, Supply Chain Management: An International Journal, Vol. 17 No.
5, pp. 544-555.
Steever, Z., Karwan, M., and Murray, C. (2019), “Dynamic courier routing for a food delivery
service. Computers & Operations Research, Vol. 107, pp. 173-188.
Suhartanto, D., Helmi Ali, M., Tan, K. H., Sjahroeddin, F., & Kusdibyo, L. (2019), “Loyalty
toward online food delivery service: the role of e-service quality and food quality”,
Journal of foodservice business research, Vol. 22 No. 1, pp. 81-97.
Thamaraiselvan, N., Jayadevan, G. R. and Chandrasekar, K. S. (2019), “Digital food delivery
apps revolutionizing food products marketing in India”, International Journal of Recent
Technology and Engineering, Vol. 8 No. 2, pp. 662–665.
Van Alstyne, M. W., Parker, G. G., and Choudary, S. P. (2016), “Pipelines, platforms, and the
new rules of strategy”, Harvard business review, Vol. 94 No. 4, pp. 54-62
Vinaik, A., Goel, R., Sahai, S., and Garg, V. (2019), “The study of interest of consumers in
mobile food ordering apps”, International Journal of Recent Technology and
Engineering, Vol. 8 No.1, pp. 3424-3429.
Viossat, L. C. (2019), “How will work look in the era of digital platforms? The challenges of a
new social contract”, Futuribles, Vol. 6, pp. 63-79.
Walsh, P. R. (2011), “Creating a “values” chain for sustainable development in developing
nations: where Maslow meets Porter”, Environment, Development and Sustainability,
Vol. 13 No. 4, pp. 789-805.
Wang, O., and Somogyi, S. (2018), “Consumer adoption of online food shopping in China”,
British Food Journal.
Wowak, Kaitlin D., and Boone, Christopher A. (2015), "So many recalls, so little research: A
review of the literature and road map for future research", Journal of Supply Chain
Management Vol. 51 No. 4, pp. 54-72.
Yeo, V. C. S., Goh, S. K., and Rezaei, S. (2017), “Consumer experiences, attitude and behavioral
intention toward online food delivery (OFD) services”, Journal of Retailing and
Consumer Services, Vol. 35, pp. 150-162.
Yildiz, B., Savelsbergh, M., (2018), “Provenly High Quality Solutions for Meal Delivery
Routing Problems”, Optimization Online.
Yildiz, B., and Savelsbergh, M. (2019), “Service and capacity planning in crowd-sourced
delivery”, Transportation Research Part C: Emerging Technologies, Vol. 100, pp. 177-
199.
Zambetti, M., A. Lagorio, and R. Pinto. (2017), “A network design model for food ordering and
delivery services”, XXII Summer School Francesco Turco, Palermo, September, pp. 1-7.