University of Groningen Faculty of Economics and Business

(1)

University of Groningen

Faculty of Economics and Business

MSc thesis: Technology and Operations Management

January 28, 2019

An extension of the CLSR-μhub in slum-like areas

A study in the context of the Physical Internet

Abstract

From a logistical perspective, last mile deliveries are one of the key challenges faced by shippers. In particular, within underdeveloped countries last mile delivery prob-lems have been and continue to be very persistent. Underdeveloped countries, as opposed to developed ones, are densely populated by small, family owned stores (nanostores) for which supply models are strongly inefficient. Drawing on the knowl-edge of the Physical Internet (PI), this paper aims to propose a system’s design, concerning an extension to the envisaged structure of the PI-system, by posing a μ-hub as an intermediate staging point in slums. By using Design Science Research (DSR), the μ-hub will stage solely fresh food, adhere strongly to safety requirements, use M-payments, and at last, be responsible to fulfil order placement. Accordingly, subtracted Critical Success Factors (CSF) illustrate that for increasing the systems’ successfulness, one shall need to involve local authorities, reduce tasks of the intra-favela transporters to the sole act of transporting and utilize end-of-life cooling containers to foster a cold chain enablement.

Keywords: Design Science Research, Physical Internet, last mile delivery, fresh produce, underdeveloped countries, nanostores

(2)

List of Figures

2.1 Supply models to nanostores . . . 13

8.1 Hierarchy diagram: A.dash.2 PI-based preliminary context functions . 28 8.2 Interaction diagram: A.dash.2 PI-based preliminary context functions 30 8.3 Hierarchy diagram: A.dash.1 Enable PI-based supply . . . 31

8.4 Interaction diagram: A.dash.1 Enable PI-based supply . . . 32

8.5 Hierarchy diagram: A.0 Reach nanostores via μ-hub . . . 33

8.6 Interaction diagram: A.0 Reach nanostores via μ-hub . . . 34

8.7 Hierarchy diagram: A.1 Manage safety requirements . . . 35

8.8 Interaction diagram: A.1 Manage safety requirements . . . 36

8.9 Hierarchy diagram: A.2 Perform last mile delivery . . . 37

8.10 Interaction diagram: A.2 Perform last mile delivery . . . 38

8.11 Hierarchy diagram: A.21 Hold, consolidate and pick-up from μ-hub . 39 8.12 Interaction diagram: A.21 Hold, consolidate and pick-up from μ-hub . 40 8.13 Hierarchy diagram: A.3 Place order(s) at supplier(s) . . . 41

8.14 Interaction diagram: A.3 Place order(s) at supplier(s) . . . 42

A.1 Supply model DSD . . . 60

A.2 Supply model via distributor . . . 61

A.3 Supply model via wholesaler . . . 62

B.1 Gatos in slums . . . 67

D.1 Code tree: Delivery specifications . . . 77

D.2 Code tree: Transport requirements . . . 77

D.3 Code tree: City hubs . . . 77

G.1 BPMN model: optimistic scenario, order placement by nanostore owner 83 G.2 BPMN model: alternative scenario, order collection by phone . . . 84

G.3 BPMN model: alternative scenario, order collection face-to-face . . . 85

H.1 Hierarchy diagram: Alternative diagram of phone ordering . . . 86

H.2 Interaction diagram: Alternative diagram of phone ordering . . . 87

H.3 Hierarchy diagram: Alternative diagram of face to face ordering . . . 88

(6)

List of Tables

6.1 List of nanostore and slum characteristics . . . 22

(7)

Abbreviations

BPMN Business Process Modelling and Notation

CR Critical Requirements

CSF Critical Success Factors

DC Distribution Center

DSD Direct Store Delivery

DSR Design Science Research

ETA Estimated Time of Arrival

FMCG Fast Moving Consumer Good

JIT Just In Time

LSP Logistic Service Provider

PDA Personal Mobile Device

PI, π Physical Internet

RQ Research Question

SLB Smart Locker Bank

UCG Urban Consolidation Center

(8)

Glossary

Decapsulation

The act of emptying the (PI, π)-container(s) to obtain ordered goods. Encapsulation

The act of filling matched (PI, π)-container(s) at the supplier sight, with goods ordered.

Informal authorities

Actors active in slums, which have any form of informal power. These are inter-changeably termed social non-state actors.

Intra-favela transporter

A resident of a slum that is employed by the μ-hub to perform the act of replenish-ment between the μ-hub and nanostores.

Macro-hub

A detachment centre, in which arriving containers are detached from other containers to be transported to their assigned μ-hub.

M-payment

Stands for Mobile-payments, indicating a transfer of (virtual) money by means of mobile devices.

Nanostores

Small family-owned stores, also known as mom-and-pop stores or kiranas. Physical Internet

An envisaged transportation system that is designed as a hub and spoke network. It functions on four defining pillars, namely openness, modularization, automation and decentralization.

(PI, π)-container(s)

A shipment device, which entails smart, autonomous features and enables activeness of an interlocking mechanism. This interlocking mechanism ensures similar shipment devices to be transported as a single unit.

μ-hub worker

A resident of a slum that is employed at the μ-hub to prepare shipments by the act of consolidation.

Slums

(9)

1 Introduction

Nowadays, cities all over the world are confronted with the consequences of the ever-expanding worldwide distribution system. These consequences consist of cost inefficiencies, high percentages of gas emission, increased pollution and high lev-els of congestion (Montreuil, 2011). As an act of reducing these consequences, the concept of the Physical Internet (PI, π) is introduced. When expected to be imple-mented in 30 years, it will be functioning on four key pillars, namely automation, decentralization, openness and modularization (Montreuil, Meller, & Ballot, 2012). However, in order for PI to be implementable, multiple parts of the concept need to be researched and developed beforehand.

One of these PI parts is the delivery over the last mile before reaching a final cus-tomer. From a logistical perspective, last mile deliveries are one of the key challenges faced by shippers. The reason for this is that urbanization has increasingly affected traffic congestion, narrow and blocked streets, cargo theft and the deficit between in-frastructure demand and supply (Kin, Spoor, Verlinde, Macharis, & Woensel, 2018). Especially in underdeveloped countries less attention has been oriented towards sup-ply solutions in the context of PI. To our knowledge, not a single thought is formed on how PI and its four fundamental pillars will be applied here. This is limiting, as last mile delivery problems have been and continue to be very persistent in these regions.

For instance, Mexico City is formed by a small population of city residents living in affluent neighborhoods, whereas the largest portion of the population live within extremely poor areas (e.g. slums, favelas, comunas) (Blanco, 2014). These foremost crowded and unsafe areas are densely populated by nanostores, which are low ca-pacity, family owned stores (Blanco & Fransoo, 2013). For acts of frequent stock replenishment singular suppliers have to travel long distances between Distribution Centers (DC) and nanostores, leading to insufficient supply models (Kin, Ambra, Verlinde, & Macharis, 2018). In many areas acting as the only food source, these nanostores offer overall highly processed, low-nutrition items with high sugar and fat levels (Ben´ıtez, Mej´ıa Argueta, Fransoo, & Salinas, 2018). This is partly due to infrastructural factors and available resources that do not support the supply of fresh produce in slums. One can state that the current presence of a push supply model is likewise restrictive in fostering the supply of fresh produce in slum-like areas.

(10)

(LSP). Customers can then use self-pickup to retrieve their goods, either by open-ing the container and pick-up their ordered goods or by bropen-ingopen-ing the whole (PI, π)-container home (Faugere & Montreuil, 2017). Utilizing this approach in slums could potentially reduce the distance between supply (DCs) and demand (nanos-tores) significantly, fostering a faster stock replenishment. An extension of the SLB, a so called City Logistic Smart Rack (CLSR)-μhub as proposed by De Vries (2018) could offer even more potential, as it provides the possibility for cross-docking (which might potentially reduce the number of daily stops at nanostores), for handling larger shipments and for assuring a safer container location.

Consequently, this research will propose a generic design, that for one exploits the current μ-hub features and secondly proposes an extension to the μ-hub. This will provide the following:

1. A foremost Just In Time (JIT) staging point, containing solely fresh produce as a supply for nanostores.

2. Safety features that can deal with the extreme characteristics of the application environment.

3. The usage of M-payments, in elimination of non-cash-based transactions. 4. The act of order placement, shifting from a push-based to a pull-based delivery

mode.

(11)

2 Theoretical Background

In this chapter, the generics on the Physical Internet (PI, π) is explained by means of its four defining characteristics. Moreover, it is described how last mile delivery in a nanostore network is organized nowadays. Subsequently, emphasis is put on the relevance of using PI characteristics in underdeveloped countries, in a general sense. Where-after, recent developments are provided to get insight in the previous con-ducted research on this topic. Afterwards, the theoretical background is concluded by putting emphasis on the specific research direction taken.

2.1 The context of PI

Environmental, economic and social challenges which arise from the logistical system as organized nowadays, are a constant concern for actors operating in this system (Montreuil, 2011). Some of these challenges, also referred to as symptoms of unsus-tainable current logistics, indicate a poor routing of freight, a poor usage of storing facilities, strangling of innovation, a poor synchronization with other transport types and high percentages of idle transport on a travel route (Montreuil, 2011). Conse-quently, in order to overcome these challenges, a novel concept for the organization of freight and logistics worldwide is initiated as a response. This so-called vision is referred to as the Physical Internet (PI, π), and expected to reach implementation in a timeframe of 30 years. Pan, Ballot, Huang, and Montreuil (2017) define the concept as: “A new concept of breakthrough innovation aiming to improve the eco-nomic, environmental and societal efficiency and sustainability of the way physical objects are moved, stored, realized, supplied and used worldwide.”. Accordingly, to get the outcome as stated in this definition, PI must function on its four defining characteristics. These so called pillars, can be defined as openness, modularization, automation and decentralization (Montreuil et al., 2012). As for them having their independent characteristics and functions in the PI-network, they also have impor-tant inter-linkages.

(12)

the pillar of modularization demands that physical entities are encapsulated in (PI, π)-containers, which are standard-sized, smart, modular and reusable entities of the system (Bektas, et al., 2015). Subsequently, (PI, π)-containers are provided with

fix-tures which allow individual containers to interlock, therefore generating itself into a storage structure and further easing a seamless transfer (Montreuil et al., 2012). Nevertheless, characteristics of modularization and features of (PI, π)-containers to interlock will not maintain without the pillar of decentralization. In explanation, this stimulates containers to manage their journey individually via decentralized control. To foster this process, a smart tag is added to each individual (PI, π)-container, which then inter-links with the pillar of automation. Through automation, a wide variety of handling, monitoring, storage and routing operations can be automated at an information and communication level (Montreuil, Meller, & Ballot, 2010). Such smart tagging then determines the ’best’ route towards the final destination of the physical object transported. For best route determination, routing functions in the manner of hubs and spokes, indicating point-to-point transport. Subsequently, when a (PI, π)-container passes a hub, new transportation protocols for each container can be applied, leading to a search for the best match between the required service and the following travelling step of the (PI, π)-container (Sarraj Rochdi, Ballot Eric, Pan Shenle, Hakimi Driss, & Montreuil Benoit, 2014). Consequently, in practical terms this would demonstrate itself in a combination of inter-modal transport and shared facilities (Sarraj Rochdi et al., 2014).

(13)

2.2 Last mile delivery in underdeveloped

coun-tries

The traditional retail channel makes up for many small family-owned stores, which serve between 100-500 customers in their direct neighborhood and which are called nanostores because of their limited size (Boulaksil & Belkora, 2017). Additionally, this limited size indicates limited product assortment and shelf space (Blanco, 2014). Blanco and Fransoo (2013) found that nanostores have a low barrier to entry, which aligns with their appearance. As a part of a residents’ house can easily be converted into a store (Blanco & Fransoo, 2013).

Currently, nanostores are estimated at an amount of around 50 million individual stores (Blanco & Fransoo, 2013). These stores remain viable due to numerous reasons. First, they play a significant role as a community hub, making them an important economic force (Ben´ıtez et al., 2018). Second, they are viable due to high income disparities within underdeveloped countries (Blanco, 2014). A large share of the population lacks access to public transport or cash (Blanco & Fransoo, 2013). As a response, nanostores offer a solution to this by operating on informal credit for unbanked inhabitants and by maintaining a small assortment. In a nanostore it is therefore no exception to buy a single cigarette or cookies in a two-pack (Blanco & Fransoo, 2013).

Besides the assortments being low in quantity, content wise 90 percent of sold prod-ucts contain durable, highly processed, low-nutrition items with high levels of fat and sugar (Ben´ıtez et al., 2018). The reason that few fresh products are available in these small stores is mainly due to an underperformance of the fresh food distribu-tion system in terms of cost, quality, sustainability and food security (Waldhauer, Van Der Burgh, Van Der Vorst, Bing, & Scheer, 2015). This is noticeable and caused throughout the entire chain. In Mexico City and Cairo alone it is determined that product quality suffers and food losses are major, because of a low transparency in the food chain, poor food packaging, rough handling and insufficient cooling (Waldhauer et al., 2015). In addition, for the current situation, adherence to com-munity demand is low. Therefore, Fast Mover Consumer Good (FMCG) suppliers maintain free power of pushing their products on nanostore shelves.

With regards to utilized supply models, Kin, Ambra, et al. (2018) addresses the existence of three of these models for nanostores which are found globally. These are visualized in figure 2.11_{. The first one they describe as direct store delivery (DSD),}

indicating that currently FMGC suppliers perform a direct supply to nanostores either from their own plant or from a nearby distribution center. The second one, which is most commonly used, indicates that supply to nanostores is performed by means of a regional distributor. Whereas in the last model, inventory of stores is replenished by means of a wholesaler which the store owners have to visit one self.

1_{To notify, the current process functions as the basis of improvement. Therefore, in fuller}

(14)

Furthermore, in the application of these supply models, the majority of goods are distributed by informal, lightweight vehicles, like motorized motorcycles and small trucks or vans (Kin, Ambra, et al., 2018).

Figure 2.1: Supply models to nanostores Source: (Kin, Ambra, et al., 2018)

Overall, these supply models have to cope with the following complexities. First, the close proximity in which these stores are located increases density, whereas the small assortment decreases store capacity. As a consequence, fragmentation in logistics increases. This fragmentation, in combination with the limited size of these stores, causes high frequent stops accompanied by small drop sizes. Second, many nanos-tores are located in difficult to reach neighbourhoods, due to an underdeveloped infrastructure. Therefore, transporters often struggle in finding a parking space and thus must park illegally (Boulaksil & Belkora, 2017). Third, in order to pay trans-porters for their delivery, cash is needed. However, because owners typically offer an informal credit to their customers managing cash availability becomes challeng-ing (Boulaksil & Belkora, 2017). Consequently, empty deliveries occur when cash quantities are insufficient.

2.3 Relevance of PI in underdeveloped countries

As noted, it is believed that the importance of implementing the key PI pillars in underdeveloped countries is to be called profound. This section will further elaborate on the actual relevance of defining PI requirements in these areas.

First, the vision of a PI-system is ideal when a worldwide application is considered, as a full architectural coverage is then achieved. Nonetheless, there are no thoughts on how the four key pillars will be applied in a violent, unsafe slum-like geographical area. Therefore, no ideas are yet formed on how to deal with the characteristics of a nanostore network, in the context of PI. This is limiting, as last mile delivery problems have been and continue to be very persistent in these regions. Accordingly, as an underlying objective of this research, taking such an extreme example to tackle is deliberate, as it might prove that outcomes in this research will pertain in any less extreme urban environment.

(15)

system. From an economic perspective, this illustrates itself in the fact that the flow of goods in these markets remain very costly. Next to inefficiencies in drop-offs, extra costs are generated when addressing to unsafe areas and frequent cargo theft (Vieira, Fransoo, & Carvalho, 2015). Taken measures occur in the form of extra vehicle insurance, overtime work, escort vehicles and replacement of stolen products (Vieira et al., 2015). Moreover, defining requirements of PI in underdeveloped countries would initiate a cheaper supply of goods to these markets.

Third, urban freight, in dense city areas, in underdeveloped countries make up for a disproportionate share of several externalities, such as congestion, pollution, greenhouse gases and infrastructural damage (Blanco, 2014), generating large en-vironmental issues. Likewise, extending the overall PI infrastructure towards un-derdeveloped countries forms an important tool in empowering a sustainable city logistics network. Helping to transport physical objects into, through and out of cities, while minimizing negative impacts on citizens quality of life in an efficient manner (Bektas, et al., 2015).

Fourth, access to fresh and healthy products is limited to customers. Extending the PI-framework to these underdeveloped countries fosters the expansion of the range of products accessible. In addition, the characteristics of PI can help foster a cold chain due to encapsulation in cooling containers. Being in line with the social perspective, implementation of PI in these areas nurtures the possibility of increasing the accessibility of objects that enhance quality of life and are valuable to the population (Montreuil, 2011).

Consequently, if requirements in terms of openness, modularization, automation and decentralization would be defined for these underdeveloped countries then the posi-tive notes as mentioned would be open for exploitation.

2.4 Recent developments in a PI context

Up until this point, the relevance of addressing PI in underdeveloped countries has been clarified. This section further elaborates, by first introducing recent develop-ments out of a PI-context to shortly illustrate general findings of last mile deliveries in a nanostore network. Next, research is addressed which covers architectural so-lutions in a PI-context.

First, out of the context of PI, research on improving last mile deliveries in un-derdeveloped countries have been mainly exploratory in nature. Findings in general illustrate that for high density areas deliveries over smaller distances offer more ben-efits in supply, therefore valuing the need of a closer orientation of supply (Boulaksil & Belkora, 2017; Kin, Spoor, et al., 2018). Despite these findings, researches lack in providing a general framework which addresses a solution that goes beyond ex-ploratory purposes.

(16)

delivery of e-commerce orders, while simultaneously mitigating the ”last mile de-livery problem”. Last mile dede-livery problems are mitigated, as the SLB exploits the idea of customer self-service. Instead of performing delivery to a customer’s house, pick-up by the customer is requested. In addition, the idea of including PI suggests the use of (PI, π)-containers instead of fixed lockers, therefore turning these lockers into smart ones (Faugere & Montreuil, 2017). When implemented, customers could either pick-up their ordered goods by opening the container and removing the content or by bringing the entire (PI, π)-container home (Faugere & Montreuil, 2017). SLBs can be beneficial to cities by reducing the city logistics flow, the number of failed deliveries and the amount of vehicles needed to cover a geo-graphic area (Faugere, Montreuil, & Stewart, 2017). Relating it to underdeveloped markets, utilizing this approach in slums could potentially reduce the distance be-tween supply (DC’s) and demand (nanostores) significantly, fostering a faster stock replenishment. Additionally, De Vries (2018) extented the SLBs architecture in her thesis. The extension was forwarded as City Logistic Smart Rack (CLSR)-μhubs. These CLSR-μhubs function as a cross-docking point, splitting medium shipments into small decapsulated shipments to the final customer. The extension De Vries (2018) proposed, included larger sized shipments to foster a B-to-B environment, decapsulation to reduce dead-weight transportation, usage of intra-city transporters and temporary stock keeping options. From the author’s performed research it is expected that μ-hubs will be advantageous.

Nonetheless, direct application of the CLSR-μhub in slum-like areas is problematic. This is the case as the base of this architecture is founded on the modern retail channel instead of the traditional one. Therefore, this architecture is unable to ad-just to requirements concerning security assurance, the different nature of payment procedures, parking to supply hubs in less developed infrastructures and the lack of available technologies. It is not excluded that some requirements of these models will be beneficial to underdeveloped countries. The CLSR-μhub offers advantageous opportunities to seize. Next to fostering a faster stock replenishment, it could offer a strong reduction of daily stops at nanostores (due to cross-docking functions), be attractive for fresh food suppliers (extending the range of supply), increase economic development of slum residents and potentially increasing the efficiency of supplier processes.

(17)

3 Design details

This research proposes a functional design on the operational process of the last mile delivery in any given slum. Whereas the operational process centres on the application of a μ-hub as an intermediate staging point. Accordingly, the μ-hub will be built on an application environment containing the highest extremes, as explained in section 2.3.

With the functional design in place it would be expected that the following adjust-ments will be obtained, while nevertheless the feasibility of these adjustadjust-ments will be investigated. The μ-hub is expected to provide:

1. A foremost Just In Time (JIT) staging point, containing solely fresh produce as a supply for nanostores.

2. Safety features that can deal with the extreme characteristics of the application environment.

3. The usage of M-payments, in elimination of non-cash-based transactions. 4. The act of order placement, shifting from a push-based to a pull-based delivery

mode.

Next to these μ-hub characteristics and preliminary to the design, some additional design details are presumed in advance. These are as follows.

(18)

4 Research Questions

Research question (RQ) 1: ”Which Critical Requirements (CR) and Critical Success Factors (CSF) are required, within the PI-system, to propose a generic architecture of a μ-hub’s operational process, applicable to slum-like areas?”

Theoretical analysis

(a) What are the current characteristics of a nanostore network that form the application environment?

(b) What are the current characteristics of slums, which use a nanostore network to foster supply and that form the application environment?

Practical analysis

(c) What opportunities can be identified in the current process of processing small shipments?

System Design

(d) Who are the stakeholders involved?

(e) What is the objective of each of these stakeholders?

(f) What are the Critical Requirements that can be deduced from the stakeholders’ objectives, which are necessary to design such an infrastructure?

(g) What does the functional architecture of the design look like?

(h) What are the Critical Success Factors that can be deduced from the functional architectures?

Validation

(19)

5 Methodology

In this chapter, the chosen research method is introduced in order to answer the research questions proposed. It is then reasoned why this specific research method was chosen, and how data will be collected, measured and analyzed.

5.1 The design of a generic architecture

The ultimate goal of this research is to identify CRs and CSFs and thereby answer

RQ 1. By means of answering RQ 1, the desired outcome of this paper can

be obtained. Namely, a generic architecture concerning the operational process of involving an extended μ-hub in slum-like areas, that by these means forms a utility and adds to the envisaged structure of the PI-system.

According to Hevner (2007), Design Science Research (DSR) is motivated by the desire to improve the environment by introducing new and innovative artifacts. Ac-cordingly, the ultimate design of this research strives to obtain this ”innovative artifact” making DSR a well qualified fit for this paper. Therefore, it can be empha-sized that the research performed will be qualitative, as no simulation or executable models will be included in delivering the final design. However, it is not excluded that the functional architecture might form a base on which such computational experiments might be built.

5.1.1 Theoretical analysis

In order to collect required data, an analysis is performed to answer the following sub-questions.

(a) What are the current characteristics of a nanostore network that form the application environment?

(b) What are the current characteristics of slums, which use a nanostore network to foster supply and that form the application environment?

(20)

alternative sources. This takes place in the preliminary phase to that of the design in which CRs are formed.

To highlight, in order to define these requirements, a literature review is performed. A literature review provides the opportunity to compare slums, that foster supply through a nanostore network from different continents based on their characteristics. By means of this review, multiple important characteristics can be distinguished, creating the possibility to extract drawbacks that form an understanding of the actual application environment.

Furthermore, validity is increased due to the comparison of multiple slums. By comparing slums from different continents and extracting characteristics that are perceived as most extreme, a broader perspective is taken into account. This ap-proach fosters the establishment of a design that is utmost generic and therefore applicable to various slum-like environments.

5.1.2 Practical analysis

However, this will only provide a clear understanding of the subtracted require-ments which evoke design decisions. Nonetheless, it won’t directly provide means in dealing with these decisions. Therefore, semi-structured interviews are performed at Fietskoeriers.nl, which is a Dutch courier service, delivering small shipments in dense areas by means of a bicycle. Here, it is not necessarily limiting in that the company is active in a different geographical area. The reason for this is that the company will function as a source of best practices. Their practical knowledge and experience might help providing proper argumentations in making design decisions. Therefore, the following sub-question will be answered.

(c) What opportunities can be identified in the current process of processing small shipments?

5.1.3 System Design

Subsequently, to get to the point of defining CSFs of the system. The requirements, as well as found opportunities should be evaluated simultaneously with the systems’ stakeholders. By defining the stake of each stakeholder involved, the system might better support the actors using it. Therefore, the following sub-questions should be answered next.

(d) Who are the stakeholders involved?

(e) What is the objective of each of these stakeholders?

(21)

(f) What are the Critical Requirements that can be deduced from the stakeholders’ objectives, which are necessary to design such an infrastructure?

According to van Aken, Chandrasekaran, and Halman (2016), up to this point the final product can be obtained, which they denote as an innovative generic design dealing with obtained field problems or opportunities. To get this final product, the last sub-questions must be tackled.

(g) What does the functional architecture of the design look like?

(h) What are the Critical Success Factors that can be deduced from the functional architectures?

5.2 The validation process

In order to illustrate validity, Hevner (2007) addresses that the DSR output must be returned into the environment for study evaluation. Furthermore, van Aken et al. (2016) add that in DSR validity is centered around effectiveness. They especially put a nuance on pragmatic validity, questioning whether the generic design which is received as an output of research, will indeed work after contextualization and implementation.

Because of the non-implemented state of the PI-system, actual observations of the utility added by means of this research are off. It is therefore not possible to fully validate the design after implementation. However, in order to still test on the extent of the pragmatic validity RQ 2 is formulated.

(22)

6 The application environment

This chapter fosters a detailed view on the application environment for which the final design is built, answering subquestion (a) and (b), chapter 4. By comparing characteristics of multiple slums in different continents, extreme characteristics can be distinguished. This is as according to the underlying objective as introduced in section 2.3, allowing for the generation of a design with a proper fit to its application environment.

6.1 Analysis on slums

In comparison of multiple slums in multiple cities worldwide, eventual subtracted requirements will provide a higher level of inclusiveness. This is the case, as the final architecture provided through this research will have more strongly defined generic characteristics. In comparison, multiple continents where slums are characterized by a nanostore network, were chosen to be reviewed. This led to the inclusion of Brazil, Mexico, Egypt and India.

By reviewing literature on nanostore networks and slums, different characteristics could be distinguished. Table 6.1 shows a summary of these characteristics. For a broader explanation and argumentation, appendix B can be consulted accordingly. Consequently, the most extreme comparisons are highlighted and illustrate which characteristics form the basic application environment, aligning with subquestion (c), chapter 4.

6.2 Company visit

(23)

Table 6.1: List of nanostore and slum characteristics

Country Brazil Mexico Eqypt India

Nanostore characteristics

Cold storage facilities Limited cold storage facilities Limited cold storage facilities Limited cold storage facilities No cold storage facilities

Payment options Cash-based Cash-based Cash-based Largely cash-based 10 percent accepts mastercards Fresh food availability Mostly holds

processed foods Fresh food access to farmers markets

Few fresh products Few fresh products Barely holds fresh food

Consumer awareness Buying behavior is connected to customer awareness

Level of education Low Low Low Low

Level of logistical knowledge Limited Limited Limited Limited

Technology used Personal mobile phone (often Xinglings) Personal mobile phone Mobile penetration is steadily increasing Personal mobile phone Personal mobile phone

Price of goods Higher then modern retail Higher then modern retail Higher then modern retail Higher then modern retail

Vehicles used Various small

vehicles

Small and not cooled, informal

Small and not cooled

Various small vehicles Slum characteristics

Road availability Heavily congested Heavily congested Lack of loading and unloading spaces Streets are chronically congested Poor city planning

Heavily congested Freight parking spaces are limited No integrated freight policy Lack of freight policy Lack of parking spaces

Cold chain A limited amount

of wholesalers use cold storage

Lack of cold chain infrastructure Internet connection Internet access is

increasing Access at CTC’s

Limited internet connection for the poor Access through Computer Kiosks No full internet coverage Access through Cybercaf´es Safety conditions Robbery, murder

and drug trafficking is common

High lack of safety High risk of freight theft

Unsafe and unplanned areas

High lack of safety Fear for robberies is high

Slums are controlled by an informal ”system of laws”

Corruption Two power sources

in slums: state actors and social non-state actors Corruption among local, legal authorities is common Corruption among local, legal authorities is common Corruption is a standard in slums Equipment failures Equipment is

exposed to more extreme environmental factors More extreme environmental factors are present

More extreme environmental factors are present

Equipment is exposed to more extreme environmental factors Power outages Subject to frequent

power outages Illegal connections and energy theft

Subject to frequent power outages Illegal connections and energy theft PET logistics flow A lacking PET

reverse logistic flow

(24)

7 Stakeholder objectives and CRs

This section will elaborate on all stakeholders involved in the optimistic and alter-native main design scenarios of this research, answering sub-questions (d), (e) and (f), chapter 4. These stakeholders are connected to the functioning and outcomes of the system to be, and therefore important generators of CRs to which the system should adhere.

7.1 The stakeholders

1. The suppliers

After suppliers receive an order request they become responsible for sending an order confirmation, for order processing and for registering the shipment through an online PI web-portal. Furthermore, the proposed design shifts the range of suppliers from FMCG suppliers towards fresh food suppliers. Suppliers, supplying the μ-hub at large, will be nearby farmers and fresh food producers. Moreover, due to the principles of PI any supplier can be utilized, therefore other supply sources are not excluded. However, prioritizing nearby sources will expectantly benefit reach and thereby shelf life, as well as foster economic value to the region.

2. The matchmaker

The matchmaker can be seen as a PI-system operator. Matchmakers search in [the PI-engine] for an available (PI, π)-container in [the PI-database]. Compar-ing the containers with the shipment registered for transport by the supplier, a suitable match is sought. After a match is found, details are communicated with the supplier.

3. The (PI, π)-container owner

PI-container owners are responsible for maintaining the availability and the condition of containers in the system. Due to the openness of the PI-system, container availability is known to the matchmaker mechanism. 4. The (PI, π)-container transporters

Container transporters are responsible for performing inter-modal transport and for loading and unloading of shipments in the PI-system. This physical distribution of goods will be carried out in between hubs that make up the route from supplier to final delivery destination.

5. The hub owners

(25)

rout-ing, with hubs at intermediate points. Inside hubs, PI-containers are option-ally interlocked to optimize further routing of containers involved. Within this paper the final hub that is located before the μ-hubs functions mainly as a detachment centre, it is therefore further denoted as a macro-hub.

6. The μ-hub owners

μ-hub owners offer the possibility to temporarily store goods in the μ-hub. They provide the services necessary to stage and consolidate goods in a proper manner. Offering this service includes, providing some sort of energy supply, data storage functions, connection to a broader network (internet) and online and offline security purposes.

7. The intra-favela transporters

Intra-favela transporters are responsible for performing the last mile from a designated μ-hub to the nanostores on their route. They pick-up shipments at the μ-hub and then perform deliveries decapsulated. Moreover, they are part of the slum community as their roots lay in the slum.

8. The μ-hub worker

μ-hub workers are residents of a slum and perform sole consolidation activities within the μ-hub.

9. The nanostore owner

Nanostore owners enable goods to reach their point of consumption by per-forming the act of selling. Additionally, nanostore owners are responsible for placing orders and maintaining product freshness.

10. The slum residents (the community)

Slum residents are considered to be nanostore customers. Due to the design purposes of this research their health might be increased, as their access to fresh products is enabled.

11. The Authorities

As is following from table 6.1, within slums it is common that two types of authorities are present. These authorities are generally the actors who seek compliance to rules which they have founded themselves. Subsequently, two types of authorities can be distinguished.

(a) Legal authorities

Legal authorities are statutory bodies with a source of legal power. Within slums these bodies are denoted as members of the police force. Due to high levels of corruption, these actors are often collaborating with other informal authorities active in slums.

(b) Informal authorities

(26)

7.2 Subtraction of CRs

Table 7.1 shows the stakeholders as mentioned in section 7.1, with their correspond-ing objectives and CRs.

Table 7.1: Stakeholder objectives CRs

Stakeholders objectives Explanation CRs

The suppliers

1. Make a profit by selling fresh products

(a) Enable efficient transport, by utilizing the μhub

Profitable Efficient (b) Decrease spoilage, by

utilizing μhub facilities 2. Reach the buyer in the

designated time-frame

Fast Reachable Responsive 3. Get the right products at

the right destination

(a) Enable a sufficient picking mechanism between fresh products, and corresponding destination Accurate Reliable Available Open (transparent) 4. Deliver products, at the

μhub in a fresh state

Fresh

The matchmaker

1. Match a suitable container to the requested shipments

(a) Collect real-time data on container availability and container specifications (b) Ensure container fits requirements, to ensure product freshness Accurate Reliable Available Open (transparent)

The PI-container owner 1. Make a profit out of each PI-container utilized

(a) Keep PI-containers in the right condition

Profitable Secure (b) Stimulate fast return of

empty containers in the PI-system

Safe Authorized Sustainable (c) Prevent containers from

abduction

(d) Prevent containers from unauthorized handling

(e) Recycle or sell containers, which are end-of-life

2. Keep (different type of products) fresh within the PI-container

(a) Keep availability of

sufficient containers up to date (b) Maintain right conditions inside the container

Fresh Availability

(27)

Table 7.1 continued from previous page

1. Make a profit out of each shipment carried out

Profitable

2. Reach the (intermediate) hub as fast as possible

(a) Respond accurately to communicated ETA’s

Fast Reachable Responsive 3. Ensure a safe PI-container

handling and transport

Safe Reliable The (macro) hub-owners

1. Let containers pass the (macro) hub as soon as possible

(a) Collect real-time data on ETA of containers

(b) Provide PI-transporter with real-time ETA of container

Accurate Responsive Open

(transparent)

2. Ensure safe container staging

(a) Prevent containers from unauthorized handling

(b) Prevent from unauthorized (macro) hub entry

Safe Secure Authorized

The μ-hub owners

1. The μ-hub should function as an ordering mechanism

(a) Collect incoming order demand (regular and special) (b) Search for a qualified supplier, when special demand is requested

Open (Trans-parent) Pull-based

2. The μ-hub should maintain a forecasting function

(a) Collect historical order data

(b) Collect sales data (c) Monitor lead times

Accurate Responsive

3. The μ-hub should function as close as possible to a staging point

(a) Reduce length of stay of supply and containers

(b) Ensure fast decapsulation (c) Ensure fast consolidation

Accurate Reliable Sustainable

4. Ensure μ-hub location, to be strategically decided

(a) Space for (un)loading purposes (b) In closest possible proximity to nanostores Reachable Strategic location Spacious for (un)loading 5. Ensure μ-hub can only be

entered by authorized individuals

(a) Prevent from unauthorized μ-hub entry

Safe Secure Authorized 6. Get the right products at

the right nanostores

(a) Enable a sufficient picking mechanism

Accurate Reliable User friendly

(28)

Table 7.1 continued from previous page

7. Maintain fresh state of products

(a) Maintain right conditions inside the μ-hub

(b) Reduce length of stay

Fresh Fast

8. Ensure that routes are optimized and adapted to available driver capacity

Responsive Capacity restrictions The intra-favela

transporter

1. Collect a payment from the nanostore owner after every delivery

Connected Organized

The nano store owners 1. Collect profit out of selling goods

(a) Enable more efficient supply, by utilizing the μ-hub

Profitable

2. Always have supply available to sell to customers

(a) Order the right fresh products

(b) Order the right amount of fresh products

Available Reliable Customer satisfaction 3. Keep (different types of

products) fresh, when stored

(a) Enable cold storage (b) Provide an energy supply to the cold storage function

Fresh Cold facilities Energy supply The slum residents

1. Develop ones need to a healthier food supply

(a) Receive education on the benefits of a healthier lifestyle

Education

2. Have access to a healthier food supply

(a) Have access to food subsidies

(b) Have access to affordable, fresh products Healthy Quality of life Affordable Formal authorities

1. Collect fees Profitable

2. Enforce the law Informal authorities 1. Provide protection, in return for fees

(a) Guard replenishment by (external) PI-transporters (b) Guard the μ-hub

Safe Secure

(29)

8 Functional architectures

In this chapter, the functional architectures and interaction diagrams of the μ-hub are developed and presented, answering sub-question (g), chapter 4. In advance of modelling the system through functional architectures, three design scenarios are distinguished, in accordance with chapter 6. These are illustrated by means of a Business Process Modelling and Notation (BPMN) approach. The depicted BPMN models show the involved actors, their tasks and decisions in support of the μ-hub’s functioning. For further reasoning and a visualization of these scenarios, appendix G can be consulted.

Subsequently, in accordance to the design scenarios, the functional architectures are depicted. These functional architectures are derived and adjusted from previous de-veloped architectures of dr. N. B. Szirbik and his MSc. students, at the University of Groningen. In addition, these diagrams depict the operational process, the system’s (sub)-functions which form this system, the interaction between sub-functions and necessary controls. As these diagrams model the operational process, they especially put emphasis on the higher levels of the system. Furthermore, in a better under-standing of this chapter, it is recommended to read chapter 9 in parallel. Whereas at last, in line with the alternative design scenarios, functional alternative diagrams are developed and depicted in appendix H.

8.1 Function A-2 PI-based context functions

(30)

The PI-based preliminary context functions, as decomposed in figure 8.1 and further shown in figure 8.2 indicate preparatory means for enabling PI-based supply of fresh products to μ-hubs. These preparatory means consider among others an ETA to external PI-transporters and information on μ-hub locations.

(31)

(32)

8.2 Function A.dash.1 Enable PI-based supply

Figure 8.3: Hierarchy diagram: A.dash.1 Enable PI-based supply

Figure 8.3 and 8.4 provide a more narrow illustration of the systems’ context. Whereas the decomposition of A.dash.2 entails a higher level approach, A.dash.1 evolves around the lowel level of the supply chain.

Subsequently, in enabling PI-based supply of fresh products to slums, these products need to reach the closest DC near the slum as for A.dash.11. Whereas earlier denoted, this DC is called a macro-hub as it substantiates itself from a regular DC by performing merely the task of container detachment and managing the flow of incoming and outgoing containers, highlighted as A.dash.12. Consequently, when subfunction A.0 is fulfilled, A.dash.13 illustrates alignment with assumption 15. Namely, that slum residents have developed significant demand for the consumption of fresh products, as brought forward by the μ-hub.

(33)

(34)

8.3 Function A.0 Reaching nanostores via μ-hub

Figure 8.5: Hierarchy diagram: A.0 Reach nanostores via μ-hub

Concerning figure 8.5, A.0 illustrates the design of the system to be and its sub-functions which basically control the operational process. From this operational process point of view, A.3 indicates high critical importance, whereas it is however not seen as the most important sub-element. This is the case, as it functions in equal way to the initiator of order placement, turning the μ-hub in a pull-based system. Hence, due to its importance, A.3 is decomposed in section 8.7.

To the contrary, within this decomposition, A.1 is regarded as the most important sub-element for the system to be, as it is believed that the system cannot be main-tained without the placement of security measures. This is viewed in two ways. First, security should be enabled, as otherwise last mile delivery would be viewed unsafe due to the presence of bad safety conditions and corruption, seen in table 6.1. Second, security should be enabled, as otherwise PI-based context functions would be at risk. To explain, safety hazards might indicate larger amounts of either container theft or encapsulated supply. This risks profitability of PI-transporters, suppliers and other actors involved, therefore reducing incentives to engage in the utilization of a μ-hub.

(35)

(36)

8.4 Function A.1 Manage safety requirements

Figure 8.7: Hierarchy diagram: A.1 Manage safety requirements

(37)

(38)

8.5 Function A.2 Perform last mile delivery

Figure 8.9: Hierarchy diagram: A.2 Perform last mile delivery

The hierarchy diagram of figure 8.9 illustrates the last mile delivery. This function is decomposed in A.21, the holding, consolidation and redistribution from, and to the μ-hub until all deliveries are performed. Whereas it additionally covers the sole act of transportation in between the μ-hub and nanostores, the management of M-payments and the act of keeping and selling the fresh delivered products at the nanostores.

(39)

(40)

8.6 Function A.21 Hold consolidate and pick-up

from μ-hub

Figure 8.11: Hierarchy diagram: A.21 Hold, consolidate and pick-up from μ-hub

(41)

(42)

8.7 Function A.3 Place order(s) at supplier(s)

Figure 8.13: Hierarchy diagram: A.3 Place order(s) at supplier(s)

Order placement, as the most critical function regarding the operational process is decomposed in four sub-functions as depicted in figure 8.13.

(43)

(44)

9 Design decisions and CSFs

In this chapter, arguments are provided for the design decisions made, in the sce-narios and functional architectures. Multiple alternatives are considered, whereas reasoning is provided to support decisions on the optimistic scenario. It has to be noted that these design decisions already keep into account some of the physical components that might be considered as important for the system’s functioning. Moreover, all design decisions as utilized in the optimistic scenario are marked in green, if no marking is present in a section decisions are considered optional. Furthermore, these design decisions support in the argumentation of CSFs for the system to be. In answering sub-question (h), chapter 4, these CSFs are mentioned beneath each section.

9.1 Order demand

Due to the principles of PI, the μ-hub will solely function as a pull-based system, indicating that sole nanostore demand will generate supply.

When orders are collected they are shared with the μ-hub that falls in its region. Subsequently, each μ-hub is accommodated with a coupling mechanism, enabling all nanostore orders in its region to be consolidated and then treated as a single order accordingly.

In addition, a two-way distinction is made between the types of order demand, containing a:

1. Regular order demand

A regular demand contains demand for fresh products that are considered fast moving to nanostores. These products, as well as suppliers of these products are known to the μ-hub, due to historical orders.

2. Special order demand

Special order demand illustrates items which are not considered regular. Ir-regularity indicates that demand for these items is sporadic, providing variety in lead times. The μ-hub distinguishes between two types of special order demand:

(a) Fresh food consumables

(45)

cheeses (Edam, Gouda) or a range of Japanese fishes). This type of special demand is generated by nanostore customers.

(b) End-of-life cooling containers

A second form of special order demand indicates fresh food storage equip-ment. To keep fresh food in a non-hazardous state cooling should be con-trolled up until the point of sale. Nonetheless, it is impossible that fresh food will be sold immediately after arrival at the nanostore, necessitating a cooling function. As a solution cooled end-of-life containers are utilized, see section 9.8, from which demand is generated by the nanostore owners. CSFs:

• Incoming regular, as well as special demand shall always pass the μ-hub as an intermediary staging point for nanostore replenishment purposes.

• No supply is pushed through the μ-hub, it shall therefore solely function as a pull-based system.

• The μ-hub can only sustain if there is supply, which is able to pass the hub. Therefore, educational programs shall be set in place, to generate knowledge about, and demand for fresh products if latent.

• To foster fresh food safety, and alignment with fresh food protocols, end-of-life cooling containers shall be utilized at the point of sale.

9.2 Storage in the μ-hub

To avoid turning the μ-hub into a safety hazard, the μ-hub should hold as minimal inventory as possible. Preferably, this would entail the μ-hub to function on a JIT mechanism. Due to the openness of PI, one can directly buy from a producer, reducing lead times and increasing product freshness (Riikka, Dukovska-Popovska, & Loikkanen, 2013). However, a full in place JIT mechanism is expected to be non-realistic in a slum-like area. Due to the fact that PI may reduce lead times, these lead times might still be uncertain especially for special order demands. For example, due to the utilized PI-characteristics, if a special kind of cheese is demanded, like Edam or Gouda, the μ-hub might order the cheeses at any suitable supplier among whom Unilac, a dutch certified cheese supplier (Unilac Holland BV, n.d.). This act of ordering special demand provides varying lead times due to the range of suppliers and products. Whereas, for a JIT mechanism to function well, all parts in the process of delivery need to be certain. Nonetheless, intra-favela transporters are capacity bounded, requiring multiple μ-hub pick-ups, making it non-realistic to maintain direct pick-up after arrival.

(46)

function can maintain incoming supply in the μ-hub by generating the right order quantities and overcoming lead time uncertainties.

CSF

• The μ-hub shall function as close as possible to a staging point, holding as little inventory as possible.

9.3 Ordering mechanism

Moreover, in order for the μ-hub to enable order placement, order information needs to be collected beforehand. In terms of this order collection process the following alternatives are considered:

1. The intra-favela transporter places the order

(a) By phone; In practice, ordering by phone would entail that each nanos-tore owner can send a message to the intra-favela transporter, before a specified time. No received message triggers the intra-favela transporter to respond by a phone call, for which assumption 14 holds. Subsequently, for this to function well the time of order placement is crucial and should be kept to a minimum, as intra-favela transporters are bound to transport the rest of the day. Collecting orders by phone is advantageous in the sense that these time restriction can be met by collecting orders before a shift, whereas personal contact is fostered. However, on the downside, this method demands more time from the intra-favela transporter, which is not directly related to delivery purposes. As for the experienced view of Fietskoeriers.nl, this should be avoided as much as possible, figure D.3. (b) Via direct contact; Alternatively, when a intra-favela transporter is al-ready on route, the possibility arises to take orders in parallel. In addi-tion, this lengthens the intra-favela visit to the nanostore while reducing dedicated time to performing deliveries. However, this fosters the social interaction between transporter and nanostore owner.

2. The nanostore owner places the order

Due to assumption 3, appendix E, the possibility arises for nanostore owners to place orders themselves. This might be beneficial, as orders might be di-rectly processed due to the eliminated necessity of a gatekeeper (intra-favela transporter). Consequently, in this scenario faster order processing, as well as placement is expected. Additionally, this scenario enables intra-favela trans-porters to safe time, which in turn can be dedicated to delivery purposes. Due to these opportunities, this design decision is utilized in the optimistic scenario.

CSFs

(47)

orders is most critical with respect to initiating the operational pro-cess.

• To enable order placement, the μ-hub shall have enough capacity to collect and process all information required.

• To allow order placement by nanostore owners, access to the order-ing platform for these actors shall be enabled.

9.4 Consolidation activities

In terms of consolidation as a design decision two options are considered. Consoli-dation in the hub is either performed by:

1. A μ-hub worker

If consolidation is performed by a μ-hub worker, then tasks are separated for the intra-favela transporter and his priority shifts to the execution of deliv-eries solely, as advised by Fietskoeriers.nl, figure D.3. This is seen as most advantageous, as separating tasks reduces variability and therefore potential for errors (Adeyoyin, Agbeze-Unazi, Oyewunmi, Adegun, & Ayodele, 2015). Moreover, in fostering the direct processing of incoming supply, utilizing μ-hub workers would enable this due to their presence at the μ-hub. Additionally, by employing slum residents, like μ-hub workers, more economical value is added to the community. Whereas the acceptance of a μ-hub implementation would be stimulated. By means of these reasons, a μ-hub worker is utilized in the optimistic scenario.

2. An intra-favela transporter

An alternative solution is for the intra-favela transporter to perform the act of consolidation oneself. This is the case, as the intra-favela transporter might be better capable of estimating its own capacity restrictions.

CSFs:

• The μ-hub shall enable consolidation tasks to be performed by a designated μ-hub worker.

• The μ-hub shall provide access to planning and routing information, for consolidation purposes.

• The μ-hub shall utilize a nearby energy supply to temporarily hold containers, in enabling consolidation.

9.5 Vehicles used

(48)

a higher carrying capacity is required, as it should carry multiple containers that each carry a consolidated load. Two options arise:

1. Each μ-hub owns a van

Utilizing a fixed van per μ-hub would indicate less capacity restrictions. As only the capacity of at least one μ-hub should be carried per run. Nonetheless, this also indicates higher investment costs, maintenance and higher efforts in terms of safety.

2. Vans are shared between μ-hubs

Utilizing a shared van would increase the probability of not fulfilling a sin-gle run to supply all μ-hubs due to capacity restrictions. Multiple runs are necessary therefore pressuring the JIT mechanism. However, investment and maintenance costs would be lower, whereas safety efforts would be required in less quantities.

3. A PI-transporter outside the slum is used

Another alternative, as utilized in the optimistic scenario, provides the oppor-tunity of utilizing a PI-transporter outside the slum. This means no investment in vans or maintenance is required, as transport between the macro-hub and μ-hubs is outsourced. This option might be pertained as most fruitfull and best fitting the principles of PI. Additionally, safety hazards might be reduced by exchanging a fee in return for security activities from informal authorities. Moreover, from the μ-hub to the nanostores, much smaller vehicles need to be uti-lized. Two options can be distinguished:

1. Informal vehicles are utilized

In the current situation, intra-favela transporters utilize a variety of small, informal vehicles, appendix B.8. Each of these vehicles has a different carrying capacity, posing a restriction on the load size. Nonetheless, continuing to use informal vehicles would indicate less investment costs, as well as safety measures necessary.

2. A fixed set of vehicles is utilized

By utilizing a fixed set of vehicles more control can be directed on the maxi-mum load size carried and speed reached. It is highly probable that a higher capacity can be maintained, see options as by Fietskoeriers.nl, figure D.2, therefore reducing the number of runs required. However, the probability of stricter safety measures would be expected, as vehicles become more advanced. Nonetheless, this option is adapted in the optimistic scenario due to its ben-eficial elements. Whereas an optional safety measure is discussed in section 9.8.

9.6 Payment options

(49)

1. The cash-based method is continued

Maintaining a cash-based method would indicate that for the intra-favela transporter safety is at risk due to the carrying of actual money at hand. As countermeasures, investment costs arise for the implementation of tracking systems, as well as other safety measures for intra-favela transporters. Addi-tionally, keeping the situation as it is, would probably maintain the no cash, no sale policy. Therefore not reducing the level of empty deliveries.

2. The cash-based method is discontinued

As an alternative solution, it is theorized that the digitization of cash-payments would bring more advantageous means. In order for digitization to be applica-ble, it is expected for assumption 3 to be set in place. Digitization is believed to be applicated through branchless banking. Branchless banking enables access to financial services outside traditional bank branches, which does not require a formal bank account, by using agents among which informally licensed insti-tutions (Mas, 2009). This act enables the transfer of M-payments, a so called payment of an electronic store of value, by means of a mobile device (Mas, 2009). Especially for the extreme poor, branchless banking is an academically backed solution. This is the case, as the barriers to entry are perceived to be low, whereas limited legal boundaries exist to connect participants to accounts (Ivatury & Mas, 2008). The large gains that can be achieved, causes branch-less banking to be utilized in the optimistic scenario. Nonethebranch-less, as agents, devices and platforms need to be developed in a match to the application environment, the actual implementation is expected to differ per country. CSFs

• The μ-hub will only function if all parties are engaged to the utilization of M-payments. Therefore, in order for M-payments to sustain, commu-nication of benefits and usage to nanostore owners shall be enabled.

9.7 Security measures

In order for the μ-hub to function properly security measures should be taken to generate a secure infrastructure for the last mile delivery process. This infrastructure might be enabled by:

1. Involving local authorities

(50)

2. Safety protocols

Safety protocols should be set in place to allow for safe handling of fresh prod-ucts in the μ-hub, as well as on delivery. Therefore some sort of safety proto-cols should be built and maintained. Safety protoproto-cols adhere to among others locking bicycles and supply components, as mentioned by Fietskoeriers.nl, ap-pendix D, utilizing secure μ-hub entering systems and entering the μ-hub only when required.

3. Authorization controls

Authorization controls should be enabled for μ-hub entering. Ingoing and outgoing supply should be authorized to avoid unauthorized access and con-tent’s safety. Multiple authorization options are optional for this enablement. Nonetheless, the extremes of safety hazards, as subtracted in table 6.1, ask for mechanisms and controls which are difficult to copy, like individual character-istics inputted in biometric systems. Biometric systems are promising forms of identification enabling technologies that allow for authorization by screening an individual’s physiological or behavioral characteristics (Tan & Schuckers, 2005). These include fingerprints, irises, face recognition, face thermograms, hand geometry, retinal patterns and speech (Anil, Hong, & Pankanti, 2000). Moreover, these solutions form promising authorization and control mecha-nisms for the μ-hub, whereas they are not without risks. For example, a fingerprint recognition system might be subject to database or sensor level attacks (Tan & Schuckers, 2005), whereas speech-based recognition systems are sensitive to background noises or to the speakers physical state (Anil et al., 2000). In addition, as noted in table 6.1, equipment should be able to en-dure extreme environmental factors, whereas these technologies are foremost sensitive.

CSFs:

• To secure incoming and outgoing supply, one shall need to involve (in)formal authorities in doing so.

• To ensure safe handling and delivery of fresh products, safety pro-tocols shall be built and maintained

• The μ-hub shall be secured with authorization mechanisms and con-trols, which are difficult to copy and fitting to the environmental factors present in the direct application environment.

9.8 Utilize end-of-life containers

(51)

that can add value when deployed elsewhere. Consequently, in order to ensure deployment for a different purpose, the following acts should be incorporated:

1. A detection function should be set in place, to detect the end-of-life status of a container.

2. The (PI, π)-container owner should be willing to sell off the container, when the end-of-life status of the container is reached.

When these acts are incorporated, different applications of end-of-life containers can be ensured in slums. These applications might subsequently foster the establishment of a μ-hub and are considered to be as follows:

1. Cooled storage application

As for table 6.1, no cold storage facilities are present in nanostores. However, in order to maintain products in a fresh state, a cold storage adaptation is required. As a solution, a specialized end-of-life container can be utilized. Due to the requirements for cooling, a so called reefer container could propose a qualified fit. A reefer container includes a cooled aggregate as a panel module, keeping enclosed products at the required temperature (Moesker, 2018). 2. Security boxes, for fixed vehicles

As for appendix B.8, the option arises to use a fixed set of vehicles for intra-favela transporters. For this option the downside indicates a higher probability of vehicle theft. However, by adapting an end-of-life container into a security box, probabilities of theft might decline. This is the case, as containers contain characteristics of strength and durability (Landsch¨utzer, Ehrentraut, & Jodin, 2015). Whereas vehicles are believed to be better secured due to an attached S-sensor (Sallez, Pan, Montreuil, Berger, & Ballot, 2016).

3. Building blocks for the μ-hub

At last, end-of-life (PI, π)-containers might function as building blocks for the μ-hub due to their characteristics of strength, durability and identification (Landsch¨utzer et al., 2015). These characteristics might form an important protective function to prevent from equipment failures, a hazard as identified in table 6.1. Furthermore, (PI, π)-containers allow for continuous communication features to be active, enabling the container to be connected to the internet in a stable and constant manner (Moesker, 2018), required to place orders. CSFs:

University of Groningen Faculty of Economics and Business