University of Southern California Dissertations and Theses

Mobile apps and services for development: What can we learn from the non-smartphone era in ICT4D?

(USC Thesis Other)

Mobile apps and services for development: What can we learn from the non-smartphone era in ICT4D?

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 1
MOBILE APPS AND SERVICES FOR DEVELOPMENT:
WHAT CAN WE LEARN FROM THE NON-SMARTPHONE ERA IN ICT4D?
by
Melissa Loudon
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfilment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMMUNICATION)
December 2015
Copyright 2015 Melissa Loudon
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 2
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 3
Copyright
This work is available under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0
International (CC BY-NC-ND 4.0) License.
You are free to:
• Share — copy and redistribute the material in any medium or format
The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
• Attribution — You must give appropriate credit, provide a link to the license, and indicate if
changes were made. You may do so in any reasonable manner, but not in any way that suggests
the licensor endorses you or your use.
• NonCommercial — You may not use the material for commercial purposes.
• NoDerivatives — If you remix, transform, or build upon the material, you may not distribute
the modified material.
• No additional restrictions - You may not apply legal terms or technological measures that
legally restrict others from doing anything the license permits.
Notices
You do not have to comply with the license for elements of the material in the public domain or where
your use is permitted by an applicable exception or limitation.
No warranties are given. The license may not give you all of the permissions necessary for your
intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you
use the material.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 4
Acknowledgements
It takes a village, it really does.
The elders of this particular village, my committee members Mike Annany, Manuel Castells
and François Bar as chair, deserve unrestrained thanks for draft-reading, admin-wrangling, direction,
support and inspiration throughout my time at USC. I feel very privileged to have worked with you.
François, in particular, I could not have asked for a better faculty mentor and chair.
In direct relation to the research components, I am grateful to the anonymous respondents who
agreed to be interviewed for this research, as well as to Dimagi for allowing me to use their data for the
secondary data analysis. Neal Lesh and Rashmi Dayalu helped me derive sensible hypotheses and
output measures, Jon Jackson interjected with ideas, and Scott Lee provided feedback on the mixed
effects design.
I've been fortunate to be part of a fantastic community of PhD students at Annenberg. You are
the only thing I'll miss about the dungeon. Within and outside of Annenberg, friends who knew when
not to ask how it was going (and when to ignore me and ask anyway) thank you for listening and
advice and everything else. You never made a big deal out of it, but when I needed you, there you
were.
Drew, thank you for unwavering support, even when my decisions defied logic and my
emotional state defied reasonable management. Family, the same. None of you have to read it, but it
means a lot that you offered.
Support throughout my time at USC was provided by the USC Annenberg PhD Fellowship.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 5
List of tables
Table 1: Research design summary.........................................................................................................55
Table 2: Comparison of platforms for m4d apps and services................................................................94
Table 3: Mixed-effects model for productivity measure, R notation....................................................179
Table 4: Mixed-effects model for retention measures, R notation........................................................179
Table 5: Independent variables and group counts, by model................................................................180
Table 6: Results of likelihood ratio tests for fixed effects in productivity models................................184
Table 7: Parameter estimates for significant fixed effects in productivity models................................184
Table 8: Marginal and conditional R2 for productivity models............................................................185
Table 9: Results of likelihood ratio tests for fixed effects in retention models.....................................187
Table 10: Parameter estimates for significant fixed effects in retention models...................................188
Table 11: Summary of CommCare data analysis findings....................................................................190
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 6
List of figures
Figure 1: A distributional model of infrastructure (Bowker at al 2010, p.101).......................................37
Figure 2: Infrastructures, platforms and actants in m4d apps and services.............................................47
Figure 3: Research components mapped to framework..........................................................................50
Figure 4: GSM Network Architecture (Hillebrand, Rosenbrock & Hauser 2013)..................................66
Figure 5: Actors in Standards Setting after GSM Phase 2+ (Hess & Coe, 2006, p. 1219).....................68
Figure 6: Person-to-person SMS transmission (Acker, 2014, p.565)......................................................74
Figure 7: Third party - WASP - operator relationship. Adapted from Brown, Shipman and Vetter (2007,
p.108).......................................................................................................................................................75
Figure 8: USSD application architecture (ETSI, 1996)...........................................................................84
Figure 9: Technology platform frequency in m4d apps and services inventory.....................................99
Figure 10: Inventory apps and services grouped by archetype.............................................................106
Figure 11: Inventory apps and services by archetype and sector..........................................................108
Figure 12: Inventory counts by archetype and mobile technology platform.........................................110
Figure 13: Language support in inventory apps and services................................................................112
Figure 14: Inventory breakdown by cost to user...................................................................................114
Figure 15: Developers of inventory apps and services..........................................................................116
Figure 16: Types of funders for inventory apps and services................................................................118
Figure 17: Funder type by technology level..........................................................................................119
Figure 18: Inventory apps and services by sector.................................................................................121
Figure 19: Intended users of inventory apps and services.....................................................................123
Figure 20: Inventory apps and services by development theory of change..........................................127
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 7
Figure 21: Median visits per project in month six, showing projects that are part of organizations that
have multiple projects............................................................................................................................178
Figure 22: Mean and standard error, visits by project, users' sixth month on CommCare....................182
Figure 23: Mean and standard error, visits by project, users' twelfth month on CommCare................182
Figure 24: Mean and standard error, visits by project, users' eighteenth month on CommCare...........182
Figure 25: Short-term retention percentage by project..........................................................................182
Figure 26: Long-term retention percentage by project..........................................................................182
Figure 27: Visits per user in their twelfth month on CommCare, by device type.................................186
Figure 28: Percent of users short-term retained per project, by device type.........................................188
Figure 29: Percent of users long-term retained per project, by device type..........................................189
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 8
Abbreviations
AIDS Acquired immune deficiency syndrome
AMPS All Media and Products Survey
ANT Actor-network theory
API Application programming interface
CAMEL Customised Applications for Mobile networks Enhanced Logic
CDC Centers for Disease Control and Prevention
EDGE Enhanced Data for Global Evolution
ETSI European Telecommunications Standards Institute
GDP Gross domestic product
GPRS General packet radio service
GPS Global Positioning System
GSM Global System for Mobile Communications
GSMA GSM Association
GUI Graphical user interface
HIV Human immunodeficiency virus infection
HST Health Systems Trust
HTML HyperText Markup Language
HTTP Hypertext Transfer Protocol
ICT Information and communication technologies
ICT4D Information and communication technologies for development
ICTD Information and communication technologies and development
IVR Interactive voice response
JVM Java virtual machine
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 9
LSM Living Standards Measure
MMS Multimedia Messaging Service
MNO Mobile network operator
MRC Medical Research Council
MSC Mobile switching centre
NGO Non-governmental organization
PBO Public benefit organization
PDA Personal digital assistant
RCT Randomized controlled trial
RMS Record management system
SIM Subscriber identity module
SMS Short message service
SMSC Short message service centre
SS7 Signalling System No. 7
SST Social shaping of technology
UCT University of Cape Town
UNICEF United Nations Children's Fund
USAID United States Agency for International Development
USSD Unstructured supplementary service data
WAP Wireless access protocol
WASP Wireless application service provider
WASPA Wireless Application Service Providers’ Association
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 10
Abstract
This project aims to describe and explain the field of mobile apps and services for development
- 'm4d' apps and services – primarily by studying its realisation in South Africa during the pre-
smartphone era. Concepts from infrastructure studies (Star, 199) and platform studies (Montfort &
Bogost, 2009) are employed to describe the relationship between technology platforms and the apps
and services built on them. Research activities include producing a historical and technical account of
the mobile network and platforms for m4d apps and services, taking inventory of m4d apps and
services in South Africa and categorizing them according to various dimensions of interest, conducting
and analysing unstructured interviews with developers and designers of m4d apps and services, and
carrying out an exploratory quantitative analysis of secondary data generated through use of a mobile
data collection app. The findings demonstrate that claims about the potential for m4d apps and services
should be understood in the context of embodied infrastructural limitations, arising from the gap
between design context and the use contexts in the developing world. The result is a field with limited
diversity, with designs for m4d apps and services reducible to a relatively small number of archetypes.
Alignment with actants in the global network for m4d apps and services, including funders, mobile
network operators and “circuits of truth” (Roy 2010, location 123) such as the GSMA's Mobiles for
Development Impact programme produces additional incentives for concentration. Finally, the
discursive construction of the m4d consensus as it relates to apps and services is not neutral. Whatever
benefits accrue to users in the developing world, frames that position technologies as offering solutions
to development problems are designed to achieve favourable public perception, and by extension
favourable regulatory treatment, for technology industry interests.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 11
Table of Contents
Copyright..................................................................................................................................................3
Acknowledgements...................................................................................................................................4
List of tables.............................................................................................................................................5
List of figures...........................................................................................................................................6
Abbreviations............................................................................................................................................8
Abstract...................................................................................................................................................10
CHAPTER 1: Introduction......................................................................................................................16
1.1 Background...................................................................................................................................16
1.2 Research Approach.......................................................................................................................20
1.3 Research Objectives......................................................................................................................22
1.4 South African case........................................................................................................................25
1.5 Scope and Limitations..................................................................................................................26
1.6 Chapter Outline.............................................................................................................................27
CHAPTER 2: Conceptual Framework....................................................................................................29
2.1 Mobile phones and development..................................................................................................30
2.2 Mobile networks as infrastructure................................................................................................35
2.3 Studying technology platforms.....................................................................................................38
2.4 Actor-network theory in ICT4D...................................................................................................40
2.5 Innovation and Appropriation.......................................................................................................43
2.6 Framework....................................................................................................................................47
CHAPTER 3: Methods............................................................................................................................51
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 12
3.1 Research design in infrastructure studies and ANT......................................................................51
3.2 Study design..................................................................................................................................53
3.3 Part 1: The GSM standard and platforms for m4d apps and services...........................................56
3.4 Part 2: Apps and services inventory..............................................................................................57
3.5 Part 3: Key informant interviews..................................................................................................61
3.6 Part 4: Quantitative analysis of operational data..........................................................................63
CHAPTER 4: The Mobile Network........................................................................................................64
4.1 Origins of GSM............................................................................................................................64
4.2 Platforms for Apps and Services...................................................................................................69
4.2.1 Voice and Interactive Voice Response (IVR)........................................................................69
4.2.2 SIM Application Toolkit.......................................................................................................71
4.2.3 The Short Message Service (SMS).......................................................................................72
4.2.3.1 Message Transmission..................................................................................................74
4.2.3.2 Billing............................................................................................................................79
4.2.3.3 Message content............................................................................................................81
4.2.3.4 Security.........................................................................................................................82
4.2.4 Unstructured Supplementary Service Data (USSD).............................................................83
4.2.5 Mobile Internet......................................................................................................................85
4.2.6 The J2ME App Environment................................................................................................89
4.2.7 Comparison of platforms for m4d apps and services............................................................92
4.3 Discussion: Mobile network platforms for m4d apps and services..............................................95
CHAPTER 5: Apps and services inventory findings..............................................................................98
5.1 Technology platforms...................................................................................................................99
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 13
5.2 Archetypes..................................................................................................................................101
5.3 User experience dimensions........................................................................................................111
5.3.1 Languages...........................................................................................................................111
5.3.2 Cost to user..........................................................................................................................113
5.4 Product development dimensions...............................................................................................116
5.4.1 Developers..........................................................................................................................116
5.4.2 Funders................................................................................................................................118
5.4.3 Mobile operator involvement..............................................................................................120
5.5 Development intervention dimensions.......................................................................................121
5.5.1 Sector..................................................................................................................................121
5.5.2 Intended users.....................................................................................................................122
5.5.3 Geographic Reach...............................................................................................................124
5.5.4 Development theory of change...........................................................................................126
5.5.5 Role of technology..............................................................................................................128
5.5.6 Measures of success............................................................................................................130
5.6 Discussion: South African m4d apps and services landscape....................................................133
CHAPTER 6: Interview findings..........................................................................................................137
6.1 Technology..................................................................................................................................138
6.1.1 Platforms.............................................................................................................................142
6.2 Projects.......................................................................................................................................148
6.2.1 Process................................................................................................................................148
6.2.2 Users and non-users............................................................................................................149
6.2.3 Funding...............................................................................................................................152
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 14
6.2.4 Operator relationships.........................................................................................................154
6.2.5 Government.........................................................................................................................158
6.2.6 Evaluation and impact.........................................................................................................159
6.2.7 Sharing and re-use...............................................................................................................161
6.3 General observations on m4d.....................................................................................................164
6.3.1 Sector concentration and lack of diversity..........................................................................164
6.3.2 Stakeholders in defining m4d.............................................................................................167
6.3.3 Smartphones........................................................................................................................169
CHAPTER 7: Findings from CommCare metadata analysis................................................................172
7.1 Data and methods.......................................................................................................................173
7.1.1 Data.....................................................................................................................................173
7.1.2 Data gathering.....................................................................................................................174
7.1.3 Cleaning..............................................................................................................................175
7.1.4 Hypotheses..........................................................................................................................176
7.1.5 Analysis and descriptives....................................................................................................178
7.2 Findings......................................................................................................................................183
7.2.1 Predicting user productivity (number of visits)..................................................................183
7.2.2 Predicting user retention.....................................................................................................186
7.3 Discussion...................................................................................................................................189
CHAPTER 8: Conclusion......................................................................................................................194
8.1 The realisation of the m4d apps and services promise...............................................................194
8.2 New technology paradigms and ICTD research.........................................................................199
8.3 Platforms and design considerations for m4d apps and services................................................200
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 15
8.4 Broadening the scope of ICT4D interventions...........................................................................203
References............................................................................................................................................206
Appendices...........................................................................................................................................227
Appendix A: Interview topic guide..................................................................................................227
Appendix B: Inventory dimensions.................................................................................................230
Appendix C: Apps and services in inventory...................................................................................235
Appendix D: CommCare data analysis code book..........................................................................243
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 16
CHAPTER 1: Introduction
1.1 Background
From 2005 to 2013, the number of mobile cellular subscriptions globally more than tripled,
from 2.205 million to 6.835 million (International Telecommunications Union, 2014). Developing
countries are largely responsible, increasing by over 400% from 1.213 million subscribers in 2005 to
5.235 million in 2013. In 2004, the number of people with voice service on the continent having
doubled over the preceding three years, the International Telecommunications Union (ITU) began
trumpeting Africa's “mobile miracle” (ITU, 2004).
Widespread adoption by users at the “bottom of the pyramid” (Prahalad, 2006) lead to strong
interest from the development community, and in 2008 the Economist described the mobile phone as
“a potent force for economic development in the world's poorest countries” (“Halfway there”, May 29
2008). Aker and Mbiti (2010, p.3) characterize the discourse on mobiles and development as “a great
deal of speculation and optimism,” sustained both by the apparent potential of mobile technology and
its unprecedented diffusion.
Over the last decade, research has provided varied evidence of development impacts. Increased
mobile penetration has been linked to economic growth (Waverman, Mesci & Fuss, 2005). Mobiles
have been shown to improve the livelihoods of various groups of small traders and entrepreneurs in the
developing world (see review in Donner & Escobari, 2010), as well as the effectiveness of individuals'
job search (Samuel, Shah & Hadingham, 2005), self-reported life satisfaction (Goodman, 2005), and
ability to maintain economically important social ties (Diga, 2008). By making it possible for
previously disconnected groups to disseminate information and communicate rapidly and across
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 17
distance, mobile phones have also been implicated in sustained, large-scale popular mobilization
(Castells, Fernandez-Ardevol, Qiu, & Sey, 2007).
In addition to work on the impacts of access to mobile communication, the field of mobiles for
development (m4d) is concerned with designing, implementing and evaluating mobiles apps and
services for the “bottom of the pyramid.” Mobile apps and services allow development interventions
reach previously unserved populations, who now have access to telecommunication in the form of
basic mobile phones. They are also used to extend the reach, capabilities or efficiency of social
services, with mobile health, education and agricultural support being established subfields. Unlike
their predecessor in telecenters, which were beset by hardware failures, low user numbers and
sustainability problems, mobiles offer an established user base, reachable at a distance, through a
device that users buy and manage themselves.
With prices for voice and text messaging falling, mobile technology shifting “from narrowband
to broadband” (World Bank 2012, p. 1), and phones with Internet access increasingly the norm, mobile
operators are also looking to apps and services as a promising revenue source. Kenya's m-Pesa, a
mobile money service with over 14 million users in March 2012 (Safaricom, 2013) operated by
dominant mobile network operator Safaricom, is frequently cited as an example of the potential for
market-led development through mobile apps and services.
The World Bank's Information and Communication for Development Report for 2012,
explaining the institution's shift in focus from basic connectivity to mobile apps and services, reasons
that “mobile applications not only empower individual users, they enrich their lifestyles and
livelihoods, and boost the economy as a whole” (p. 3). The GSM Association, an industry trade group
representing GSM operators, takes a similar position. They have programmes that promote and support
mobile services in health, agriculture, financial services, job search, community services, disaster
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 18
response and women's issues in the developing world, and envisage their role as “Serving the
underserved through mobile: Bringing together our mobile operator members, the wider mobile
industry and the development community to drive commercial mobile services for underserved people
in emerging markets.” Their website describes mobiles as “one of the world's most potent development
tools” (GSMA, 2013). The overarching discursive frame for m4d apps and services is that, although
domain-specific nuances exist, apps and services offer the possibility of scalable, market-based
technological solutions to various development problems.
Over ten years since the proclamation of the 'mobile miracle', mixed evidence from the field
suggests a more complicated reality. Widespread adoption of mobile phones represents an important
gain for previously disconnected users. There are also significant examples of mobile apps and services
attracting low-income consumers, some of which make explicit claims about tackling development
problems. Yet, with the exception of youth-oriented social networking and mobile money in some
countries, there are few examples of apps and value-added services in widespread use among low-
income consumers. A survey of low-income mobile users in Bangladesh, India, Pakistan, the
Phillipines and Thailand, (Zainudeen & Ratnadiwakara 2011) found very low levels of mobile service
use, and only slightly greater awareness of mobile services. Mobile services users were generally better
connected (most of their contacts have a phone) and more technologically savvy (likely to also be
Internet users) than non-users. Research ICT Africa (2012) found similarly for low-income users in
South Africa, with notable use of social media and chat services among (particularly urban) youth but
little use outside this group.
The mobile communications industry and the development community have advanced a shared
vision in which mobiles – and particularly mobile apps and services developed by the private sector
with a for-profit model – enable solutions to diverse development problems. The paradigmatic nature
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 19
of this vision is in many ways analogous to the telecenter era in ICTD. Telecenters and other public
computer access venues, set up by government, NGOs or development agencies but often sponsored by
ICT companies, for a time represented consensus on how to mitigate the 'digital divide' (Heeks, 2008).
The telecenter paradigm dissolved in the face of high failure rates and sustainability problems at
telecenter projects, combined with the rise of the market-based “bottom of the pyramid” model of
development (Gurumurthy, 2010; Heeks, 2008) that underlies commercially-oriented m4d
interventions. Public access venues remain important as part of the ICT ecosystem in the developing
world, but studies such as Walton and Donner's (2012) “ecological” investigation into the interplay
between mobile Internet use and public access venues among Cape Town teens take an integrative,
context-specific perspective distinct from the utopian thinking of the telecenter paradigm.
As with telecenters, consensus on the potential of mobile apps and services has emerged from a
particular constellation of interests and incentives. As the mobile ecosystem matures, we can
investigate how these ideas have translated in particular contexts. Where early optimism about the
potential for mobile apps and services to contribute to development was based largely on mobile
penetration statistics and a few 'paradigm cases' such as M-Pesa, it is now possible to consider
evidence from an established mobile services ecosystem. Empirical evidence is available to support
both in-depth studies of particular models, application domains and user experiences, and broader
perspectives on the field.
Understanding how the relational, institutional and technological specificity of the m4d
ecosystem has affected the kinds of apps and services that are developed, which ones succeed, and
what evidence of impact they offer is part of a critical engagement with m4d. It is also relevant to ICT
for development generally, as successive claims of technological potential fuel continued optimism
without necessarily encouraging reflection. The mobile network is undergoing a major reconfiguration
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 20
as mobile operators are joined by Internet companies such as Google and Facebook, with smartphones,
tablets, and mobile Internet replacing pre-smartphone technologies such as SMS. As such, this is a
good time to consolidate learning from the pre-smartphone era in m4d.
1.2 Research Approach
This research takes a social shaping of technology (SST) approach to studying mobile apps and
services for development. SST encompasses various approaches to opening the 'black box' of
technology, studying the interplay between technology and social relations. SST research interrogates
the material features and effects of particular technologies alongside discursive frames, social relations,
processes of innovation and anticipated and unexpected use in which they arise and through which
they become embedded into human activity (Williams & Edge, 1996; MacKenzie & Wajcman, 1999).
The social shaping and material and interpretative realization of mobile apps and services for
development is “cultural formation in the world system” (Marcus, 1995, p.99) and at the same time a
reflection of the system itself. Marcus describes a research strategy of “multi-sited ethnography” in
which the subject of study is a particular cultural formation – and by extension, the system in which it
operates – rather than a single physical site, and identifies studies in this tradition in both development
studies (Escobar) and science and technology studies (Latour, Haraway). Examples that combine
development studies and SST include Burrell's study of youth practices at Ghanaian Internet Cafes
(Burrell, 2012) as well as Roy's ethnography of how ideas about microfinance circulate in the
development community (Roy, 2010).
While multi-sited ethnography begins with a “baseline” understanding of the subject of study,
it is also a process of constructing, revising and refining that develops a particular argument around the
subject (Marcus 1995). In Dourish's words “ethnographies are texts, not veridicial representations of
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 21
the world” (Dourish, 2006, p.543). This shift in perspective from description to argument is similar to
Latour’s definition of what makes a good “ANT account” of a particular actor-network: like a
laboratory experiment, Latour says that a text should be valued for what it allows us to do (Latour,
2005, p. 123).
Thus Roy (2010, location 845) traces the emergence of particular ideas about microfinance for
development, but uses her account to document the structural imperatives underlying “permission to
narrate” (Spivak, 1999, p. 191) and the production of knowledge about the lives of the poor. Burrell
studies Internet practices and the social meaning of the Internet for a particular group of Ghanaian
youth, developing an analysis that challenges commonly accepted ideas about technology transfer,
development and the digital divide.
The research described here approaches m4d apps and services as a particular cultural
formation, shaped by technology design (of the mobile network, mobile app and services platforms and
the apps and services themselves), discourses of mobiles and development, relationships between
organizations, and use, appropriation and non-use by both service designers and end-users. While my
methods are ethnographic only in the broadest sense, Marcus' suggestion that the researcher 'follow' a
particular subject that is at the same time reflective of, and embedded in, a wider set of social relations
seems well-suited to an SST study. It also accommodates the normative orientation of development
interventions (and the study thereof) while demonstrating the limits within which they operate.
In arguing for a “social process”, “human-centred”, “social constructivist” model of
technology in development, May, Waema and Bjåstad (2014) recognize that technology needs to be
understood in relation to social systems. Ethnographic studies of mobile adoption and use often use
social constructivist approaches to understand the practical and symbolic affordances of the mobile
phone to a particular group of users. However, Burrell (2012) also demonstrates the value of “strong
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 22
materialism” which seeks to understand how people construct technological worlds, but holds the
technology central because of its relative persistence in space and time. In m4d, mobile technologies
are the central organizing idea of a broad range of development interventions. A technology-centred
perspective – one which takes into account the historical development, social and institutional
arrangement and specific accordances of mobiles as a platform for m4d apps and services – is a
necessary component of an ecosystems perspective on m4d.
Ultimately, my intention is to build an argument through “reconstructing technologies as social
practices” - undertaking critical analysis of actually existing technologies, while at the same time
“imagining alternative possibilities” (Suchman et al, 1999). On the one hand, it seems pragmatically
useful to establish technological and social factors relevant to the current landscape of m4d apps and
services. At the same time, we need to critically consider the limits of the most recent in a long
succession of technology solutions to development problems. And finally, perhaps alternative
possibilities exist, in which mobile technologies, both more malleable and less governable than their
creators intended, support additional modes of development than the market-based, individualized
‘fortune at the bottom of the pyramid’.
1.3 Research Objectives
The research described here aims to produce an empirical account of m4d as a field, both in the
direct sense of developing an inventory of m4d apps and services for a single country, and as part of a
larger explanatory analysis that considers actors, relationships and discursive formations that shape and
sustain the field. From this it is possible to revisit the mobiles for development consensus, aiming to
moderate some of what has been asserted about the development impacts of mobile apps and services,
and gain an expanded understanding of emergent issues.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 23
The theoretical framework draws from infrastructure studies. Star's idea of infrastructure (Star,
1999) accommodates complex interactions between people, institutions and technologies, and can do
so at different levels simultaneously (Horst, 2013). In relation to development, is useful because it can
trace power through multiple sites, including the social, economic and political context in which
mobile networks have developed as well as local innovations and appropriations by users. As such, it
recognizes the “social embededdness” (Avgerou, 2010) of m4d apps and services without losing the
possibility of normative statements about their role (promised or actual) in promoting development.
Specifically, the conceptual framework places mobile networks as providers of platforms for
mobile apps and services. The short message service (SMS), unstructured supplementary service data
(USSD), the mobile web, and mobile app platforms such as Java Micro Edition (J2ME) are all
examples of platforms on which m4d apps and services have been built. The platform is “whatever the
programmer takes for granted when developing, and whatever, from another side, the user is required
to have working in order to use particular software” (Bogost and Montfort, 2007, p.1).
Montfort and Bogost's (2009, p. 3) intention for platform studies is to examine “how the
hardware and software of platforms influence, facilitates or constrains particular forms of
computational expression.” This requires a technically detailed focus on the features of the platform
itself, and on how they have been exploited (or subverted) by the digital artefacts that build on them –
in this case, m4d apps and services. Understanding a platform as infrastructure also considers how
particular technologies developed and came to be adopted, what motivated design decisions, and how
the discursive and cultural realization of technologies interacts with design and physical properties to
produce a particular set of affordances and constraints.
Producing an overview of a large and diverse field, more so with the detailed and context-
specific approach required by infrastructure studies and platform studies, is a vast task. The objectives
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 24
of this study are also intentionally broad. In attempting to document and analyse the realisation of m4d
apps and services in the first 10 years of the “mobile miracle”, I have chosen to adopt a multi-level,
mixed methods approach that integrates related but distinct components. The four components are as
follows:
1. A technical history of the development of the GSM standard, which specifies the underlying
architecture for mobile networks around the world, as well as various technologies that serve as
platforms for m4d apps and services. For each platform – Voice, SMS, USSD, the SIM toolkit,
Mobile Internet, Mxit, J2ME - implications of the characteristics of the platform for m4d apps
and services are discussed.
2. A structured inventory of existing m4d apps and services in South Africa, which aimed for
broad coverage across platforms and stages of development. For this and much of the interview
component, the scope is limited to a single country (South Africa) to allow the data to be
considered in a well-described context.
3. Key informant interviews with 15 individuals working in the m4d space, and/or involved in the
design and operation of m4d apps and services. Interviews were coded thematically with a view
to developing a reflective commentary on the field, as well as informing aspects of the other
research components.
4. A data-driven, exploratory quantitative study of nearly 300 m4d projects using Dimagi's
CommCare software, which provides configurable mobile data collection and reporting for
remote extension workers. Operational metadata for 300 projects and over 3000 users provides
an opportunity to test hypotheses, defined with reference to research components, that relate to
implementation support and technology choice. The inclusion of a data-driven research
components reflects a growing interest in data-driven research for m4d project evaluation,
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 25
demonstrated in both inventory and interview findings.
The resulting account is inevitably incomplete, but provides opportunities for insights at
infrastructure level. The idea of infrastructure, as well as related work in actor-network theory,
emphasises the interplay of macro, micro and contextual spheres as well as historical and perceptual
(discursive) factors. A multi-level and mixed-methods study with a final synthesis to m4d as a field -
and ICTD more generally - has opportunities for interactional insights less readily available without
multiple levels of analysis.
1.4 South African case
The m4d apps and services inventory, and to a lesser extent the interviews, are geographically
focused on South Africa as a field site. As well as being the my home country and the site of my
professional experience, South Africa is an appropriate place to study m4d apps and services because it
is at the same time a leading African technology innovator, and host to the kinds of development
problems – and development interventions – m4d apps and services are proposed to solve. ‘First world’
business conditions share space with pervasive poverty and un- and under- employment. In 2008,
54.2% of rural and 21.9% of urban households were below the poverty line (Amstrong, Lekezwa &
Siebrits, 2008, based on the Statistics SA ‘lower bound’ poverty line). HIV prevalence is among the
highest in the world at at 17.8% (UNAIDS, 2012), with poor rural areas bearing most of the burden.
Despite notable achievements in service delivery since the country's first democratic elections in 1994,
progress is hampered by political stagnation, limited fiscal room to manoeuvre and occasionally
effective but uneven civil society engagement (Habib, 2013).
The mobile services ecosystem reflects an unequal society. Mobiles are ubiquitous, with
coverage over most of the country and 81.8% of the population (South African Audience Research
Foundation, 2013), including 80.1% of those below the poverty line, reporting regular use (Statistics
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 26
South Africa, 2012). Internet penetration is tiny by comparison at around 20% (World Wide Worx,
2012), with broadband figures perhaps half of this (Research ICT Africa, 2012). For a small proportion
of the population, smartphones, apps and two-year postpaid contacts are the norm. The rest rely on
prepaid access and use a mix of basic and feature phones with varying app and service offerings. Their
access to mobile services is hampered by cost and technical barriers, as well as general lack of use in
social networks outside of urban youth (Research ICT Africa, 2012). However, services such as 'please
call me' messages, which are advertising-supported and free to send, and youth-oriented chat network
Mxit, have been widely adopted by some groups of prepaid users. 'Please call me' messages may have
generated an estimated 4.5 billion US$ for Vodacom alone (“'Please Call Me' plaintiff got R5m boost”,
12 August 2013). Mxit boasted 6.5 million active users in South Africa in 2013, although by 2014 this
number had dropped to 4.9 million (Thomas 2015).
1.5 Scope and Limitations
In addition to being geographically centred on South Africa, the research described here is
limited to a particular set of technologies. Specifically, the GSM standard, which describes the first
widespread digital cellular mobile networks and underlies the mobile network architecture of much of
the world, is the main focus. Platforms for m4d apps and services on GSM networks are included,
some of which involved other standards processes such as the Java Community Process, but
technologies that originated in third generation (3G) mobile networks are not. Similarly, although
smartphones are becoming common in many developing countries, the intention of this study is to
produce a retrospective overview of m4d apps and services in a pre-smartphone world. The final
research component – the analysis of CommCare metadata - and the conclusion begin to consider the
transition to smartphones, but take a comparative perspective aimed at drawing lessons from the pre-
smartphone ecosystem.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 27
The theoretical framework and research design impose additional limitations. By choosing to
study the mobile network as infrastructure, I am intentionally focussing on technologies and the
processes, actors and institutions that shape them. ICTD often (rightly) advocates focussing on
contexts and people rather than technology, as an antidote to rampant technological determinism that
underlies many of the promises made and broken in the field. By looking within the 'black box' of
technology, I hope to avoid technological determinism while critically considering the place of
technology in ICTD, the mobile ecosystem, and specifically in m4d apps and services.
In terms of research design, and appropriate to an ecosystem-level study, I have chosen breadth
over depth. The four research components employ different methods, but also work at different levels
of the m4d ecosystem and are complementary rather than directly comparable. Within each, the
emphasis is on taking advantage of the retrospective nature of the study to capture the variety of
platforms for m4d apps and services, the breadth of experience of key informants, the range of apps
and services operating in South Africa, and patterns emerging from metadata covering hundreds of
mobile data collection projects. Because of the timing and unique availability of data, it is possible to
produce an overview study that complements the many case, country or sector -focused research
projects that constitute m4d as a field.
1.6 Chapter Outline
Chapter One provides an introductory overview to mobiles for development, and specifically
the ideas surrounding mobile apps and services as development interventions. An overview of the
research approach and the rationale for a multi-level, mixed-methods study are also presented.
Chapter Two describes the conceptual framework, covering the theoretical basis in
infrastructure studies, platform studies and actor-network theory as well as concepts and theories from
the application domain of mobiles for development.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 28
Chapter Three is a general methodology chapter, covering the rationale for and organization of
a mixed-methods approach and describing the applicable research methods for each of the four
components. Where appropriate, additional methodological detail covering data acquisition, analysis
and/or statistical tests is provided in the findings chapters.
Chapter Four describes the historical development of the GSM standard, mobile networks that
implement it, and various platforms for m4d apps and services. The focus is on identifying affordances
and constraints arising from particular technical decisions.
Chapter Five describes and analyses data from the inventory of m4d apps and services
operating in South Africa. In addition to breakdowns by technology, language, target audience and
stage of development, the main finding from this component is that m4d apps and services are mostly
variations on a small number of archetypes. There is some diversity in content, but much less in form.
Chapter Six reports findings from the key informant interviews. It also links interviewee
comments to some of the findings reported in previous chapters, identifying broadly relevant issues for
m4d apps and services in South Africa.
Chapter Seven uses operational metadata from a group of m4d projects corresponding to a
single archetype - data collection, management and reporting for mobile remote workers – to test a set
of hypotheses about how technology choices and implementation support and affect outcome
measures. The main purpose of this component is to consider how secondary datasets from m4d
projects might be used. Data acquisition, cleaning and statistical analysis are described in some detail
to facilitate this.
Chapter Eight restates the findings in the previous chapters into four key ideas about how the
mobiles for development consensus has translated in practice. I then consider how findings from this
retrospective study might inform future ICTD research and practice.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 29
CHAPTER 2: Conceptual Framework
This chapter develops the conceptual framework used to design, conduct and analyse the
research. The framework draws on theory from science and technology studies, including
infrastructure studies, platform studies, actor-network theory, as well as literature from development
studies and ICTD that describes the application domain. The four bodies of literature are used as
follows:
● Work on mobiles and development provides an outline of the “cultural formation” of
m4d apps and services, including different ideas about how (or whether) mobiles are related to
development. Additionally, published and grey literature on m4d mobile apps and services was
consulted to arrive at a preliminary set of factors related to success in mobile apps and services,
which in turn informed the design of the interview guide and inventory categories.
● Infrastructure studies (Star) and platform studies (Bogost and Montfort) suggest ways
to think about how the mobile network, as platform for mobile services, might affect the kinds
of services that are created and the forms they take. Chapter Three describes various platforms
for m4d apps and services, with their history and affordances related to the findings of the
inventory and interview components.
● Actor-network theory (ANT) describes specific cases of technological innovation in
terms of evolving relationships and collaboration between a network of human and non-human
actants. Two particular ANT frameworks - local-global networks and moments of translation –
are described, but ANT ideas such as equal treatment of human and non-human actants and
attention to standards, formats and design complement an infrastructure studies approach and
are generally visible in the conceptual framework.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 30
● Both user-driven innovation/technology appropriation and studies on non-use
complicate the relationship between users and technology objects. This literature mostly
considers end-users as the primary actors, but also can inform studies of platforms where
designers of apps and services innovate through use. Where the design decisions and
affordances of technology platforms represent the “embodied” structure of the mobile network,
user-driven innovation and appropriation provides a way to capture “emergent” properties
(Orlikowski, 2000) that are defined and redefined through use.
Drawing on the material developed here, a visual model of the m4d apps and services
ecosystem is presented at the end of the chapter.
2.1 Mobile phones and development
Links between mobile phones and development have been drawn at multiple levels, with
development understood in different ways (Donner, 2008; Burrell & Toyama, 2009; Heeks, 2010). In
the simplest case, mobile penetration rates are used to demonstrate improved access to communication
as an end in itself. Macro-economic studies, such as Waverman, Meschi and Fuss (2005), have
demonstrated relationships between that mobile penetration and macro-economic indicators such as
GDP growth, while work on small, medium and micro-enterprise (for e.g. as reviewed in Donner &
Escobari, 2010) has considered how mobile use improves livelihoods of small traders and
entrepreneurs in the developing world. Individual livelihoods may also be improved through more
effective job search (Samuel, Shah & Hadingham, 2005), better access to financial services such as
mobile money (Morawczynski, 2009), and more successful engagement with formal institutions and
administrative processes (Skuse & Cousins, 2008).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 31
Demonstrating the utility of mobile phones for maintaining social ties can also imply a
livelihood model, as evident in Diga's finding that mobiles “provided a means of timely responses,
reduced surprises with available information, allowed the ability in multi-task and plan during shocks,
engaged less time to physically search individuals and less emotional stress during really difficult
ordeals” (Diga, 2008). However, there is also a broader well-being argument for the value of
“expanding possibilities for connectedness between people” (Smith, Spence & Rashid, 2011), which
Smith et al make with reference to Sen's concept of development as expansion of freedom by
enhancing human capabilities (Sen, 1999). Goodman (2005) finds simply that, controlling for income,
age, amount of face-to-face family contact and membership of social groups, there is a positive
correlation between mobile ownership and self-reported life satisfaction.
By making it possible for previously disconnected groups to disseminate information and
communicate rapidly and across distance, mobile phones have also been implicated in sustained, large-
scale popular mobilization. In the 'People Power II' protests in the Philippines, mobile phones were
used to mobilize extremely rapidly, and respond flexibly to unfolding events (Castells et al, 2007,
p.184-242), ultimately leading to the ousting of corruption-accused President Estrada. Similarly,
despite social media dominating the discussion in the international press, SMS was the most used peer-
to-peer communications technology during Egypt's January 2011 pro-democracy protests (Wilson &
Dunn, 2011). Without getting into the many nuances of information and communication technologies
and mobilization, these examples imply a way for marginalized groups – particularly youth and the
urban poor - to amplify their grievances in public discourse and demand change (Castells, 2012;
Mason, 2012).
Work in the areas of m-health, education and basic service provision holds that mobiles can
contribute indirectly to development by improving the quality, affordability and accessibility of social
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 32
services. The outcomes – better health, better education, personal safety, time and cost saving –
contribute to both individual/household livelihoods and community wellbeing. Improvements are
generally incremental rather than radical, and downstream impact is predicated on taking particular
services to scale. For example, mobile reminder services for community health workers have
demonstrated efficiency and care quality benefits (De Renzi et al, 2012). However, achieving the
promised contributions to wellbeing requires the incorporation of the community-based care model
within the larger context of the health system (Estrin & Sim, 2010). There are also few cost-benefit
analyses on mobile services interventions in under-resourced settings (Rashid & Elder, 2009). While
mobile services clearly can improve aspects of healthcare and education, their proposed contribution to
development, because it requires that interventions achieve sustainably and reach scale, is contingent
on a host of factors well beyond the technology itself.
This brief review is necessarily limited, but should serve to demonstrate how general
statements on the significance of mobile penetration in the developing world are translated into
specific accounts of user practices, and particular research activities and interventions. These are
linked, implicitly or explicitly, to distinct ideas about development. Authors such as Liang (2010) and
Mercer (2006) are critical of this in itself. Mercer terms the presumption that, unlike the rest of the
world, Africans' use of ICTs should be in the service of development “African exceptionalism”, and
Liang is likewise concerned that “developmentalism” sees the poor restricted to using technology in
instrumental ways. In a related position, Donner (2008) suggests a reframing in which mobiles act as
“capillaries” in Castells' network society (Castells, 1996), “asymmetrically amplifying power in some
places and further diminishing it in others.”
Avgerou (2010) restates the first position as “ICT does not necessarily result in development
for all: it is subject to the power dynamics of information systems (IS) innovation action” and the
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 33
second as “ICT does not necessarily result in development for all: the transfer and diffusion of ICT
leads to uneven development”. She notes that both contain a recognition of the “social embeddedness”
of ICTs and ICT4D. Like Heeks (2008), who cites the demise of the telecenter model as an example,
Avgerou perceives the field as facing “the falsity of widely held technology-deterministic expectations
that ICT, by virtue of its technical properties, will have this or that development effect.”
At the same time, both authors find grounds to retain the belief that ICTs have the potential to
improve the lives of marginalized groups, and that research can be designed to promote this. This
position is somewhat analogous to how Schuurman (2000) responds to what he sees as an “impasse” in
development studies. Post-development scholars such as Escobar (1995) and Rahnema (1997) had
challenged the idea of development as progress, and the presumed homogeneity of the third world,
while globalization has called into question the role of the “hollowed out” state as the primary frame
for, and agent of, development. Schuurman accepts both criticisms, but remains convinced – not least,
he says, by the demands of popular movements – that citizens of the global South still aspire to
progressive improvement through equitable economic, social and political development.
Implications for development research have been drawn in different ways. Schuurman, echoing
Sen's idea of development as freedom, suggests we conceptualize development as addressing
“inequality of emancipation” (p. 14). Brett (2009) urges greater focus on how institutions can be
shaped to improve the lot of the poor. Like Stiglitz (2007), he shows that context matters and supports
“institutional pluralism”, while on the other hand recognizing the challenge posed to institutions by
globalization and related inequality within and between countries. There are many other authors and
many other ideas in this debate, but the key issue is that despite serious reservations about both the key
tenets and 'track record' of development – and, for ICT4D, the contribution of information and
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 34
communication technologies – as a frame it remains widely used in practice and can be productive as a
subject of normative as well as critical research.
The research described here engages with mobile apps and services in terms of discursive
framing, design and implementation, and use and non-use by low-income groups. By doing this, I hope
to contribute to understanding the affordances and limitations of mobile networks as infrastructure for
this group, as well as to the conversation that links mobile apps and services to development. I find
support for this idea in Bebbington (2000):
If research engaged with questions of practice - both popular and bureaucratic - it might become
apparent that the goals, meaning, and power relationships underlying development often differ
from those imputed by much development theory. Power, meaning, and institutions are
constantly being negotiated, and these negotiations open up spaces for potentially profound
social and institutional change. Understanding how these spaces open and how they are used is a
critical research challenge, and will take us beyond some of the oppositions that haunt much
development theory.
(Bebbington, 2000, p.497)
The next section explores what it means to think about mobile networks as infrastructure.
Although not widely used in studies of mobiles and development, Star's work on infrastructure
provides a framework that accommodates complex interactions between people, institutions and
technologies, and can do so at different levels simultaneously (Horst 2013). In relation to development,
is useful because it can trace power through multiple sites, including the social, economic and political
context in which mobile networks have developed as well as local innovation and appropriation by
designers and users. As such, it recognizes the “social embededdness” (Avgerou 2010) of mobile
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 35
services without losing the possibility of normative statements about their role (promised or actual) in
promoting development.
2.2 Mobile networks as infrastructure
The conceptual frame for the proposed research on mobile apps and services considers
different platforms – Voice, short message service (SMS), Unstructured Supplementary Service Data
(USSD), the SIM toolkit, J2ME apps, the mobile web - as part of the technological infrastructure
(Star, 1999) of mobile networks. It is through this infrastructure that the mobile network influences the
design and operation of mobile services in specific ways. This in turn may influence adoption and
forms of use among marginalized users, and the potential for mobile services to contribute to
development.
Star's idea of infrastructure accommodates complex interactions between people, institutions
and technologies, and can do so at different levels simultaneously (Horst, 2013). In relation to
development, is useful because it can trace power through multiple sites, including the social,
economic and political context in which mobile networks have developed as well as local innovations
and appropriations by users. As such, it recognizes the “social embededdness” (Avgerou, 2010) of
mobile services without losing the possibility of normative statements about their role (promised or
actual) in promoting development.
Understanding a technology system as infrastructure implies deep embedding in social
practices – a state of being “sunk into and inside of other structures, social arrangements and
technologies” (Star ,1999, p. 381). In addition to embedding, Star's definition of infrastructure requires
transparency, such that users do not have to consciously “assemble” or “reinvent” the technology
system for each task. Further, an infrastructure must have spatiotemporal reach beyond a single site or
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 36
event (p. 381). Mobile networks, with global reach, widespread use and ubiquity in daily life, clearly
meet this definition.
As a technology system that has become infrastructure (albeit unevenly), mobile networks are
both analogous to, and constructed in relation to, physical infrastructure such as the electrical grid,
road networks and international telecommunications cables. Being “built on an installed base” (p. 382)
enables and constrains an infrastructure in particular ways. Mobile phone networks provide
unprecedented coverage to large parts of Africa, but the location of base stations is constrained by road
accessibility and available electricity supply. Furthermore, a technology system can operate as
infrastructure for some groups but not others. Functions such as text messaging may become intuitive
(perhaps ultimately transparent) for those with basic literacy, but exclude those without.
Bowker, Baker, Millerand & Ribes (2010) propose a model of infrastructure in terms of two
perpendicular “distributions”, with local-global on one axis and technical-social on the other. The
model (Figure 1) provides a useful summary of aspects of infrastructure previously discussed.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 37
There is no 'greenfields' site for infrastructure development, and the process cannot be
understood without historical, technological and social context. This perspective is already at odds with
the kinds of evidence - such as coverage and penetration statistics – that are used to promote mobile
networks as the optimal information and communication platform on which to build services for
previously disconnected users in the developing world. However, as Star (2002, p.111), the point of
studying infrastructure “is not to yell 'aha, a bias! I knew it all along”, but rather to uncover the various
technological and social path dependencies that lead to an infrastructure being understood in a
particular way in relation to the social world.
Figure 1: A distributional model of infrastructure (Bowker at al 2010,
p.101)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 38
Star and Griesemer (1989, p. 390) liken the process of becoming infrastructure to a progressive
“funneling” of interests - “reframing or mediating the concerns of several actors into a narrower
passage point”. Once a passage point is passed, alternative design possibilities appear increasingly
closed off. Eventually, a technology system is becomes stabilized, an embodiment of the actors and
interests that shaped its evolution. To give an example from the mobile network ecosystem, the
vacuum left by the failure of standards such as the Wireless Access Protocol (WAP), which attempted
to transition the mobile network to an open platform for mobile services, had as much to do with
operator and manufacturer reluctance as technical limitations (Palomäki, 2004; Funk, 2007). The
resulting mobile content 'walled garden' embodied these interests.
Looking at the mobile network as infrastructure implies a focus on technology, but should not
be confused with a technologically determinist position. Layered over the “embodied” structure of the
mobile network also “emergent” properties that are defined and redefined through use (Orlikowski,
2000), This is evident in the user-driven innovation literature (Bar, Pisani & Weber, 2007; Sey, 2008;
Horst, 2013), which describes how users and designers have worked around the limitations of mobile
networks in interesting, creative and sometimes instrumental ways. However, ICT4D has tended to
focus on adaption, appropriation and the potential or “emergent” properties for development rather
than critical accounts of the histories and relationships embodied in the technology itself. The taken-
for-granted nature of infrastructure requires explicit work in order for this embodied structure to be
made visible. Interrogating the technology platforms that are common currency in the ICT4D
community is one way to do this.
2.3 Studying technology platforms
Bogost and Montfort's (2007, p.1) definition of a technology platform provides a way to
conceive of the specific relationship between mobile networks as infrastructure and mobile services.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 39
The platform is “whatever the programmer takes for granted when developing, and whatever, from
another side, the user is required to have working in order to use particular software.” Montfort and
Bogost (2009) point out that platforms generally exist in layers – for example, a personal computer
comprises a hardware platform on which an operating system is layered, on which execution
environments for programs written in high-level languages such as Python or Java depend in turn.
Here, I choose to examine the mobile network as infrastructure for mobile services by defining as
mobile services platforms the shared technology protocols that define how devices structure, display
and solicit information from the user, and communicate with the mobile network. Smartphone app
environments, SMS, unstructured supplementary service data (USSD), the mobile web, and non-
smartphone mobile apps are examples of technology platforms on mobile networks.
Montfort and Bogost's (2009, p. 3) intention for platform studies is to examine “how the
hardware and software of platforms influence, facilitates or constrains particular forms of
computational expression.” This requires a technically detailed focus on the features of the platform
itself, and on how they have been exploited (or subverted) by the digital artifacts that build on them.
Understanding a platform as infrastructure also considers how the platform was developed and came to
be adopted, and how particular design decisions came to be made. As Star and Griesemer (1989)
demonstrate, studying the process of becoming infrastructure is an attempt to open the 'black box' of a
technology and expose the actors, relationships and interests that led to its existence in a particular
form. Neither the historical and relational background nor the technical characteristics of the platform
define what can be built on it – creative appropriation is always possible. However, both influence how
designers, and in turn users, encounter it.
Combining platform studies' emphasis on technical detail with the contextual focus of
infrastructure studies is particularly appropriate to the study of m4d apps and services. The
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 40
combination makes it possible to consider not only how platforms interact with material forms of
mobile services, but also the discourses of development in which they it is embedded. Promoting a
technology system as a platform can function as “discursive work” that attempts to funnel diverse
interests into the particular interactions it supports (Gillespie, 2010, p. 347). Examining the “discusive
work” involved in promoting mobile services as tools for development on the one hand, and profitable
new products on the other, could consider how this affects the kinds of services that are available, and
the ways they are or are not used (Maurer, 2012). Also at issue here is how different actors interpret
development. Together with the accumulated embodied structure and emergent practices of users,
discourses and interpretation “format” (Maurer, 2012, p. 594) the mobile services market in particular
ways.
2.4 Actor-network theory in ICT4D
Actor-network theory (ANT) offers tools for understanding collaboration in innovation
processes, giving equal weight to the involvement of technological and non-human objects (for
example, standards and procedures) and human/organizational actors. There is a small body of this
work in ICT4D, with most authors using a particular ANT-based framework to understand the
trajectory of a project in they have been directly involved. The wider IS literature provides a greater
number of examples, although these also generally from analyse a single case.
In addition to exploring drivers and barriers that exist in specific ICT4D projects, approaches
derived from ANT are attractive because they resist attributing the result of innovation processes to
single actors. For example, Andrade and Urquart (2010) find multiple misaligned incentives in a rural
telecenter project, culminating in cascading failures that eventually doomed the project. This leads
them to characterize the project as a failure to “impose a network over other networks” (Andrade &
Urquart, 2010, p. 370) in recognition of the competing priorities of key implementing partners, who
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 41
proceeded with the project despite not having a common goal. Heeks and Stanforth (2007), writing
about a failed eGovernment project in Sri Lanka, find the project stalled for similar reasons, with
heterogeneous actors and diverse interests unable to be reconciled into a shared project definition.
ANT also demands that authors recognize power explicitly – including their own as creators of
a narrative account. In her study of a South African telecenter managed by a womens' co-operative,
Rhodes (2009) characterizes the unravelling of the project as resulting from a failure to match its
design to underlying social realities. Uneven gender and economic power structures are painfully
revealed as project culminates in witchcraft accusations against the telecenter management, and,
tragically, the murder of a security guard during a robbery.
While most ANT studies produce an after-the-fact account of projects for an academic
audience, Heeks and Stanforth (2007, p. 175) note that they can also help actors understand their own
role in the project in relation to the rest of the network. Project-based work is also not the only way to
use ANT (and related approaches) to study ICT4D problems – for example, it could equally be applied
to common examples of user-driven innovation such as SIM swapping (Horst, 2013). One way of
looking at this contribution is as an explanation for why innovation is not (or not just) a function of the
'marketplace of ideas' – why ideas that 'make it' are not necessarily the best by any chosen standard.
This is particularly important in ICT4D where projects routinely negotiate complex local and
international relationships, cross cultural boundaries and involve partners across government, industry,
funders and the user community.
To illustrate concrete applications of an ANT approach, I describe two specific ANT-based
conceptual frameworks. Local-global networks (Law & Callon, 1992) and moments of translation
(Callon 1986) concentrate respectively on the resourcing and trajectory of particular intervention
models.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 42
Law and Callon (1992) separate the actor-network into two components: the local network and
the global network. The local network comprises actants directly involved in implementation. The
global network is the collection of actants and relationships that “generates a space, a period of time,
and a set of resources in which innovation takes place” (Law and Callon 1992, p.21). These are the
actants that need to be mobilized to create the conditions for implementation. Both the local network
and the global network must have the appropriate actants enrolled to sustain the project, and ensure
that their interests remain aligned throughout its lifetime. In addition, Law and Callon’s example of the
failure of an aircraft carrier development project highlights the importance of having the project as a
central connector node - an obligatory point of passage in ANT terms - between the local and global
networks. In the case of the aircraft carrier, relationships existed beyond the project between certain
actors in the local and global networks, allowing them to circumvent formal project management
structures.
The separation of actants into local and global networks provides a conceptual vocabulary for
the effects of network composition and structure on project outcomes. Callon’s moments of translation
(Callon,1986) provide a parallel vocabulary for stages in a projects’ trajectory. Callon defines four
moments of translation, summarized by Rhodes (2009, p. 6) as follows:
• During problematization, “an actant or actants defines the problem so that other actants
recognize it as their own problem.”
• During interessement, “actors convene around an issue [...] using devices to detach actants from
elsewhere and attach them to this point of view. It also involves translation, strategic
compromise and persuasion to lock allies into the proposed roles.”
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 43
• Enrollment is the process of growing and strengthening attachment in the network, including
translating the interests of actants to accept a particular spokesperson or group, and
“maintaining alignment through constant attention”.
• Finally, mobilization entails ensuring the continued position of the project as an obligatory
passage point, and, as outcomes are generated, maintaining the legitimacy of the spokesperson.
While perhaps best understood as iterative or parallel processes rather than a linear
progression, each moment defines a standard for successful completion, and suggests strategies
employed by actants to achieve this standard. Over time, stabilized networks are those that
“persistently act as one” (Faik, Thompson & Walsham, 2013) to the extent that they become viewed as
a 'black box', taken for granted such that the actants and relationships within are obscured. An ANT
approach requires opening the black box to uncover how the network developed, and the frameworks
described above provide a conceptual language to do so in terms of actants, interests and relationships.
Walsham (1997) points out that as described by Latour, ANT prescribes not just an analytic
approach but also a particular orientation in research design and methodology. In particular, Latour
(2005) advises that description, as much as possible in the attributable to the actors themselves, should
be preferred to explanation in any pre-conceived form. Like an experiment, the resulting analysis
should be valued for what, from its coherence as a text, it allows us to do. ANT studies in IS (and
ICT4D) often allow the author to explain the success or failure of a project in which they have been
directly involved, as a form of participatory action research.
2.5 Innovation and Appropriation
Both literature and interventions dealing with ICT use in the developing world have been
criticized for presenting users as applying ICTs solely in pursuit of 'developmental' ends, such as
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 44
increased crop yield, economic productivity, or access to services (Mercer, 2005; Burrell, 2012). While
such uses of mobiles are common and significant for at least some populations, ethnographic studies in
particular (Horst & Miller, 2005; Hahn & Kiborra, 2008; Skuse & Cousins, 2008; Ureta, 2008; Sey,
2011) reveal a much wider array of uses, negotiated in complex and contingent ways. For example Sey
(2011) argues against separating social and economic uses, and notes that for most respondents in her
study in Ghana, connectivity for multiple uses was the norm.
In addition to anticipating diverse and overlapping uses, the literature on appropriation and
user-driven innovation identifies creative and transformative uses of mobile phones, including by users
in under-resourced settings. While recognizing that severe constraints remain, it challenges the 'digital
divide' narrative of technology transfer by acknowledging the agency and creativity of the poor.
Authors such as Bar and Riis (2000) and Heeks (2008) also point out that, consistent with Von Hippel's
findings (2006) in the US, users in the developing world have originated innovations that have gone on
to be very successfully commercialized as mobile services. Mobile money transfer, originating in the
practice of sharing airtime voucher codes good for a certain value by text message, is a well-known
example, as are 'please call me' messages, which commercialized calling but hanging up before the call
is answered as a way to communicate without being charged.
Bar, Pisani and Weber (2007) suggest understanding appropriation and user-driven innovation
in the form of the “technology evolution cycle”, comprising adoption (users do “old things in new
ways”), then experimentation (users “do new things in new ways”) and re-configuration (“social
practices are transformed”). Deliberately designing technologies for appropriation provides a natural
entry point for users into the technology evolution cycle. Creating a social space where innovation and
appropriation is encouraged – the authors use the example of Brazil's Pontas de Cultura telecenters –
can yield a similar result.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 45
The implication of design for appropriation is that technology is not neutral, and embodied
power relationships can and do constrain user-driven innovation. The act of appropriation can imply a
challenge to established power structures as embedded in consumer technology (Bar, Pisani & Weber,
2007), but the characteristics of particular technology platforms can restrict certain innovative practices
at the same time as they enable others. Users in resource-limited settings seem more likely to arrive at
mobile technology innovations through “bricolage”, or using existing technological artifacts in new
ways (Van Der Boor, Oliveira & Veloso, 2013). How this translates to m4d apps and services, which
typically anticipate a relatively narrow set of uses, provides an interesting counterpoint to the
sometimes extremely successful user appropriation of basic mobile communication.
One way to explore this may be to study non-use of mobile services. Selwyn (2003, 2006)
points out that the discourse of the digital divide denies agency to non-users by presuming that their
non-use is simply a result of restricted access, to be progressively overcome. Satchell and Dourish
(2009) elaborate other possible reasons for non-use. These include lagging adoption as envisaged by
the digital divide but also active resistance (they use the example of living 'off-grid'), disenchantment,
disenfranchisement, disinterest, and displacement, of which Sey's study of the informal shared
payphone model is one example (Sey, 2008). Investigating non-use of, for example, mobile money
services might reveal motivations not envisaged in the promise of providing banking services for the
'unbanked'. Like appropriation and user-driven innovation, studying non-use is a way of understanding
the way users negotiate their relationship to technology, both working within and going beyond the
forms of use imagined by designers.
User practices exist in relation to the wider context of the mobile services ecosystem. Horst
(2013) uses Star's idea of infrastructure as a lens on mobile user practices, understood as part of a
larger socio-technical system. To the extent that particular practices are widespread in a particular
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 46
group, she suggests that users' negotiation of technology use (and, we could add, non-use) creates an
“informal infrastructure” that exists in relation to the “formal infrastructure” of the mobile network.
In particular, I consider how service designers and developers, as users of platforms for m4d
apps and services, engage in activities that produce informal infrastructures. Platforms for mobile apps
and services are sites of innovation and unanticipated use as well as sources of (more or less malleable)
affordances and constraints. Knowledge-sharing and collaborative activities in the mobiles for
development space are in themselves part of the socio-technical context through which infrastructure is
formed. As with user-driven innovation, we might ask if it is possible to design for – or otherwise
support - appropriation at the platform level.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 47
2.6 Framework
Figure 2: Infrastructures, platforms and actants in m4d apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 48
Figure 2 is a conceptual representation of the ideas presented in this chapter, relating them to each
other and to the realization of m4d apps and services in the pre-smartphone era. Specifically:
• The mobile network is an infrastructure for mobile communications. It is “sunk into
and inside of other structures, social arrangements and technologies” (Star 1999, p.
381), including earlier and contemporary telecommunications systems, the electrical
grid, the road network and many others, and its design and function are mutually
constituted with them.
• Platforms for mobile apps and services provide interfaces to mobile network
functionality. Platforms, through their design, realization and management, variously
enable and constrain the artefacts – in this case m4d apps and services - that are built
on them.
• At the same time, platforms for m4d apps and services are not fixed; rather, they are
shaped through innovation, adaptation and appropriation by both users and service
designers/developers
• M4d apps and services develop through the actions of a network of human and non-
human actants, which the ANT principle of symmetry requires be considered together.
• Actants can be separated into local and global networks. The local network comprises
actants directly involved in implementation, such as developers and designers,
implementing partners and users. The global network is the collection of actants and
relationships that “generates a space, a period of time, and a set of resources in which
innovation takes place” (Law and Callon 1992, p.21), including funders, infrastructure
partners, policymakers, and others.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 49
This high-level framework is useful in locating the various research components in relation to each
other and the ecosystem as a whole. It also identifies ideas from the literature that have been used to
guide the research design described in Chapter Three. While the findings reported in Chapter Four
through Chapter Seven are organized as appropriate to the research design, elements of the framework
return in the discussion at the end of each chapter to link them into a single analysis.
The research components map to the framework roughly from the bottom up, with Part 1 dealing with
mobile network technologies, Parts 2 and 3 with m4d apps and services and the various human and
non-human actants that shape their development, and Part 4 with use patterns and evaluation
techniques for a single m4d app. Figure 3 relates research components to the framework.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 50
Figure 3: Research components mapped to framework
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 51
CHAPTER 3: Methods
3.1 Research design in infrastructure studies and ANT
Infrastructure is by definition difficult to study. It is obscured by familiarity, and exists in
relation to people, technologies, organizations and institutions and other infrastructures. It is “messy”
and “boring” (Star, 2002, p.109). It extends across space and time, defying a clearly bounded field site,
but is encountered through local and specific interactions.
Research approaches to the study of infrastructure generally involve following a particular
'thing', such as a standard or classification scheme. Star and Griesemer (1989) particularly encourage
the study of “boundary objects”, where interdisciplinary collaboration in the development of objects
that “inhabit several intersecting social worlds” (p. 393) reveals disciplinary assumptions and tensions.
Examples of boundary objects include repositories (e.g. library, museum), ideal types (e.g. diagram,
atlas, description), coincident boundaries (map layers), and standardized forms ("methods of common
communication across dispersed work groups") (p. 410-411).
Star (2002) is similarly concerned with the contested meanings that emerge at “fringes” where
different spheres of experience intersect. Broadening beyond the study of particular objects, she offers
two examples of systems development projects. In the first, a contested understanding of the system
results from differences of meaning between users, designers and the organizational and institutional
processes through which they encountered the system. In the second, a digital library project,
homonyms and appropriated/reclaimed meaning in words such as “nigger” and “queer” reveal
disciplinary disparities as system designers unfamiliar with sociological and library science approaches
encounter them for the first time.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 52
Like Star's systems design projects, m4d apps and services are sites of contested meaning. The
promise of m4d originates at a confluence of interests, including those of the development industry,
mobile technology companies, ICTD researchers and previously unserved users of mobile
communications. The multiple relationships involved defy representation in a 'top-down' model of
development intervention. A network, shifting over time, is a better model (Heeks, 2013, p. 15). The
“fringes” that emerge in the realisation of m4d apps and services illuminate the various interests at
play, and the implications of their enrolment in the network.
Research on infrastructure is usually ethnographic in the sense that it is oriented towards
observation and explanation, rather than hypothesis-testing, and to sense-making though the
convergence of detail. Uncovering “second and third-order” (Star, 2002, p.116) meanings underlying a
relational formation of multiple actors, simultaneously local and global, human and non-human,
technical and social, requires detailed data, multiple angles and extended immersion on the part of the
researcher. While no particular method is prescribed, the focus on intertwined contexts and particular
“boundary objects” means historical research and documentary sources feature prominently,
supplemented by interviewing and observation.
The concept of the field site is reconfigured along the lines of Marcus' “multi-sited
ethnography” (Marcus, 1995). In addition to extending across space and time, infrastructure is
coexistent with the “world system” such that it (the world system) “is not the theoretically constituted
holistic frame that gives context to the contemporary study of peoples or local subjects closely
observed by ethnographers, but it becomes, in a piecemeal way, integral to and embedded in
discontinuous, multi-sited objects of study” (Marcus, 1995 p.97).
In both ANT and infrastructure studies, the researcher is often closely involved in some aspect
of the work being described. They may position themselves as a participant-observer or draw on
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 53
personal experience to produce a retrospective analysis. The ethnographic requirement of reflexivity
applies, in the general sense of recognizing how the researchers' place in the social world shapes their
understanding of it as well as in relation to their involvement with specific subject of study.
Given the origin of science and technology studies in laboratory studies – essentially, research
on doing research – it is not surprising that reflexivity is extended to the process and product as well.
Where Hammersley and Atkinson (2007, p.15) point out that reflexivity in ethnographic research
means understanding how the findings may be used – for example to support a particular political
agenda – Latour (2005 p.128) sees an ANT account as a text in which each point “may become a
bifurcation, an event, or the origin of a new translation.” An ANT account is to be judged by what it
allows us to do (p.123).
3.2 Study design
Methodological aspects of ANT and infrastructure studies are important to the research design
for this study because, in contrast to the usual separation of theory and method, both are concerned
with how to do research as well as how to interpret and analyse the results. As such, development of
the conceptual framework in Chapter 2 requires a parallel argument on methods. This is particularly
true because most research on development interventions is normative to some extent. The underlying
idea is that development theory and development practice should inform each other. ANT by contrast
has been criticised as lacking a moral or political agenda, being “seen as a tool to understand the
world, not to change it” (Heeks, 2013, p.9).
In relation to theory, I aim to pragmatically employ theories and concepts from SST to
understand m4d apps and services specifically and the role of technology in ICT4D more broadly. In
the tradition of development studies I am also interested in producing recommendations for practice.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 54
This involves translating theoretically driven insights into clearly described, situated ideas accessible
to people with varied prior experience. Both are compatible with a broadly ethnographic stance, and
with developing research products that are theoretically and/or practically actionable.
The question of scale is more complicated. Most ANT accounts describe a single project in
great detail, following actants and providing extensive description. Similarly detailed, intensive work
is associated with infrastructure studies, generally centred on a single object. The way Roy (2010)
identifies her subjects of study is closer to my approach to studying m4d apps and services: a set of
commonly accepted ideas about a particular model of development intervention, realized in material,
discursive and relational forms. I am attempting to describe a composite subject, at multiple scales
simultaneously and through different research activities. As such the main focus of the research design
and the analysis is identification of commonalities and cross-cutting issues.
Table 1 summarises the methods, level of analysis and research questions for each of the
research components, and indicates how they relate to each other.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 55
Part Methods Level of analysis Research Questions Links to other parts
1 The GSM standard
and platforms for
m4d apps and
services
Documentary sources Mobile network
technology/infra
structure/platfor
ms
What design logic shaped the GSM
network, and platforms for m4d
apps and services? How does this
influence the functionality and
limitations of the platform as
encountered by designers and
developers of m4d apps and
services?
Used to formulate
interview questions about
affordances/constraints of
particular technology
platforms. Raised issues of
operator control and cost-
to-user, included in
inventory questions.
2 Apps and services
inventory
Qualitative description,
classification
Realisation of
m4d apps and
services in a
single country
What does the pre-smartphone m4d
apps and services landscape look
like in South Africa? What points of
concentration, exclusions and
commonalities exist?
Source of contacts for
interviews. Used to
identify competitors /
similar projects to discuss
with interview
participants.
3 Key informant
interviews
Semi-structured
interviewing,
participant observation
Overall m4d
ecosystem,
mainly focused
on South Africa
Why does the pre-smartphone m4d
apps and services landscape look the
way it does? How are projects
initiated, how do they develop, how
are they evaluated? What are the
forces shaping the field?
Identified apps and
services for inclusion in
inventory. Discussions
about implementation
process lead to one of the
hypotheses in part 4.
4 Quantitative analysis
of operational data
Statistical analysis Single
app/model – data
collection and
reporting for
remote workers
What opportunities and pitfalls exist
in the widespread practice of using
operational data to evaluate m4d
apps and services? What might we
learn, and what might we kit cause
us to miss?
Tests two quantitative
hypotheses derived from
findings on affordances of
different m4d platforms
(part 1) and
implementation issues
(part 2).
Table 1: Research design summary
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 56
Finally, I should note that although none of the research components reports on participant-
observation as a main source of data, as my research stance is that of a participant-observer. I have
worked in m4d as both a software developer and researcher, as well as writing practice-oriented pieces
on mobiles and social change for MobileActive.org from 2008-2012. During 2014, I developed a
research data platform for Dimagi, through which I gained access to the operational data analysed in
Part 4. For the key informant interviews (Part 3), I have drawn on experience building and
implementing m4d apps and services in South Africa to build the initial contact lists and draw up the
interview topic guides. The inventory exercise (Part 2) followed a similar process, and in researching
platforms for m4d apps and services (Part 1) I started from a general understanding of the issues from
an implementer perspective.
3.3 Part 1: The GSM standard and platforms for m4d apps and services
The history and significance of the GSM standard is relatively well described, and literature
research provided a straightforward entry point to the first research components. In addition to
academic work, technical overviews of the architectures developed in various GSM phases were
consulted, as well as a particularly thorough account of the standards-setting process (Hillebrand,
Rosenbrock & Hauser, 2013) sponsored by the European Telecommunications Standards Institute
(ETSI). ETSI also makes available an full database of standards documents relating to GSM, which are
useful in establishing the original reasoning behind various design decisions.
Post-GSM platforms – the mobile web, Mxit and J2ME – are mostly described in technical
rather than historical or social contexts. My main sources for J2ME were technical documents targeted
at developers, which explain the limitations of the platform from a developer perspective. For Mxit,
South African trade website MyBroadband (http://www.mybroadband.co.za), which maintains a large
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 57
archive and regularly publishes articles on South African mobile companies, was a useful starting
point.
3.4 Part 2: Apps and services inventory
One of the central motivations for this study was the realization that ICT4D is nearly always
forward-looking. Research that investigates (and interventions that demonstrate) how a particular
technology could contribute to development, or draws lessons from a single case or a set of cases is
relatively common. Cross-sectional retrospective studies are harder to find. By aiming to produce a
relatively complete account of m4d apps and services in South Africa I hope to be able to draw lessons
from patterns across multiple cases, some of which may generalize to m4d in other countries and to
ICTD generally.
Several similar projects are worth mentioning. Hellström (2010) describes various m4d apps
and services across agriculture, health, education, financial services and governance sectors. Hellström
uses a few case studies from each sector to demonstrate innovation and explore current barriers to use
and requirements for future scaling. A World Bank report (Qiang, Yamamichi, Hausman, Miller &
Altman, 2012) focusing on mHealth compares selected mHealth apps and services in three countries –
Haiti, Kenya and India – by age, type of health intervention and implementing organization. The report
is oriented towards forecasting and recommendations for future expansion of the sector, and does not
provide information on how the included apps and services were selected. For South Africa, Research
ICT Africa (2012) provides an overview of apps and services for the bottom of the pyamid (BoP)
including user surveys and focus groups identifying drivers and barriers for individual use. Like
Hellström, they give examples of m4d apps and services across various sectors, but do not attempt full
coverage.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 58
Donner (2009) reports on an inventory exercise that does attempt full coverage, geographically
restricted to Africa but limited to “mobile livelihood services” for farmers and small entrepreneurs. The
data comes from two online databases of m4d apps and services as well as reviews of the field by
development agencies and a small number of published research studies. At the time of his analysis,
many of the services were in the pilot phase, and the discussion is focused on technical and
implementation issues affecting scaling up.
The GSMA, as part of their Mobiles for Development Impact programme, has complied a
dataset of “mobile-enabled products and services in the developing world” (Hatt, Wills & Harris, 2013)
which is used to produce descriptive statistics for the field of m4d. They list 1870 deployments in total,
with 128 in South Africa. Their focus seems to be on listing as much as possible rather than
establishing a clear definition of a deployment (or a 'mobile-enabled product or service'), so it includes
projects that use the mobile network for connectivity but are not accessed on a mobile phone as well as
operator-provided services that do not make any claim in relation to development. Breakdowns by
technology, sector, business model and type of organization are presented. In keeping with the GSMA's
function as an industry body for GSM operators, much of the analysis is forward-looking, identifying
trends of interest to mobile operators looking to value-added services in the developing world as a new
revenue stream.
The present study aims to contribute a comprehensive inventory of m4d apps and services in
the non-smartphone GSM ecosystem, across all sectors but for South Africa only. For this purpose, I
define an m4d app or service as any app or service that makes either an implicit or explicit
development claim. This includes apps and services that improve services to poor communities e.g.
healthcare, education, basic infrastructure as well as apps that target materially or otherwise
disadvantaged groups with the aim of ameliorating some aspect of the challenges they face. The unit of
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 59
analysis is an app or service rather than each site at which it is deployed, although for software-as-
service architectures where significant customization occurs this is sometimes a difficult line to draw.
In these cases, I have counted each deployment only where deployments that are publicly represented
as distinct products, noting that counts for the category that is most affected – mobile data collection
and reporting apps – need to be examined with this in mind.
Building the initial list of apps and services for the inventory process involved a combination
of existing databases, grey literature (particularly reports from development agencies) and
contributions from interview participants. For the databases, I accessed a MobileActive's curated
database list m4d apps and services by obtaining data from the now-defunct MobileActive.org site
through the Internet Archive (https://archive.org/web/), and scraped the GSMA's list of South African
apps and services. For Mxit apps, the in-app navigation was used to identify m4d apps and services,
which are mostly made available under the Mxit Reach menu item. For this step, the goal was to
produce an exhaustive list rather than to strictly apply exclusion criteria.
To construct the inventory, a Google Form was developed for the dimensions in Appendix B.
For each app or service I conducted a web search by name, as well as a literature search using Google
Scholar. Where possible, I also accessed each app or service directly, although this obviously works
better for those whose intended audience is the general public. In addition, interviewees with
knowledge of the various apps and services were asked to confirm certain information. While I could
not always verify information found online, it was usually either authored by organizations involved or
by journalists or researchers who had talked to them, thus representing at least the public face of the
project. In conjunction with a confirmation of the basic functionality by accessing the relevant app or
service, this was generally enough to establish most of the dimensions required.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 60
The dimensions themselves are of three types:
• Basic descriptives, including a free-text description of the service, organizations involved,
sector and current status (pre-pilot, pilot, at scale) are commonly used to compare and/or
enumerate m4d apps and services, including in the similar work previously described.
• Technology and platform issues, including the technology platform(s) used and particular
platform-specific features as well as mobile operator involvement. Dimensions of interest were
drawn from interviews and my own experience as well as particular affordances of platforms
described in Chapter 4.
• Characterization of the app or service as a development intervention, including intended users,
development 'theory of change' (“an articulation of how and why a given intervention will lead
to specific change” [Stein & Valters 2002, p.2] ), and claims/measures of success. A judgement
on whether, based on the claims presented and the change it sought to achieve, the app or
service could be considered successful was included in the initial draft, but turned out to be
unworkable given the limitations of publicly available information.
Given that the main outcome measure (degree of success) could not be reliably determined from the
information made available by most projects, the dimensions do not provide a foundation for causal
claims. They do, however, allow for description and selection of cases (for example, considering all the
apps and services build on a particular technology platform) as well as relationships between
dimensions such as technology platform and intended users. As Bailey (1994, p.11) points out,
classification exercises are the only method in the social sciences where relationships can be presented
together with the data itself. By combining free-text description with exclusive or non-exclusive
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 61
categorization, it was also possible to iterate on the dimensions to explore patterns emerging in the
data, thus involving classification in the analysis as well as the data collection.
In addition to the various descriptives discussed in the findings, during the inventory process it
emerged that many m4d apps and services can be viewed as variations on a small number of
archetypes. There is some diversity in content, but much less in form. The archetypes can also be
related to sector, technology platform and project structure. For example, Mxit apps are mostly
information services, while SMS is used for reminders, and Android and mobile java for data
collection. Mxit projects mostly work with intermediaries outside the m4d space while mobile data
collection and SMS projects usually use one of a small number of intermediaries with experience
navigating the challenges of the platform and of development projects. To explore this idea further,
categorization by archetype was performed during the analysis phase. The archetypes themselves were
derived by collecting similar apps and services together, not requiring that every app or service match
an archetype.
3.5 Part 3: Key informant interviews
The initial set of key informant interviews took place mostly in February 2014, with a second
set of interviews in February 2015. Fifteen interviews were completed from an initial contact list of 27
potential interviewees. Interviewees were selected based on involvement with long-running or well-
known m4d apps and services, or because of their professional knowledge of m4d as a whole. The
inventory task, which ran concurrently with the first set of interviews, was also used to identify
potential interviewees.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 62
As expected, among non-participants refusal was less common than non-response. People I
could establish a personal connection to from my time working in South Africa were more likely to
respond and to agree to be interviewed.
In addition to the interviews, I attended a two-hour open session hosted by the GSM
association (GSMA) at the ICTD conference in Cape Town in December 2014. The session was titled
“Scaling mobile for development: the importance of partnerships to successful business models” and
included participants from the GSMA as well as developers and implementers of m4d apps and
services, with discussion mostly in a round table format.
Interviews took place either in person or via Skype, and were recorded on obtaining consent
from participants, then transcribed. A general topic guide was prepared (see Appendix A) and adapted
for each participant prior to the interview. Notably, the topic guide initially anticipated asking most
participants about a single app or service with which they were directly involved. While some
interviews did follow this format, I was also able to interview several people with long experience and
high-level knowledge of the m4d ecosystem, adjusting the interview content accordingly. Although
representatives from both the MTN and Vodacom Foundations were approached, as well as from
WASPA, the industry body representing value-added service providers, none agreed to be interviewed.
Analysis of the interview data was conducted using Dedoose
1
, an online content analysis tool.
Each interview was thematically coded so that related content could be considered together when
developing the findings. The topic guides were used to derive an initial set of codes, which were then
revised based the content of the interviews themselves as well as emerging issues in the inventory
findings.
1 Dedoose Version 5.0.11, web application for managing, analyzing, and presenting qualitative and mixed method research
data (2014). Los Angeles, CA: SocioCultural Research Consultants, LLC. Available at http://www.dedoose.com
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 63
3.6 Part 4: Quantitative analysis of operational data
The fourth research component is a narrowly focused investigation into whether secondary data
from m4d apps and services can be used to test hypotheses about technology choice and project design.
The dataset consists of operational metadata from several thousand projects using Dimagi's CommCare
platform (http://commcarehq.org/), of which 280 projects were included in the cleaned dataset for
analysis. The hypotheses were derived by considering the variables available in the dataset alongside
early findings in the interview and inventory research components. There is reasonably good published
evidence that CommCare and similar mobile data collection/remote worker support apps and services
can improve outcomes in programmes with remote workers (see Chatfield, Javetski, Fletcher & Lesh,
2014 for a summary). It is therefore appropriate to consider the conditions under which improved
outcomes are most likely to be achieved.
Access to the dataset came through contract work I did for Dimagi from June-December 2014,
which involved building a data platform for internal research and business process improvement. In the
Chapter 7 I describe the data, hypotheses and analysis performed. Research involving secondary use of
data is both demonstrably tricky (Botsis, Hartvigsen, Chen & Weng, 2010; Ancker et al, 2011; Hersh et
al 2013) and relatively new in ICTD. Because of this, I also include a detailed description of the
process. The dual function of this research component – testing specific hypotheses while also
exploring the idea that secondary analysis of operational data from m4d apps and services can be used
to evaluate interventions – makes it preferable to present methods and findings together.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 64
CHAPTER 4: The Mobile Network
This chapter provides an overview of the mobile network, highlighting the features that
constitute platforms for m4d apps and services. In accordance with an infrastructure approach, the
analysis begins from the historical and institutional context from which modern mobile networks
emerged. The globally dominant GSM standard is particularly important in this regard.
Economic, technical and historical perspectives on GSM are well-represented in the existing
literature (Agar 2003; Farley 2005; Dunnewijk & Hultén 2008; Redl, Weber & Oliphant, 1998), with
significantly more detail than presented here. The main focus of this chapter is to highlight key
attributes of GSM networks that constitute platforms for m4d apps and services. The sections that
follow consider how platforms for m4d apps and services developed and with particular focus on how
they were conceptualized in, and shaped by, the GSM standards process.
4.1 Origins of GSM
Launched in 1982 as the Groupe Spécial Mobile (later Global System for Mobile), the GSM
standards process is widely regarded as a successful model of multinational public-private co-
operation. It also demonstrates the role of political and economic imperatives in infrastructure design.
While mobile telephone systems and the cellular concept were already in use elsewhere in the world, it
was the interoperability imperative that set GSM apart and arguably lead to its global success. Driven
by the European Commission, it reflects a founding logic both symbolically and practically invested in
creating a unified communications system for a unifying Europe. It was also a successful
demonstration of the competitive advantage of a single European market, superseding fragmented
national systems (Agar, 2003).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 65
Where initially membership comprised mostly incumbent monopoly operators from each
country, technical complexity and the innovation required by a digital rather than analogue standard
(including associated intellectual property issues) lead to equipment manufacturers being included in
the pre-rollout phase (Rovinen & Ylä-Anttila, 2004). This was facilitated through 'basket model' of
standardization that specified interfaces between components rather than their internal functioning.
In contrast to the US, where the AMPS standard described communication between the handset
and the base station only, the GSM standard was also a complete specification covering all aspects of
the network (Agar, 2003). The technical challenge was thus concurrently deep and broad, and both
Agar and Rouvinen and Ylä-Anttila document informal references to the 'Global Software Monster'.
The European Commission was able to mitigate the private sector risks resulting from the complexity
of the proposed system with political will in the form of a Memorandum of Understanding signed in
1987, in which operators in thirteen European countries agreed to implement GSM by 1991. This, in
conjunction with the basket model of standardization, made it possible for both manufacturers and
R&D co-operations in Europe to commit to the standard (Dunnewijk & Hultén, 2007).
To ensure participating operators were able to roll out functional networks by 1991, a phased
approach to the specification was agreed on. The first phase (Phase 1) included basic call functions, as
well as international roaming which was still the showpiece of the standard (Hillebrand, Rosenbrock &
Hauser, 2013). The successful first-phase rollout and agreements leading up to it gave GSM the
momentum it needed to retain commitment from equipment suppliers. A phased approach meant that
technical issues arising from the implementation of a complex and significantly novel standard could
be addressed in the Phase 2, along with outstanding lower-priority aspects such as SMS (Hillebrand,
Rosenbrock & Hauser, 2013). The basic architecture agreed in Phase 1, on which Phase 2 was also
based is shown in Figure 4.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 66
Completion of Phase 2 was accompanied by the realization that, as a result of technology
advances and also the successful rollout, at scale, of GSM networks, enhancements to the now
completed standard where nevertheless required. Successive releases of the updated standard occurred
between 1997 and 2000 under the umbrella of GSM Phase 2+. Reflecting the rapid expansion of GSM
networks beyond Europe, Phase 2+ was notable for a change in terminology, with Digital Cellular
System replacing references to a European System in standards documents (Redi, Weber & Oliphant,
1998). In fact, some of the Phase 2+ work items were a result of pressure from Nokia, whose global
expansion required features not prioritized by the European market such as extended character sets for
SMS (Hillebrand, Rosenbrock & Hauser, 2013). The developing world (with emerging Asian
economies leading the way) embraced mobile largely to provide cost-effective access to
telecommunications for areas where fixed line services where non-existent or extremely expensive,
Figure 4: GSM Network Architecture (Hillebrand, Rosenbrock & Hauser 2013)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 67
although cost-effective international roaming was also seen as politically attractive (Redi, Weber &
Oliphant, 1998, p. 15).
The introduction of data services in the form of the General Packet Radio Service (GPRS) and
later Enhanced Data Rates for the GSM Evolution (EDGE) were perhaps the biggest story of GSM
Phase 2+ (Hillebrand, Rosenbrock & Hauser, 2013). GPRS made it possible to bill users for the
amount of data used rather than time online, and was developed (although by a separate working
group) in conjunction with the Wireless Access Protocol (WAP). WAP was meant to make Internet
services usable with limited bandwidth on a small-screen device, although in practice is was never
particularly successful (Agar, 2003). Nevertheless, by 2000 it was clear that demand for mobile
Internet was and would remain high.
From the point of view of third parties wanting to offer services on the mobile network, the
addition of Intelligent Networks (IN) components was another significant addition in GSM Phase 2+.
IN was developed to support value-added services in landline networks, including toll-free numbers,
reverse billing, and premium information and entertainment services, often provided by third parties.
In mobile networks, IN concepts were standardized as CAMEL, with releases each year of GSM Phase
2+. Prepaid billing models, which require that the user's balance be checked in real-time prior to and
during a call, are possible because of the provision in CAMEL for queries to external systems such as a
prepaid balance management system.
Sauter (2010, p. 58) likens CAMEL to the content-unaware design of HTTP in that it provides
a standard way for services on the mobile network to be developed and to communicate with other
network components but does not concern itself with the specifics of any particular service. The
introduction of CAMEL was motivated partly by operators' desire to differentiate themselves in the
market through provision of value-added services (Hillebrand, Rosenbrock& Hauser, 2013). As a result
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 68
– in fact by design – available value-added services vary by network, as do the possibilities for new
services built on operators' proprietary IN platforms.
With GSM networks established around the world and convergence between telephony and
data services and equipment the site of much innovation and speculation, standards-setting beyond
GSM Phase 2+ was a game with much higher stakes. Not only did the new standards require new
spectrum allocation, but both the number of actors in both public and private sectors had greatly
expanded, as evident in Figure 5. Hess and Coe (2006, p. 1219) describe the situation as a “pluralistic
model of standard setting, shaped by complex networks of state and corporate entities interacting at
different spatial scale”, driven by an increasingly globalization and deregulated telecommunications
sector as well as new models introduced by convergence.
Figure 5: Actors in Standards Setting after GSM Phase 2+ (Hess & Coe, 2006, p. 1219)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 69
In the sections that follow, I cover particular features of the mobile network that comprise
various platforms for m4d apps and services. However, the historical development of the GSM
standard and rollout of GSM networks in general is relevant to m4d for several reasons. First, it should
be clear from the above that the main target market in the initial GSM concept was European
businesses users. Success in the developing world was partly a result of liberalisation and innovation in
pricing models and handsets, but mostly a recognition of great latent demand for telecommunications
that mobile was able to address. As Hess and Coe (2006, p. 1224) point out “value enhancement and
capture are still predominantly realised within the established markets of the advanced economies.”
Second, mobile telephony was in the beginning fundamentally a telephony project, with data services,
messaging and computing functionality on handsets an afterthought in the initial design. Even today,
platforms for m4d apps and services reflect the somewhat ad-hoc nature of non-telephony functions.
4.2 Platforms for Apps and Services
Various sets of features of GSM networks constitute platforms on which m4d apps and services
are built. While each platform exists in relation to the GSM network discussed in the previous section,
the main focus here is on particular features that are relevant to m4d apps and services. In each case,
much more detail is possible – the features of a platform are determined by multiple interactions
between hardware, operating system, network and user. The primary aim of this set of short overviews
is to develop an argument about the various trade-offs involved in selecting a platform. To this end,
Table 2 at the end of the section presents a comparative view.
4.2.1 Voice and Interactive Voice Response (IVR)
Given the primacy of telephony functions in the original GSM specification, voice-based apps
and service seem an obvious choice particularly for low-literacy intended users. Interactive voice
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 70
response (IVR) systems, in which key presses on the phone's numeric keypad transmit small amounts
of data to the remote system, are commonly available for operator services such as airtime loading and
balance checking. IVR systems can also be used to record audio from the caller, or to connect them to
a human operator.
The main disadvantage of voice as a platform for m4d apps and service is generally cost. In
most countries, voice calls are significantly more expensive than data services (or even SMS). India,
which has very low rates for voice calls, is an exception, and much of the interest in m4d voice
services is focused there as a result. Early pilots of m4d voice services, such as the CDC's toll-free
HIV/AIDs hotline in the Democractic Republic of Congo described in Corker (2010), were often
implemented by operators on a pro-bono basis, avoiding the need to negotiate special billing
arrangements. However, scaling and diversifying voice services ultimately requires dealing with the
cost issue.
Third-party services wanting to offer low-cost voice services need an agreement with the
operator to allow any kind of non-standard billing. Both the business case and technical interfaces for
premium-rated incoming lines managed by third parties are well developed. Revenue sharing between
the operator and the service provider ensures a high margin for such services despite relatively low call
volumes. However, reverse-billed lines are seldom easily available, and any line managed by a third
party is generally charged a rate to which various operator promotions (cheaper in-network calls, free
minutes, late-night calls etc.) do not apply. In addition, each operator is generally only able to provide
third-party services to their own subscribers. From the point of view of the service, enabling access for
users on different operators requires multiple agreements and possibly differing technical requirements.
Unless operators agree among themselves, users will generally also have to dial a different number
depending on their operator.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 71
Despite these issues, voice remains attractive to m4d projects, that various strategies have been
developed to mitigate high costs and lack of third-party access. A DIY approach appropriate to projects
with low call volumes might involve the use of a GSM modem, with calls forwarded to and managed
by an ordinary PC. One or several SIM cards are inserted into the GSM modem, and calls to those SIM
cards are answered by the IVR system on the PC. This strategy allows users to call the service as if it
were an ordinary number, including taking advantage of operator promotions if multiple SIM cards
from different operators are used. The disadvantage of multiple SIM cards is that different numbers for
each network (or just multiple numbers) need to be advertised – each SIM can only handle one call at a
time. Operators also sometimes prohibit 'telemetry applications' on their networks, and may disable
SIM cards suspected of such activity.
Voice services can also be designed to avoid the need for an incoming call by having the
service call back a user who has initiated contact by other means. This might include sending an SMS,
entering a USSD code, sending a please call me message, or calling but hanging up before the call is
answered (missed call/ 'beeping'). Each adds technical complexity to initial development, but avoids
dealing with operators and makes it possible to offer a free or low-cost service to end-users.
4.2.2 SIM Application Toolkit
The Subscriber Identity Module (SIM) is a small smart card supplied by the network operator
and inserted into the mobile handset. The SIM was initially envisaged primarily as an identification
and encryption component, able to uniquely identify the subscriber to the network and encrypt
communications over the air. However, by GSM Phase 2+, with growing interest in non-telephony
functions, the GSM standard began to make provision for other uses of the computational and storage
capabilities of the SIM. The SIM Application Toolkit, intended for data and value-added services,
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 72
specified the interface between the SIM and the handset. Initial applications were SMS storage and
phone book storage, as well as mobile banking (Hillebrand, Rosenbrock & Hauser, 2013).
SIM Toolkit applications are able to store and retrieve data, display basic menus and
communicate over the network, including setting up voice calls, sending key-press events, sending and
receiving SMS and using the data network if available. They are usable on even basic phones, and
present a coherent interface and user experience that deviates little from native menus on the phone.
Applications are constrained to an extent by limited storage and computational power on the SIM, but
the more significant constraint is that only the SIM manufacturer (or their mobile operator client) can
load applications to the SIM directly. Operators are also able to load or update applications over the air,
but third parties cannot distribute SIM toolkit applications without assistance from an operator.
The ability to run on any phone is extremely desirable for m4d apps and services. Probably the
most successful developing country mobile service to date, m-Pesa, makes extensive use of SIM
toolkit applications in its home market of Kenya, and as late as 2011 operators were reportedly
considering making Facebook accessible via a SIM toolkit application (Ray, 2011). Small pilot projects
such as Cell-Life's mobile data collection system for adherence to antiretroviral medication (Anand,
2005), have also developed SIM toolkit applications in conjunction with an operator partner. However
the inherent requirement that an operator be involved in any SIM toolkit project has undoubtedly
limited the variety and scale of m4d apps and services on this platform.
4.2.3 The Short Message Service (SMS)
SMS is a popular choice for m4d services because of universal handset support, extensive
coverage over all operators and handsets and (relatively) low cost. It also demonstrates the sometimes
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 73
significant distance between how a technology is conceived and its translation to socially and
technologically mediated forms of use.
Two-way messaging via SMS was a largely unintended success. The original design
anticipated that most use would fit the far narrower 'paging' model of operator to end-user messaging,
analogous to pager devices that were popular in Europe at the time (Hillebrand, 2010). The 'paging'
use-case of SMS was intended for notifications from a network operator to their own users. For
example, an early use of SMS was to notify a user of a waiting voicemail (Hillebrand, 2010). Another
proposal was to add SMS sending capabilities to full-keyboard terminals like the Minitel device in use
in France (Hillebrand, 2010, p. 37), which would allow end-users to send SMSs to mobile devices.
Users were not expected to type and send SMSs themselves, not least because of the difficulty of
typing on the numeric keypads common on phones at the time (Taylor & Vincent, 2005).
SMS was also anticipated mostly as an in-network service. As person-to-person messaging
grew in popularity, inter-network messaging was accommodated through protocols designed for
roaming (Holley, 2010, p.94). Third party access to operators' SMS infrastructure was an afterthought,
however (Harris, 2010, p. 112), with resulting lack of standardization across operators and equipment
manufacturers.
Each of these design factors has implications for m4d apps and services built using SMS. SMS
has been immensely profitable for operators, with a near-zero marginal cost due to its use of an
existing control channel for message transmission (Song, 2009). At the same time, third party access,
which is required by any m4d service not run by a mobile operator directly, has been layered over the
SMS system in non-standard, operator-dependent and sometimes convoluted ways. The sections that
follow cover various technical issues commonly encountered by m4d apps and services using SMS,
tracing their origin and implications and discussing mitigating strategies.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 74
4.2.3.1 Message Transmission
Message transmission for third-party-originated, mobile-terminated SMS (in which a third-
party service sends SMSs to users on one or more networks) is relatively straightforward because it
mirrors the person-to-person case. The network components involved are shown in Figure 6.
In the person-to-person case, the sender's message is delivered to the recipient via a component
called the short message service center (SMSC). Every network has an SMSC, but message delivery
from sender to recipient is the responsibility of the SMSC on the sender's network. The sender's SMSC
obtains routing information for the recipient from home location register (HLR) component on the
recipient's network (Holley, 2010, p.95), and then tries to deliver the message to the mobile switching
center (MSC) on the recipient's network, which passes it on to the base station at the recipient's current
location, which in turn transmits it to the recipient's mobile device. The SMSC handles retrying the
Figure 6: Person-to-person SMS transmission (Acker, 2014, p.565)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 75
message if it cannot be delivered immediately, continuing either until the message is delivered to the
MSC or the validity period expires (Acker, 2014, p.565).
The sender is billed by their own network, which in any case has no contract with the recipient
nor the information required to bill them directly (Harris, 2010, p.112). With the exception of services
requiring very large concurrent messaging volumes (such as emergency alerts), third party services can
send mobile-terminated SMS using the same network infrastructure as person-to-person messaging.
To transmit SMSs to to the relevant SMSC, third party services generally employ a Wireless
Application Service Provider (a WASP, also called a value-added service provider) – a company that
maintains a direct communication link to the SMSC of one or more networks, with which they also
have a billing agreement. Figure 7 shows the relationship between third-party services, WASPs, and
networks.
Figure 7: Third party - WASP - operator relationship. Adapted from Brown,
Shipman and Vetter (2007, p.108)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 76
Commercial WASPs such as Clickatell and BulkSMS allow third parties to send SMSs through
online and desktop software well as by connecting to application programming interfaces (APIs).
Dealing with a WASP gives third parties access to discounted bulk messaging rates, and also makes it
possible for a service to send messages across operators and even across national borders provided all
the operators concerned have an agreement with the WASP.
Third party services that need to receive incoming messages from end-users face a more
complicated situation. As part of the GSM architecture, the sender's SMSC is defined as being able to
handle a special class of SMS addresses called shortcodes or short numbers, which are generally five
or six digits long (for more on shortcodes see Brown, Shipman & Vedder, 2007). Shortcodes can be
configured in the SMSC to forward messages to an external address, such as a WASP's SMS gateway.
Messages sent to shortcodes are charged either at normal SMS rates or at higher premium rates, with
revenue from premium-rated SMSs being shared with the WASP and the third-party service provider.
SMS donations, premium services such as horoscopes, and competition entries are usually handled in
this way (Vodafone UK, 2013).
The major drawback of shortcodes and other network-defined special numbers is that they are
configured in a specific operator's SMSC. Since only the sender's SMSC is involved in message
transmission, incoming SMSs to special numbers must be handled within the sender's own network.
This means that for a special number to work on multiple networks, it must be set up separately with
each operator. A voluntary industry agreement has been established in many countries to overcome this
technical issue, with operators agreeing a shared system of common shortcodes available to third
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 77
parties by application or through WASPs (Brown, Shipman & Vetter, 2007). Common international
shortcodes would require similar agreements.
To mitigate the limitations of shortcodes, mobile operators in a few countries implement a
specialized network component that mimics the message delivery infrastructure of the recipient's
mobile network in the person-to-person case (Harris, 2010, p. 113). To the sending network, the 'virtual
mobile number' provided by this service appears to operate as a standard mobile number on the
receiving operator's network. Operators can then rent the virtual numbers to WASPs and/or third-party
service providers. Even in countries where shortcodes are supported by all operators, virtual numbers
may be preferred for some purposes because they are internationally accessible (albeit at international
rates for subscribers outside the country).
Issues receiving incoming messages are a common consideration for m4d apps and services.
There are multiple African countries without any operator-independent options for mobile-originating
messaging (Pottier, 2011). For services that require multi-country coverage, the problem is
compounded by country-specific WASPs. WASPs need to maintain service agreements, payments and
technical interoperability separately for each operator, and most have limited geographical coverage as
a result. m4d apps and services requiring multi-country coverage will generally need to deal with
several WASPs, applying and paying separately for shortcodes or virtual numbers in each country
where they are available.
Operators and their partner WASPs also control the allocation and regulation of special
numbers to third-party services. Requirements may include minimum usage levels, cost-to-user
restrictions, and restrictions on content deemed objectionable by the service provider. Unlike web-
based services, where data transmission is content-neutral, content restrictions from mobile operators
can be quite onerous (Brown, Shipman & Vedder 2007). In the US, for example, Verizon's terms and
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 78
conditions for SMS services at one time prohibited not only adult content and profanity but
information about organ transplants, adoption services and copyright protection circumvention, and
speech that “trivializes historical events” (Lasar 2008).
SMS services targeting low-income users in developing countries - particularly in the pilot
phase - are sensitive to the financial and time cost of the shortcode or virtual number application
process, as well as geographical coverage limitations and possible usage minimums. Two alternative
strategies for receiving incoming messages are popular in the m4d community. Both involve procuring
a SIM card for one or more mobile networks in the country where the service will operate, which is
used to send and receive messages directly rather than dealing with a WASP.
In one strategy, popularized by desktop applications such as FrontlineSMS, the SIM card is
inserted into a phone or a specialized hardware device called a GSM modem, which is connected via a
data cable to a computer that processes messages. A more recent variant involves receiving and
processing messages on an Android smartphone. The advantage of both approaches is that they require
no operator or WASP involvement, and can be set up quickly and cheaply. However, both work only
for low message volumes, as SMS throughput is low on phones and only slightly higher on GSM
modems. Some operators also technically prohibit (but seldom shut down) automated devices that send
and receive messages on their network.
Another strategy, used when no local WASP is available in a particular country but high
throughput, reliability and a local number are required, involves procuring a local SIM card that is
enabled for international roaming and then housed in an international WASP's data centre. End-users
sending messages to this number pay local rates, but the service operator pays for hosting in the data
centre as well as additional roaming charges.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 79
4.2.3.2 Billing
Issues related to SMS billing – specifically, managing the cost of messages to end-users – also
particularly affect m4d services. SMS billing options vary between mobile networks. Mobile-
originated billing (the end-user pays to send an SMS) and mobile-terminated billing (the end-user pays
to receive an SMS) are commonly available and widely used commercially. Numbers that use so-called
reverse billing, or other options for free-to-user SMS with the cost assumed by the receiving third-
party service provider, are much harder to find. Where they do exist, they must be negotiated and
configured per operator. There are also unresolved concerns around abuse of reverse-billed numbers.
WASPs in South Africa are still hesitant to offer reverse billing despite being its introduced by
Vodacom in 2012, as the WASP through which the reverse-billed SMS number is provisioned also has
to accept responsibility for all SMSs sent to the number (Muller, 2012).
Difficulties with reverse billing and other free-to-user options are part of a general history of
ad-hoc arrangements for SMS billing. Once mobile operators realised that person-to-person messaging
was becoming popular and profitable, it became necessary to establish technical interoperability and
billing procedures. Technical interoperability was achieved using an existing messaging infrastructure
designed to handle roaming users (Holley, 2010, p.94). Billing procedures were based on the
realization that only the SMS sender's network had sufficient information to bill for the message since
the sender's SMSC delivers to the recipient directly (Harris, 2010, p. 112).
Agreements among operators to deliver messages from other networks might take the form of a
monthly reckoning (Le Bodic, 2005, p.58), or (as was until recently the case in South Africa), an
agreement not to charge for delivering messages from other networks since person-to-person traffic is
roughly symmetrical (McLeod, 2013). Application-to-person traffic is not symmetrical, however, and
operators may either opt to only deliver application-to-person messages to their own subscribers, or
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 80
they may attempt to identify and charge a fee to other operators for application-to-person traffic
terminating on their network. The first scenario is most common, and again leads us to the likelihood
that users on different networks will require separate negotiations and technical accommodations for
third-party access, particularly where a special request, such as reverse billing, needs to be entertained.
The low commercial priority of free-to-user SMS makes sense because outside of m4d, there
are few business models where the cost of sending a text is a barrier to use for the intended market.
WASPs concentrate on standard and premium-rated SMS, attracting third party services whose users
are either indifferent to the cost of an SMS, or willing to pay premium rates for a service (or as a
donation). M4d SMS services, on the other hand, often need to solicit information from prepaid users
with little or no airtime and significant price sensitivity. They are forced to negotiate free-to-user SMS
numbers with operators individually, and frequently fail to obtain them.
One alternative, generally found in SMS data collection projects, is to arrange to automatically
reimburse senders for the airtime they use to submit data. This depends on third party services being
able to access operator's online credit purchase mechanism in an automated way. Another strategy is to
provide a pre-defined amount of airtime to service users earmarked to send messages to the service
(ict4chw, 2010). Neither strategy solves the problem of providing access to the service for users with
no airtime credit. Additionally, neither is really scalable beyond a pre-defined group of service users
who have an ongoing, employment-based relationship with the service.
For larger groups of unknown or once-off users, free-to-user channels such as Unstructured
Supplementary Service Data (USSD), described in 4.2.4, or 'please-call-me' messages, can replaces
the mobile-originating SMS as a means for the user to initiate contact. Services with this configuration
reply by SMS, but simplify the billing problem by sending rather than receiving SMSs. The Mobile
Alliance for Maternal Action (MAMA), a service that provides health pregnancy tips via SMS to
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 81
pregnant women who sign up for the service via a USSD code, is one example of this model (Praekelt
Foundation, 2013). At least in South Africa, where the MAMA SMS project operates, simple USSD
codes and 'please-call-me' messages are relatively well-used even by mobile users who do not send
SMSs (Velge, 2014), making them a cheap and accessible way for users to initiate contact without
incurring SMS costs.
4.2.3.3 Message content
Text encoding for short messages was based on an alphabet designed to provide a regional
paging system for Europe. The encoding makes provision for a basic 7-bit Latin alphabet and the most
common extension characters in major European languages, with the latter replacing control characters
in the base character set (Trosby, 2010). This choice, combined with the maximum message size
possible in the selected SMS transport protocol (MAP), gave a message length of 160 characters.
Unicode was by that time the standard approach in information technology, but Holley (2010, p. 94)
notes there existed “very little exchange between telecommunications and information technology
experts”, which partly explains the unusual choice for the original encoding.
With several countries in Asia and the Middle East became significant GSM users, provision
for Unicode support was added to GSM in 1996 (Hillebrand, Rosenbrock & Hauser, 2013). Non-Latin
alphabets were accommodated using a 16-bit Unicode character set. Since each character now required
double the number of bits to encode, messages in non-Latin alphabets are restricted to 70 characters.
This problem is partially ameliorated by SMS concatenation standards, which make it possible to send
a message over several SMSs (Hillebrand, Rosenbrock & Hauser, 2013).
Mobile device support for regional languages varies by device and by region. While most
phones sold in a particular region should support major regional languages,, 'grey market' phones and
older phones may not. In India, phonetic equivalents in the Latin SMS alphabet are so common that
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 82
Vodafone India offers a 'vernacular SMS' service. Users send the phonetic equivalent of the message
and the language they require to a shortcode, and receive the Hindi, Gujarati or Marati equivalent as a
picture message to forward to their chosen recipient (Vodafone India, 2013). SMS services operating in
countries where a regional language uses non-Latin alphabets still need to decide whether to choose a
non-Latin alphabet and risk some recipients not being able to read the message, or to attempt a
phonetic equivalent. Dodson, Sterling & Bennet (2013) document multiple issues with language
support, grey-market phones and basic difficulty with SMS symbols and abbreviations among low-
income users in Morocco. They conclude that in communities outside the narrow linguistic paradigm
of the mobile network infrastructure, users “are acutely aware that simple phones are not simple to
use.”
4.2.3.4 Security
Third-party services using SMS must consider that the technical specification of SMS makes
no provision for message encryption once the message reaches the sender's SMSC. The message passes
through the sender's network and the recipient's in plain text, and can be intercepted or filtered in real
time by an adversary with access (authorized or unauthorized) to mobile operators' systems. Most
commonly access is established within operators' networks, but for many years communication
between operators over the Signalling System 7 (SS7) network was unsecured except for the
assumption that only legitimate operators would be able to obtain the required hardware and software
to access the network. Once Internet connectivity became a cheaper and more reliable option than
specialized networks, SS7 over IP (Internet Protocol, essentially SS7 over the Internet) became the
preferred mode of telecommunications interconnectivity, opening the vulnerable SS7 network to
exploitation in principle by anyone with an Internet connection (National Communications System,
2003).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 83
SMS filtering under repressive regimes has been reported in several countries, and confirmed
in Syria in 2012 (Elgin & Silver, 2012). The filtering tools involved in the Syrian case are technically
identical to those commonly employed by operators to reject unauthorized messages and filter SMS
spam coming into the SMSC, blurring the line between legitimate and contested use.
As well as intercepting and filtering, SMS services are vulnerable to impersonation. The GSM
standard makes provision for senders to specify numeric and alphanumeric sender IDs as part of the
message content. Third-party services routinely do this to identify themselves and provide a reply
number. Third-party services are generally required to agree to a particular sender ID, but the message
as displayed to the recipient (or the receiving network) provides no way to positively identify the
sender.
For many m4d SMS services, security issues are a lower priority than access, and SMS is still
an appropriate choice. However, the plain-text transmission of message content and the difficulty of
verifying senders' identity may run afoul of legal requirements for certain operations, such as
transmission of medical records or financial data. Given the possibility of government interception and
filtering, services that solicit or deliver politically sensitive information may also need to avoid SMS.
4.2.4 Unstructured Supplementary Service Data (USSD)
Unstructured Supplementary Service Data (USSD) is a text-based application-to-person
communication channel designed to facilitate in-network supplementary services such as callback and
caller ID. USSD applications were initially envisaged to be housed within various GSM network
components, as shown in Figure 8, which is reproduced from the GSM standards document defining
USSD. Today USSD is sometimes used for services run by third parties. This may be restricted to third
parties providing services to the network (such as prepaid billing), or it may be available as a standard
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 84
value-added services platform. Operators who offer the latter may also offer premium-rate and reverse-
billed USSD.
Like SMS, USSD is an operator-dependent service. Unlike SMS, it was designed as an
application-to-person service, with no provision for interaction from users on other networks. Users on
different networks will generally need to use a different initial code to reach as USSD service, and any
special arrangements required must be made separately for each operator.
Also unlike SMS, which delivers messages asynchronously on a best-effort basis, USSD is
session-based. This means that once a USSD process begins, a live connection is maintained for the
duration of the interaction between the users' handset and the USSD service, or until the operator's
maximum session duration is reached. Thus, while the maximum message length for USSD is similar
to SMS at 182 characters (80 for messages in non-latin character sets), it is possible to send multiple
messages in a single session for a single rate. The user can also interact in real time with the service.
User interaction is limited to selecting from a menu of options, but USSD services can also launch
interactions on other platforms, such as an automated call back in response to USSD interaction.
Until recently, USSD apps and services were mostly provided by network operators and
partners such as mobile banking providers, with each operator retaining monopoly control over USSD
Figure 8: USSD application architecture (ETSI, 1996)
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 85
services to their subscribers. In India, operators were only convinced to enter partnerships to offer
reasonably priced USSD mobile banking services after government intervention (Telecom Regulatory
Authority of India, 2013). Operators in some countries, including South Africa, now offer USSD as a
platform for third-party value-added services. Because it provides a limited two-way interaction on
basic phones, USSD is particularly attractive for services that translate web-based applications to basic
phones. Twitter, Facebook and Wikipedia have all at one time offered USSD services in partnership
with one or several African operators.
4.2.5 Mobile Internet
Unlike SMS and USSD, mobile data services were always intended as third-party channels,
with early examples such as I-mode and the Wireless Application Protocol (WAP) offering
predominantly third-party content such as news headlines, weather and stock prices. The success of I-
mode and relative failure of WAP illustrates some of the factors at play in mobile Internet adoption.
Although the experience of (often) “mobile-centric” (Donner, Gitau & Marsden, 2011) Internet users in
the developing world is notably different to early adopters in Europe and Japan (Donner et al 2011,
Walton & Donner, 2013), the operator/user/content provider relationship remains relevant for m4d
services.
I-mode, developed by Japanese carrier NTT DoCoMo in the late 90s, was initially intended as
a business service (hence early advertising mentioned stock prices and travel bookings), but thanks to
its instant messaging function was quickly adopted by trend-conscious Japanese youth (Agar, 2013).
Data was charged by quantity used rather than time online, making the instant messaging function
economical. The I-mode offering was strongly vertically integrated despite being a third party channel.
NTT DoCoMo maintained close control of device specifications as well as scrutinizing content from
third-party providers (Palomäki, 2004).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 86
The standardization process leading to WAP was initially driven by equipment manufacturers
hoping to limit customization requests from operators. GSM phase 2+ included support for data
transmission in the form of GPRS and EDGE, and the Internet was rapidly becoming ubiquitous in
Western business. WAP was intended to offer an Internet-like experience over the slower and less
reliable connections available on the mobile network, designed with screen size limitations in mind. By
mid-2001 the WAP forum had over 500 members including all major telecommunications companies
and many software companies, and WAP was supported on most new handsets (Sigurdson, 2001). Like
I-mode, WAP was intended as a business service, but rather than charging only for data use, operators
chose a pricing model based on the duration of the connection.
Poor usability, failure to attract significant interest from content providers and the high cost of
a time-based pricing model meant that despite impressive handset and operator support, WAP suffered
from low initial uptake across the board. Palomäki singles out operators in particular for failing to
provide adequate revenue sharing for third-party content. Limited content within operator-controlled
portals was a poor substitute for the vast Internet, and as faster data transmission protocols, more
powerful handsets and better browsing software appeared, WAP languished.
The next big story in mobile Internet was the development of the Opera Mini mobile browser.
Opera Mini, initially released as a Java app for feature phones, routes web requests through a network
of Opera servers that compress web content, then sends a compressed and small-screen-optimized
version to the phone. Operator partnerships, which allowed operators to set a default landing page and
bookmarks menu as well as other branding, were a key factor in ensuring Opera Mini was (and is)
installed by default on many featurephones. In 2013, Opera Mini had 241 million unique users, and the
top ten countries by total unique users were India, Indonesia, Russia, Nigeria, China, Brazil, South
Africa, Bangladesh, Mexico and Vietnam (Opera Software, 2013).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 87
For users, Opera Mini's compression resulted in a markedly better featurephone browsing
experience despite slow connections. By reducing the amount of data transmitted, it also decreased the
cost of mobile browsing. For developers, Opera Mini and its successors shifted the burden of
accommodating mobile users from sites themselves first to Opera's servers, then to increasingly
powerful handsets. Mobile versions of websites ('mobi sites') are still important from a usability
perspective, but mobile users are increasingly operating on the web itself rather than on a limited
network of mobile-targeted sites.
In South Africa, the Mxit instant messaging app took a different approach to web content
browsing on feature phones. Mxit, after a failed early incarnation as a mobile-based multiplayer game,
gained rapid popularity among South African youth as a chat service. It offered a cheaper alternative to
SMS, and was available for a wide range of featurephone models. Several m4d services have made use
of the Mxit platform through the chat paradigm, offering real-time tutoring or advice. For others (and
for Mxit today), the Mxit platform is a way to provide targeted, discoverable web content at low cost,
through a familiar interface.
Unlike Opera Mini, which presents itself to the user as a standard web browser, Mxit
approaches web content provision through the concept of portals. Mxit 'mobi portals' (or 'apps' as they
are now described in the revised user interface) must provide web content in a particular, stripped-
down HTML format, and apply for approval from Mxit. In return, they are listed in the Mxit apps
menu and discoverable by Mxit's user base, who 'add' apps that interest them to access their content.
Mxit apps are able to access some additional chat-like functionality, such as sending notifications to
users who have added the app. In fact, Mxit, with its centralized content discovery model, notifications
and subscription functionality, is similar in many ways to the content portals envisaged by WAP and I-
mode.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 88
One of the big factors driving content consumption within Mxit's walled garden is low data
costs (Research ICT Africa 2012, p.30). Mxit apps are low-bandwidth by design, and South Africa's
third-largest operator, Cell-C, for some time zero-rated all Mxit traffic in an effort to attract Mxit users.
Mxit itself generates revenue from advertising and premium content sales (split about 50/50 in 2012),
including purchases using its in-app currency, Moola (Research ICT Africa, 2012). Although Moola
was launched with m-banking in mind, it is mostly used to buy entertainment content, with users
converting airtime to Moola by sending a premium-rated SMS.
For third parties in general, Mxit apps are attractive because they are relatively easy to develop
and update - comparable to a website rather than a native app - and have few external dependencies
beyond ongoing web hosting and initial approval by Mxit. The demographics of the user base, which
Mxit characterizes as “emerging market youth” who spend an average of 105 minutes daily using the
Mxit (Mxit, n.d), are seen as particularly well suited to m4d apps and services targeting young people.
Demographic information released in early 2014 (Mxit, 2014) suggests users are more
emerging middle class than poor, with 55% in LSM 6-8 compared to 9% in LSM 1-5 (the LSM, or
Living Standards Measure, is a commonly used consumption-based market segmentation tool that
ranks households as LSM 1-10 from least to most consumption). This is consistent with Research ICT
Africa's findings from 2012, in which while 47.7% of base of the pyramid (BoP, defined in the study as
households with a monthly income below a poverty line of ZAR 432 / USD 40 per month) social
networking users used Mxit, only 12.8% of BoP mobile users used social networking at all.
Nevertheless, Mxit has been actively supportive of m4d applications, which it develops directly with
partners through Mxit Reach.
Unlike in the WAP era, the operator is no longer central to users' experience of the mobile web.
Operators still hold some cards, however, particularly for services targeting low-income or first-time
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 89
web users. The ability to zero-rate particular sites, as Cell-C did with Mxit, allows operators to provide
free access at their discretion. Wikipedia Zero and Nokia's MoMaths service are examples of services
that have successfully negotiated zero rating in South Africa, albeit not with all operators. Operator
portals, such as Vodacom's Vodafone Live, are generally zero-rated for subscribers. In addition,
operator portals accessible from a browser splash screen or a special menu item have great advantages
in terms of discoverability. Although the mobile web offers m4d apps and services an operator-
agnostic platform, operator partnerships still have significant advantages. Similarly, social networking
sites such as Mxit have important discoverability and interactivity functions that are both harder and
more costly to build on a standalone basis.
4.2.6 The J2ME App Environment
J2ME apps are featurephone apps built to run in the J2ME (Java 2 Micro Edition) application
environment. J2ME was specified through the Java Community Process, an open standardization
process that nevertheless involved Sun Microsystems, who owned the Java programming environment,
retaining significant control. In co-operation with device manufacturer Motorola, Sun introduced the
original J2ME specification in 1999, aiming to curb fragmentation in the mobile device market
(Mahmoud 2004). Mobile and embedded software was strategically important to Sun because it
epitomized the network-oriented, 'write once run anywhere' characteristics that were the main selling
points of Java (Garud, Jain & Kumaraswamy, 2002).
Mobile devices would soon reveal 'write once, run anywhere' to be severely compromised.
Because of the differing capabilities of the kinds of devices for which J2ME was intended – anything
from a set-top box to a PDA to a mobile phone - the core requirements for a device to implement J2ME
are extremely limited. Providing access to device capabilities such as the camera, location-based
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 90
services or Internet connectivity means implementing additional specifications, so the total package of
specifications available on each device can differ significantly.
Security certification requirements for J2ME apps further complicate the situation, with
uncertified third-party apps having access to limited functionality or being required to prompt the user
for permission to access functionality such as the device camera or network connectivity. The result is
that J2ME apps require significant modifications for different devices, including devices from different
manufacturers, different models from the same manufacturer, and the same model with different
operator-specific software.
Even where developers are able to produce multiple versions of the same app to address the
fragmentation problem, distribution remains an issue. J2ME apps can be distributed in one of three
ways:
1. Direct transfer of the app to the device vis bluetooth or with a data cable.
2. Web download, in which the user visits a url in a web browser to download the app.
3. Installation on the device prior to sale.
In the first case, the user needs to first obtain the app files, either via bluetooth transfer from someone
who already has them, or by downloading them to a PC. Either physical proximity to an existing user
or access to a PC is required. Web download on the phone's web browser is more scalable, although it
requires that the user be familiar with web browsing and have a working data connection. Even so, the
third method, in which operators and manufacturers provide some apps pre-installed, is probably the
most common way users acquire J2ME apps. It is also the only method in which the user is relieved of
the task of obtaining the correct app version for their device, which otherwise relies on matching an
extensive and rapidly changing matrix of device models to each available version of the app.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 91
Although today we think of apps as largely operator- and hardware-independent, the J2ME app
environment was decisively shaped by the interests of both handset manufacturers and operators. The
complicated security certification scheme, which in addition to restricting access to phone features
allowed operators and manufacturers to dictate their own set of acceptable certification providers
(varying sometimes by device, sometimes by operator), was meant to protect operator networks and
handset operating systems from rogue or simply undesirable apps. The high cost of certification in fact
prevented many third-party apps from ever applying, excluding them from operator and manufacturer
distribution channels. Device fragmentation was driven by the need to constantly bring new devices to
market, with Nokia targeting 40-50 new devices a year in the early 2000s (Heikkinen, 2010).
Unsurprisingly, the multi-party open standards process was rapidly overtaken. Device capabilities
outpaced the standards required to access them (Qusay, 2004), and J2ME app developers were
frustrated by the limited functionality available to their apps (Tarnacha & Maitland, 2006). The sheer
number of devices meant multiple versions of each app needed to be developed and distributed,
increasing development costs and confusing users.
Despite these shortcomings, J2ME was the only pre-smartphone app environment to achieve
widespread availability in the developing world. This is changing, with cheap Android phones
increasingly available and Nokia's ubiquitous Series 40 operating system, which provided some of the
most full-featured and reliable J2ME phones, discontinued in 2014. For m4d apps and services in the
pre-smartphone era, however, J2ME was the only game in town.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 92
4.2.7 Comparison of platforms for m4d apps and services
Designed for Affords Requires Main m4d users m4d adaptations
SIM
toolkit
Data storage on SIM,
value-added services
such as banking
provided by operators
or operator partners
Native menus, simple user
experience. Works on any
phone.
Installation on SIM
requires operator co-
operation.
Large-scale pilots by
operator partners
such as Facebook
Zero.
Voice/IVR Person-to-person
business calls,
including while
roaming/across
national borders.
Primary
communication
channel on GSM
networks.
Access for low-literacy
users. Familiarity. If using
callbacks, provider-pays is
possible. Works on any
phone
Provision of free or paid
incoming lines is
operator-specific.
Most common in
India where voice
calls are cheap. Also
used for projects with
radio aspect.
GSM Modem to
receive calls, or
callback service
relying on another
channel to initiate an
outbound call.
SMS Machine-to-person
messaging, for e.g.
from a terminal
device. Paging model.
Textual data transmission.
Relatively low-cost, many
service providers (for
outgoing messages). Works
on any phone.
Provision of free or paid
incoming numbers is
operator-specific, but
often mediated by a
service provider. For non-
latin alphabets, recipients'
devices must support the
chosen character set.
Outdoing SMS
widely used by third
parties for awareness
campaigns, periodic
notifications/reminde
rs. Incoming SMS is
less common, used
for data collection,
subscription/informat
ion request.
GSM modem to
receive and send SMS,
use of hosted SIM
services for multi-
country coverage.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 93
USSD Access to network-
provided automated
services e.g. balance
enquiry
Textual data transmission,
user feedback, multiple
interactions in a single
session. Low-cost. Works
on any phone.
Provision of free or paid
USSD codes is operator-
specific, but often
mediated by a service
provider. For non-latin
alphabets, recipients'
devices must support the
chosen character set.
Sessions must be
concluded before operator
time-out, usually a few
minutes.
Information services
for basic phones,
subscription/informat
ion request.
Mobile
Web
In the WAP era,
content distribution by
operators and
approved partners.
Browsers such as
Opera brought the web
(accessed via operator
portals) to feature
phones.
Ability to present
extensive,
organized/hyperlinked
textual information. Some
user interaction e.g. asking
questions. Works on many
feature phones.
Provision of free access
(zero-rating) is operator-
specific. Users need data
connectivity enabled.
Information services
for feature phones.
Access via operator
portals to provide
discoverability, zero-
rating. Mobile web
content also often
accessible via Mxit.
Mxit Chat app – person-to-
person and public chat
'rooms', similar to the
Internet Relay Chat
(IRC) protocol.
Private text chat interaction
with live operator or
automated information
service. Low data use / low
cost. Extensive youth user
base. Works on many
feature phones.
Mxit approval required
for apps. Users must
download and install mxit
app, and have data
connectivity enabled.
Informations services
for feature phones,
interactive
counselling.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 94
J2ME Third-party apps Access to phone features,
connectivity memory and
processing power.
Available on many feature
phones. Apps are free to
use, data transmission is
inexpensive.
Installation requires
download through
browser or direct transfer
to phone. Access to phone
features varies by
manufacturer, model,
sometimes operator.
Data collection,
educational/awarenes
s-raising games.
Apps often specify a
small number of tested
handset models. Users
may request a
download link for the
app via SMS.
Table 2: Comparison of platforms for m4d apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 95
4.3 Discussion: Mobile network platforms for m4d apps and services
The GSM standard envisaged, in the first instance, a Europe-wide telephony system.
Expansion to the developing world benefited multinational telecommunications companies and
equipment manufacturers, who put pressure on the standards process to accommodate the expanded
functionality they needed. Millions of people in the developing world now benefit from access to
mobile communications. Yet the uneasy position of third-party services in the non-smartphone
ecosystem demonstrates an enduring logic of operator control, working through standards and
technology artefacts as well as the operating arrangements of the mobile industry.
The GSM standards process and subsequent technology development sanctioned mobile
operators' gatekeeper role in relation to third-party services. Furthermore, in the European context of
the standards process, third-party services were of secondary importance compared to telephony
functions. Particular emphasis was placed on telephony functions directly related to mobility, such as
roaming. In the developing world, mobile users may not have access to alternative communication
networks. Basic communication, rather than mobility, is the critical affordance of mobile phones from
the perspective of m4d apps and services. In contrast to the political will and resulting technological
and commercial success of roaming, mobile phones as basic communication devices through which
people can access third party services are variously limited.
Standards for third-party access have reinforced operator control of the platform rather than
opening it up to a third-party ecosystem. Intelligent Networks (IN), was created specifically to allow
mobile networks to differentiate themselves based on proprietary service offerings, while WAP failed
in part because operators were not willing to offer an attractive revenue sharing agreement to third
parties.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 96
Looking ahead to the interview findings in Chapter 6, operator gatekeeping around access to
network services was repeatedly identified as a constraint on the field. This was less pronounced where
the requirements of m4d apps and services aligned with commercial products such as bulk outgoing
SMS. It was more pronounced where the requirements of m4d apps and services were specific to the
needs of low-income users, as in the case of zero-rated access or complex J2ME apps.
The inventory and interview findings also demonstrate how designers of m4d apps and
services mitigate the limitations of the platform. Using please-call-me messages to initiate an
interaction, avoiding a user-pays SMS, is one example. A plugin that publishes content as a mobile web
site and a Mxit app simultaneously to improve discoverability is another. That specialist m4d
technology companies are responsible for developing the largest proportion of m4d apps and services
reflects the need to deploy a distinct set of technical and procedural adaptations to platform limitations.
Particularly in m-health, the density and longevity of specialist m4d technology companies has lead to
collective recognition of this shared set of mitigating strategies, although attempts to co-ordinate
through sharing code may be less viable than software-as-service sharing models.
Regardless of whether they are propagated through knowledge sharing or encoded in shared
code, shared technical and procedural mitigating strategies for the limitations of platforms for m4d
apps and services are part of the way the mobile network infrastructure is encountered by designers
and developers. Infrastructure is “sunk into and inside of other structures, social arrangements and
technologies” (Star 1999, p. 381). In the case of platforms for m4d apps and services, “emergent”
properties (Orlikowski 2000) arising out of the social arrangements that constitute m4d include
countermeasures against the limitations introduced by the “embodied” properties of the platform.
Many of these limitations can be traced to the gap between the design context for mobile networks and
the implementation context of m4d apps and services targeting low-income users in the developing
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 97
world. Where Heeks (2006) defines a framework of “design-reality gaps” to analyse ICT4D project
failure, this research demonstrates that design-reality gaps are also relevant at the level of
infrastructures and platforms on which ICT4D interventions are deployed.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 98
CHAPTER 5: Apps and services inventory findings
A total of 71 m4d apps and services were identified, described and classified as part of the
inventory research component. The full list, including classification by sector, archetype and
technology platform, is included as Appendix C. This chapter presents the findings in terms of the
initial dimensions of interest as well as the emergent classification into archetypes described in Section
3.5. The discussion here is supplemented by interview findings presented in the next chapter.
Returning to the conceptual framework, the inventory task address the central concern of
platform studies that “the hardware and software of platforms influence, facilitates or constrains
particular forms of computational expression” (Montfort & Bogost, 2009, p.3). Having covered
historical and technical attributes of platforms for m4d apps and services in the previous chapter, this
chapter considers how they play out in m4d apps and services as “forms of computational expression”.
Infrastructure studies offers ways to connect the material realization of mobile technology platforms,
and of the apps and services built on them, to the wider sociotechnical context. Each app or service is a
boundary object, embodying the motivations, priorities and world view of funders, developers, users,
m4d communities of practice and often the public as a mediated audience. The material realization of
m4d in South Africa is inseparable from infrastructure, institutions and social and political
arrangements. The intention of the inventory process is discover what these are and suggest how they
operate. This includes the ways designers and users of particular m4d apps and services have
encountered or circumvented the limitations of the platform.
Each step of the inventory process is clearly imperfect: identification depends on publicly
available information about the app or service, description on the particulars contained therein, and
classification on a pre-defined set of attributes and categories representing areas of interest for this
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 99
particular study. Nevertheless, patterns emerging from an imperfect dataset can suggest relationships
and biases that gain weight when triangulated with findings from the other research components.
Although the inventory data is not intended for statistical inference, it is helpful to present the
dimensions roughly in terms of dependent and independent variables. For the latter, technology
platform and archetype are the primary dimensions of interest. Breakdowns by these two dimensions
are therefore presented first. The remaining dimensions, often cross-tabulated by technology and
archetype, are organized into user experience dimensions (languages, cost), product development
dimensions (developers, funders, mobile operator involvement), and development intervention
dimensions (sector, intended users, development theory of change, role of technology, measures of
success).
5.1 Technology platforms
Figure 9 shows the inventory breakdown by technology platform.
Figure 9: Technology platform frequency in m4d apps and services inventory
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 100
Of the 71 apps and services in the inventory, 42 (59%) are accessible through Mxit and 26
(37%) through the web. There are 18 SMS-accessible services (25 %) and 6 USSD-accessible services
(8%). J2ME and smartphone apps are less common, with 10 J2ME apps (14%) and 8 smartphone apps
(11%) counted in the inventory process.
It should be re-iterated at this point that one of the biases in the inventory relates to
discoverability, which is largely determined by the platform. In particular, Mxit, which has a top-level
menu item for m4d apps and services, provides excellent discoverability. SMS and USSD provide no
discoverability through the platform at all, and apart from a few operator attempts at app stores, none
of which appear to be widely used, neither does J2ME.
Just over a third of the apps and services are accessible on more than one technology platform
– 27 of the 71 total apps and services. Mxit and a mobile web site are the most common combination,
with 17 Mxit apps also accessible as mobile web sites. From the developer perspective, Mxit apps and
the mobile web are similar platforms, both often built using common web development languages and
tools (php, html) and accessible over HTTP. Making the same content available on both Mxit and the
mobile web is therefore a relatively straightforward operation, and allows users with a web browser to
access the content without installing the Mxit app.
For users without access to a smartphone or featurephone, 17 inventory services (25%) are
accessible by SMS, USSD or both. The remainder are accessible to featurephone users via Mxit, a
mobile web site or a J2ME app. Only one smartphone-only app – the data collection system used by
ikapadata - was included in the inventory. Users of this app are generally employed by ikapadata and
provided with phones. The technology choices evident in the inventory demonstrate that
featurephones, not smartphones, are the target hardware for m4d apps and services in South Africa.
Despite near-ubiquitous claims of wide accessibility in m4d projects, attempts to make apps and
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 101
services accessible to basic phones that support only SMS and USSD face both content and technology
limitations.
5.2 Archetypes
During the inventory process it became evident that, within each sector, multiple apps and
services provided essentially the same functionality. This provided motivation for grouping inventory
apps and services into archetypes. Archetypes were defined through an iterative process, with the final
list of fourteen demonstrating some cross-sector classification but mostly a small number of shared
“forms of computational expression” within each sector. The fourteen archetypes to which inventory
apps and services were assigned were as follows:
• Surveys/market research, including apps and services that pay users for answering survey
questions or giving opinions on products. Airtime is commonly offered as payment, although
virtual currency (specifically Mxit moola), shopping vouchers and bank account transfers are
also used. In addition to self-reported surveys, several apps provide opportunities for peer
interview. InstantAfrica asks users to survey their friends, while ikapadata's original concept
paid township residents to conduct weekly panel surveys. Mobenzi Intelligence was trialled as
a microwork platform along the lines of Amazon's Mechanical Turk, with users completing
small tasks on their phones.
• Reading/literacy materials, including various reading materials on mobile platforms, whose
claim as a development intervention is literacy promotion. Most focus on teenage readers (Yoza
Cellphone Stories, Bookly, Everegg, FunDza), although Nal'ibali includes material intended for
reading aloud to younger children. FunDza also runs competitions aimed at getting young
people writing content for publication their platform.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 102
• Pill containers (a combination of an SMS service and specialized hardware) originated to
address the challenge of chronic medication compliance in the face of the HIV and TB
epidemics. Of the two pill containers tested or marketed in South Africa (SIMpill, WisePill),
neither appears to have been scaled beyond a pilot. A pilot study of WisePill (Haberer et al
2010) indicates reliability issues with consumer SIM cards and billing, as well as noting
relatively high hardware and SMS costs. A third system, SIMmed, does not require specialized
hardware but expects users to send an SMS or dial a USSD code every time they take their
medication. It too seems stalled in the pilot phase.
• Mass reach apps and services take advantage of high mobile penetration to build or reach a
mass audience for development communication. Project Masiluleke, which used advertising
space in please-call-me messages to encourage recipients to call the National AIDS Helpline, is
an early example. The use of please call me messages was made possible by mobile operator
co-operation (the operator donated the advertising space). D-Siders is a mobile web site that
publishes serialized text-based stories and encourages user engagement. The business model is
similar to Facebook advertising in the creation of a community to which advertising messages
can be displayed, but with a focus on development communication and reaching developing
world youth.
• Job search apps and services provide job listings, connecting job-seekers to employers. Uusi,
an apparently dormant job search service on Mxit, and Ummeli, a mobile web site accessible
through Vodacom's Vlive portal, are mobile-only options. Student Village began as a web site
but has added a stripped-down mobile web version and Mxit portal.
• Info services is a large category most prevalent in the health sector, with HIV/AIDS and
pregnancy – related content predominating. In addition to static content, many services operate
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 103
a question and answer section. Some, such as Hi4Life and YoungAfricaLive, mix health
information and entertainment content in an effort to engage a teen and young adult audience.
Mobile phones are seen as a way to reach a critical audience and provide access to sensitive
information through an anonymous, relatively private channel.
• General education & games includes several quiz and activity apps with content outside of
the school curriculum. Bsmrt is a quiz app with general knowledge content, while Groovy
Adventurers is targeted at parents and provides guidance and tracking for early childhood
development activities. Oxford Word of the Day and the Periodic Table app on Mxit provide
reference content. Moraba: It's your move is a more complex game incorporating informational
content about gender-based violence. 'Serious games' of this type are an established genre,
although less common on mobile platforms, and indeed Moraba seems to be most functional as
a smartphone app.
• Further education & training apps and services provide information and listings for youth –
mostly school leavers – about careers and education and training opportunities such as courses,
degrees and on-the-job training. Brainwave Careers includes a weekday afternoon live chat,
while Harambee, a government-supported project, provides job training in person with the
mobile site used for applications.
• Data collection / remote worker supervision is an established model dating back to the PDA
era. Mobile devices with an app installed are provided to remote workers, such as community
health workers or survey data collectors. Users complete interactive forms loaded in the app.
Additional features such as case management, location capture, instructional videos may be
available to users, and supervisors access various reports through a web interface.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 104
• Curricula & practice apps and services provide textbook access, multimedia content and
practice exercises based on the school curriculum. Most are targeted at students in the final
three years of high school, with maths and science curriculum content predominating.
MindsetXtra Live, a project of Mindset Education, which produces televised educational
programming, also provides a Q&A feature for curriculum content. Teachers' Corner on Mxit,
an app for teachers with advice and timed activities for the school year, has similar
functionality to the other apps in this category but targets teachers rather than students directly.
• Crowdsourcing apps and services aggregate data collected from citizens. Lungisa allows
residents to report infrastructure issues such as potholes and burst pipes, while RateMyClinic
invites patients at primary care facilities to comment on their experience.
• Counselling services mostly take advantage of discoverable private chat functionality on Mxit.
RedChatZone provides counselling about HIV, Angel counsellors are available to talk about
substance abuse and gangsterism, LoveLifeMX covers sexual health and Childline and MobieG
offer general counselling to young people. Counselling services are generally accessible during
a pre-defined time window.
• Constant contact is another large category of apps and services that send regular, mostly short
messages. Appointment reminders, medication adherence reminders, and followup messaging
after an HIV test or other medical procedure are included in this category. In the health sector,
they are used to encourage patients to continue care and to support home-based care. Cell-Life's
Medical Abortion Support service, for example, provides information for women undergoing a
medical abortion at home, while TB Free SMS reminds patients to take their medication.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 105
• Civic participation apps and services are mostly designed to encourage voting – specifically in
South Africa's national and provincial government elections, which took place during the
period inventory data was being collected. They take advantage of high mobile penetration but
low voter turnout among youth. Also in this category is Corruption Watch's Mxit portal and
mobile web site, which encourages users to be informed about and report corruption.
Figure 10 shows project frequency per archetype. Most projects were assigned a single
archetype, with the exception of a few projects that combined counselling and info services, further
education and training and job search, or data collection / remote worker monitoring with constant
contact between visits by remote workers.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 106
Counts by archetype and sector are shown in Figure 11. Reading the figure horizontally
confirms that most archetypes are confined to one or at most two sectors. In some cases this is an
artefact of the way the archetypes are defined – pill containers, for example, are clearly most likely to
be found in the health sector, and curriculum and practice apps and services in education. Other
archetypes, such as info services, crowdsourcing and constant contact, are less obviously tied to a
Figure 10: Inventory apps and services grouped by archetype
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 107
particular sector. Nevertheless, both are most common in the health sector, with info services also
addressing social issues such as drug abuse, gangsterism, teen pregnancy and gender-based violence
that affect youth. Mobile data collection and remote worker supervision models are common in the
health sector, although it should be noted that most apps of this kind are configurable and can be used
to collect many kinds of data (two mobile data collection apps actively working in multiple sectors are
coded a 'platform – applications in multiple sectors' for this reason). Further education and training
apps and services are split between attempts to connect job-seekers with employment directly and
education and training intended as a stepping stone to a job. As with the topic coverage of social issues
apps and services, this demonstrates as a field m4d apps and services is strongly youth-oriented.
Commonalities between apps and services in the same sector, sub-divided into a small number
of archetypes, suggest common factors at play despite different developers, implementers and funders.
Mobile apps and services may replicate established modes of intervention by sector, such as providing
health information to at-risk populations or improving access to reading materials to improve literacy.
Alternatively mobile platforms may provide an opportunity to address a pressing issue widely
recognized in the sector. Either way, m4d apps and services are influenced by commonly held beliefs
about the form, content and target audiences of development interventions.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 108
Figure 11: Inventory apps and services by archetype and sector
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 109
Technical possibilities at platform level, combined with usage patterns, produce a set of
affordances and constraints that shape “forms of computational expression” on mobile platforms.
Figure 12 shows archetypes differentiated by technology platform. Pill containers, which use SMS
exclusively, and data collection / remote worker supervision apps, which are primarily J2ME and
smartphone apps although some have additional functionality accessible through SMS. Curricula and
practice apps and reading/literacy materials engage their users primarily as content consumers, and
make use of Mxit and the mobile web. Counselling services are nearly all Mxit-based, taking
advantage of the platform's discoverable private chat functionality and youth-oriented user base.
Constant contact and mass reach services make greater use of SMS and USSD, appropriate to short,
frequent messages. Archetypes generally do appear to have a relationship with particular technology
platforms, or with groups of technology platforms with similar functionality – SMS and USSD, Mxit
and the mobile web, J2ME apps and smartphone apps.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 110
The sections that follow discuss the user experience, product development and development
intervention dimensions of the apps and services in the inventory, using categorization by both
archetype and technology platform to suggest relationships and possible explanations. Together with
the other dimensions, the archetypes represent the 'solution space' of computational forms of m4d apps
and services.
Figure 12: Inventory counts by archetype and mobile technology platform
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 111
5.3 User experience dimensions
5.3.1 Languages
South Africa's population is linguistically diverse, with 11 official languages. Only 9.6% of
South Africans reported English as their first home language in the 2011 census (Statistics South Africa
2012). Zulu (22.7%), Xhosa (16.0%) and Afrikaans (13.5%) were more common. Nevertheless,
English is the political and economic lingua franca, and widely spoken as a second language. In terms
of mobile platforms, Xhosa, Zulu and Afrikaans are increasingly supported as language settings on
feature phones, and in 2014, Vodacom used Xhosa and Zulu support as a selling point for its low-cost
Android tablet (“A cheap new tablet that can speak Zulu”, 24 February 2014), although most apps are
still English-only.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 112
Figure 13 shows language representation among apps and services in the inventory. As
expected, English predominates, with 47 English-only apps and services and 13 supporting other one
or more languages in addition to English. These figures exclude apps where no language information
was available and configurable software-as-service apps where implementers can supply translations.
Apps and services that support basic phones are more likely to also support multiple languages – of 13
Figure 13: Language support in inventory apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 113
basic phone apps, 6 support multiple languages compared to 7 out of 39 feature phone apps.
Examining apps and services that support multiple languages confirms that multiple language support,
like basic phone support, is part of a strategy of wide accessibility. For three apps in the reading
materials and literacy category, Yoza Cellphone Stories, FunDza and Nal'ibali, this meant providing
home-language reading materials. While data on funding was difficult to obtain, it also seems that
well-funded projects, such as Project Masiluleke, MAMA, Just Tested for HIV and VIP:Voice, are
more likely to incorporate multiple language support.
5.3.2 Cost to user
In both my interviews and prior research, cost to user is considered critical to adoption of m4d
apps and services. Specifically, apps and services that are not either free or very cheap (with free being
strongly preferred) are unlikely to be widely adopted. As discussed in Chapter 4, mobile network
billing mechanisms make providing free end-user access challenging, particularly for apps and services
without an operator partner. Nevertheless, nearly all the apps and services included in the inventory –
68 of 71 - are accessible either free or for the cost of data used only, which tends to be quite cheap. The
remaining three services require the user to pay either for an SMS or for the cost of a USSD session,
which are also relatively low-cost. The breakdown by cost to user is shown in Figure 14.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 114
The 'data charges only' category includes nearly all apps and services on Mxit and the mobile
web. Mxit is sometimes zero-rated by operators as part of product differentiation strategies aimed at
the high-churn youth market. For example Cell C, the 'third network', zero-rated Mxit traffic for all
prepaid data users for a period of around sixth months. The 'data charged or zero-rated' categories
includes some Praekelt Foundation projects, notably YoungAfricaLive and Ummeli, which are free to
Vodacom users by virtue of running through Vodacom's mobile web portal. It also includes Nokia
MoMaths, which negotiated zero-rated data with MTN and Cell-C. The disadvantage of this strategy is
that, in case of an operator portal, separate access needs to be provided to users on other networks. In
any negotiated zero-rating situation, users on other networks will need to pay data costs.
Figure 14: Inventory breakdown by cost to user
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 115
With the exception of the non-operational SIMMed pilot and Lungisa, a Cell-Life pilot service
also accessible via the mobile web, none of the inventory apps and services required users to pay SMS
costs. In most cases, only outgoing SMS is used, making it straightforward for the service (as sender)
to assume SMS costs. Services that require a transaction where the user requests information, including
several of Cell-Life's projects, avoid SMS costs by responding by SMS after contact is initiated with a
missed call or USSD session.
Project Masiluleke, which used advertising space in please call me messages to distribute
public health messaging and encourage callers to contact the AIDS helpline, is free to the user but not
user-initiated. Ironically, calls to the AIDS helpline from landlines are free, but calls from mobile
phones are not.
Finally, it should be noted that with the exception of two apps selling paid eBooks in addition
to a free offering – EverEgg and FundZa – none of the 4d apps and services operate on a
micropayment model. Evidence from the inventory suggests that at least in South Africa, users are not
a viable source of funding for m4d apps and services.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 116
5.4 Product development dimensions
5.4.1 Developers
As shown in Figure 15, the largest proportion of apps and services in the inventory – 32, or
45% - were developed by organizations specializing in ICTD. In this category, Cell-Life and the
Praekelt Foundation appear frequently, both as direct grantees of ICTD funding and as technical
partners of other grantee organizations. Several specialist mobile data collection organizations,
Mobenzi, Dimagi and Vodacom mHealth (formerly GeoMed, an independent company later bought by
Vodacom) operate with a similar funding mix. Every1Mobile operates several mobile web sites
Figure 15: Developers of inventory apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 117
(“communities”) with youth-focused, development-oriented content, funded by selling advertising to
brands looking to access developing world youth.
For the rest of the inventory, general software development contractors and in-house
developers are responsible for 11 and 14 apps and services respectively. Technology sector partners,
such as mobile operators and mobile technology companies, were found to have little involvement in
building m4d apps and services with the notable exception of Mxit Reach. Mxit Reach is a division of
Mxit specializing in building m4d apps and services with development organizations as partners, both
as a corporate social responsibility strategy and to encourage adoption of Mxit. It is responsible for 10
of the 13 apps built by a tech sector partner. The remainder include Nokia MoMaths (Nokia), Imbizo
Men's Health SMS, developed as a corporate social responsibility project by SMS aggregator Bulk
SMS, and Bookly, a corporate social responsibility showcase project of digital marketing agency
Native.
To various degrees, all the ICTD specialized organizations maintain a product or set of
products – such as a mobile data collection app, Every1Mobile's mobile web communities or Praekelt
Foundation's Vumi content delivery platform – that are customized in implementation. Mxit Reach
does the same. Together they are responsible for more than half the apps and services in the inventory.
The appearance of clusters of similar m4d apps and services – represented here as archetypes - is
influenced by product-oriented ICTD specialist organizations. In some cases multiple m4d apps and
services are implementations of the same product. In others several ICTD specialist organizations base
their business model around the same information need, such as mobile data collection.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 118
5.4.2 Funders
Nearly half of the apps and services in the inventory (29, or 41%) are self-funded, including
advertising-supported models, public relations and promotional projects, and software-as-service
providers mostly in the mobile data collection space.
Development funders, including international agencies such as the World Health Organization,
UNICEF, international and local foundations, and bilateral donors such as USAID, are involved in 28
(39%). Although the inventory coded individual projects based on specifically mentioned funders, in
practice there is a some overlap between the two categories particularly for software-as-service
Figure 16: Types of funders for inventory apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 119
providers who balance funding for specific implementations with product development that promises
sustainability by meeting the needs of software-as-service clients.
Technology companies, including mobile operators, were not a major source of funding for
apps and services in the inventory. In particular, with the exception of two projects with Vodacom
Foundation funding (Just Tested for HIV and YoungAfricaLive), mobile operators do not appear to
focus on m4d as part of their corporate social responsibility portfolio. Similarly, very few apps and
services mentioned government funding, although there are several apps and services in non-monetary
partnerships with government departments, particularly national and provincial health and education
departments.
Figure 17: Funder type by technology level
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 120
Figure 17 cross-tabulates funder type with whether an app or service requires a feature phone
or is accessible on basic phones. Although we might expect development funders to value inclusion
highly and therefore support apps and services that are accessible on basic phones, this does not
emerge as a strong pattern. One explanation is that the limitations of basic phone technology platforms
– SMS and USSD – make them less attractive to development funders despite a preference for
inclusivity. Another, explored further in the Intended Users dimension, is that projects m4d mostly
target better-connected low-income youth rather than the poorest or those living in deep rural areas.
5.4.3 Mobile operator involvement
Apps and services in the inventory overwhelmingly did not receive funding, technical support
or other assistance from mobile operators. Of 71 apps and services, 60 received no assistance. Three
received funding from the Vodacom Foundation (Just Tested for HIV, TB Free SMS, as well as
YoungAfricaLive). Two projects each received either technical assistance and preferential access to
operator infrastructure or zero-rated services, and five projects received both. Technical
assistance/preferential access included five projects accessible through Vodacom's Vlive portal (zero-
rated for Vodacom users, but only accessible to them), all developed by the Praekelt Foundation.
mCent claims agreements with operators enable it to compensate users for taking marketing surveys
with airtime, through access to airtime credit mechanisms. Cell-Life Capture has made use of a zero-
rated access point provided by Vodacom to allow users to upload data collection forms without credit.
Other zero-rated services include Nokia MoMaths, which can be accessed free by Vodacom users, and
Project Masiluleke, to which MTN donated advertising space in please call me messages.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 121
5.5 Development intervention dimensions
5.5.1 Sector
In keeping with the distribution suggested in the clustering by archetype, the inventory exercise
demonstrates a strong bias towards health and education in m4d apps and services. Figure 18 shows
the breakdown by sector. Nearly half of the apps and services in the inventory (33 of 71, or 46%)
address issues in the health sector, followed by education (20 apps, 28%) and social issues (9 apps,
13%). Within health, HIV/AIDS, pregnancy and sexual health are well represented. Education covers
mostly curriculum content, with high school maths and science predominating. Lungisa, Cell-Life's
system for residents to report infrastructure issues was the only service that addressed physical
Figure 18: Inventory apps and services by sector
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 122
infrastructure issues. Entertainment content, although frequently included as part of the presentation of
health or social issues content, was a primary focus of only two apps - D-Siders, the serialized mobile
story by Every1Mobile, and Moraba: It's Your Move, the game about gender-based violence.
5.5.2 Intended users
The previous section's breakdown by sector indicates a focus on accessing a (youth) audience
directly, rather than working with intermediaries such as local government (responsible, for example,
for physical infrastructure). Classifying apps and services in the inventory by intended users confirms
this finding. Youth are cited as the intended users of 32 apps and services (45%). Other demographic
groups specifically targeted are women (5), men (2) and individuals from low-income groups (2).
These categories are non-exclusive, and given the user profile of platforms such as Mxit and the
mobile web, it is likely that youth make up an even larger proportion of actual users, particular for an
additional 20 apps that cite no specific group of intended users beyond the general public.
In keeping with the perception of mobile phones as a personal device, only two services – Cell-
Life Communicate, which was implemented for the Treatment Action Campaign as a co-ordination
tool, and Imbizo Men's Health SMS, which worked with members of men's group Imbizo – were
designed to be used by members of a voluntary organization. Both are SMS apps with a constant
contact model. Perhaps surprisingly, all the apps and services in the civic participation sector targeted
individuals, with participation envisaged as voting in elections, or learning about and sharing an
opinion on social issues. Mobilisation and emergent group co-ordination were not represented in
inventory apps and services, despite being considered in the literature on the relationship between
mobiles and development.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 123
Of the apps and services intended to be used or managed by an intermediary – either
employees of a service provider, such as clinic workers, or clients/members, such as patients or school
learners - mobile data collection and constant contact models are most common. Most education apps
are based on a disintermediation model, where access to education content is provided to the student
directly without involving teachers. The Teachers' Corner app on Mxit, which is designed for teachers
rather than students, is an interesting exception.
The intended users of an app or service are closely related to technology choices, with Mxit
and the mobile web generally associated with youth-focused apps and services. J2ME apps, because of
poor discoverability, phone model requirements and general management overhead, are most
commonly developed for employees of a service provider who is able to provide support and supply a
Figure 19: Intended users of inventory apps and services
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 124
device. Wide accessibility is the main reason for choosing SMS and USSD, with vulnerable groups
such as TB patients, whose disease is strongly associated with poverty, presumed more likely to benefit
from a service accessible from basic phones.
5.5.3 Geographic Reach
Considering the geographic reach of m4d apps and services is relevant to scalability, and
interrelated with operator involvement. The inventory exercise was restricted to apps and services
operating in (or intended to operate in) South Africa. However, platforms differ in the geographic
access they afford. Mobile web sites are generally accessible from anywhere with Internet access. SMS
and USSD services, however, rely on country-specific, operator-specific gateways. While a third party
provider might provide a level of abstraction for outgoing messages, multi-country coverage for SMS
and USSD services is still the exception rather than the norm. For J2ME apps, the infrastructure
required to install the app and manage users and phones generally restricts access to users managed
within a particular organization. Although Mxit has users in other African countries as well as India, in
practice most users are South African.
Operator involvement also influences geographic reach. Services such as YoungAfricaLive and
Ummeli, which are free to users by virtue of a an infrastructure access arrangement with a particular
operator, are restricted to that operator's coverage area. YoungAfricaLive is in fact accessible in several
African countries through separate local arrangements, but still only to Vodacom users. Other services,
such as Nokia MoMaths, are free to users on a particular operator in a particular country, with data
charges applied to users in other countries or on other networks.
Audience, specifically whether a general audience is envisaged or whether the intended users
are linked to a particular organization or group, is also relevant. Most mobile web and Mxit info
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 125
services fall into the first category, anticipating a general audience. Constant contact services generally
extend the scope of an interaction with an organization, such as a visit to a clinic, and accordingly are
bounded in physical space. MAMA SMS, which targets pregnant women in general, is an exception,
although as an intervention it functions partly by ensuring that pregnant women seek care as necessary
at a primary care facility.
Counselling, like info services, is in theory accessible to anyone with an Internet connection,
although in practice most Mxit users are located in South Africa. Furthermore, most info services and
all Mxit-based counselling services are linked to South African organizations. Their content is most
relevant to South African users, as is their ability to link users to appropriate crisis response or
preventative services. Similarity, curriculum and practice apps and services are closely tied to the
South African syllabus for various subjects and grade levels.
The software-as-service model of mobile data collection and remote work apps supports
localization for different countries as as part of routine customization undertaken for each
implementation. Like constant contact, however, the audience is specific rather than general, usually
comprising employees of a particular organization. Mobile data collection apps such as CommCare
claim several hundred projects and several thousand total users, an order of magnitude different from
Mxit apps or mobile sites which claim user numbers in the hundreds of thousands.
Wide geographic reach is neither necessary nor sufficient for impact. However, trends in
inventory apps and services argue against making accessibility claims for m4d apps and services based
on global mobile penetration. Platforms, content and audience, as well as arrangements with mobile
operators, keep most m4d apps and services firmly anchored within national boundaries or to specific
organizations.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 126
5.5.4 Development theory of change
Categorising the inventory apps and services by a 'theory of change' answers the question 'what
is the mechanism of this app or service's intended contribution to development?'. Theory of change
approaches, which come from development planning, aim to articulate the link between an the
activities of a development intervention and its intended outcomes (Stein & Valters 2012, p.4). They
are used for project planning, evaluation and impact assessment, and can be particularly useful in
identifying barriers between outputs – such as page views or completed math exercises - and outcomes,
such as health behaviour change or school achievement.
For the inventory, I initially wrote free-text description of the implied theory of change for each
app or service. These were then grouped into (non-exclusive) categories as follows:
• Behaviour change, in which the app or service contributes to development by
providing information intended to encourage the user to change a behaviour or set of
behaviours.
• Better service provision, such as improved patient care or better teaching.
• Access to formal services in which virtual access compensates for lack of physical
access to services such as counselling.
• Skills/education/literacy in which the app or service helps users improve either
general or specific skills, which in turn improve employability or (in the case of
literacy) general communication efficacy.
• Livelihood support, comprising apps and services that provide opportunities to
improve their livelihoods by earning an income or gaining employment.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 127
• Better information – data collection, monitoring and evaluation, including apps
and services that collect data for research, oversight and process improvement
purposes
• Social networking apps and services, which help users develop and maintain social
networks for a particular purpose such as supporting young mothers or job-seekers.
• Audience-building, in which apps and services develop and provide access to a
particular target audience for development communication.
• Democratic participation, for apps and services that support participation in
democratic process.
Figure 20: Inventory apps and services by development theory of change
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 128
As shown in Figure 20, behaviour change apps and services predominate, with 29 of 71 (41%)
aiming to contribute to outcomes premised on behaviour change. These are mostly info services, and
mostly in the health sector. Similarly, the skills/education/literacy category comprises 14 apps and
services from the reading materials, general education and curricula and practice archetypes in the
education sector. Better service provision appears mostly in the health sector with the exception of the
Teachers' Corner app and Lungisa, Cell-Life's physical infrastructure issue reporting tool. Access to
formal services, in which virtual access compensates for lack of physical access to services, is the
primary goal of various counselling services. One contribution of the classification by theory of
change, then, is extending the description of archetypes.
Examining the underlying theories of change for the m4d apps and services in the inventory
also underscores the multiple contingencies involved in achieving the desired outcomes. The gap
between, for example, providing information and achieving behaviour change can be wide. Similarly,
even if monitoring and evaluation data is available to decision-makers, process improvement is
contingent on the data being suitable for, and used to, take appropriate action. Mitigating lack of
physical access to services with virtual access is limited to particular kinds of services, in this case
counselling. This is not to suggest that m4d apps and services are not useful – they are, and in some
cases perhaps decisively so. Rather, many factors intervene between outputs – the app or service being
used – and outcomes. Additionally, as will be discussed in 5.5.6, outcomes and impact are much harder
to measure than outputs, and as a result are seldom comprehensively evaluated.
5.5.5 Role of technology
In building the inventory I initially planned to classify m4d apps and services according to the
role played in each by mobile technology – increasing reach, disintermediation, decreasing costs etc. In
practice, apps and services in the inventory, and particularly apps and services on the same platform or
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 129
of the same archetype, exploit the same set of technological affordances. The remainder of this section
describes the role of technology at the level of the field as a whole, highlighting archetypes or
technology platforms for which particular aspects are most relevant.
Across every platform, reach and accessibility are highly valued. Mobile network coverage
and penetration figures are cited frequently. For Mxit, the size of the user base and the fact that most
users are teenagers and young adults provides justification for youth-focused services. Market research
and survey apps and services promise access to low- and middle- income youth, highlighting the
challenge of both providing information to, and obtaining information from, this group through other
communication channels.
In many cases, reach and accessibility is paired with disintermediation. Market research and
survey apps promise direct access to the youth market as opposed to employing someone to run focus
groups. Curriculum and practice apps let students access textbooks and exercises without the
involvement of schools to teachers. Job search services such as Ummeli connect job-seekers to
employers directly rather than involving agencies or labor brokers. Conversely, several mHealth apps,
grouped under the remote work supervision archetype, support intermediaries such as community
health workers in providing services. Teachers' Corner on Mxit is another example of an intermediary
support app.
SMS apps and services in the constant contact model rely the push capability of SMS. Unlike
USSD, the mobile web and Mxit, the user does not need to initiate an SMS interaction. Constant
contact is one of the few archetypes for which basic phones are not inherently limiting, as content
requirements are generally low (short, frequent messages) and the cost of sending outgoing messages
can be borne by the service operator.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 130
Despite reports of shared phone use in other countries, Research ICT Africa (2012, p.25) found
that South Africans across the income spectrum consider the phone a personal device. Mobile
platforms' affordance of private access to information and private communication is particularly
important to counselling apps and services as well as info services offering information about issues
that may be stigmatized.
Most m4d apps and services do not directly access on-device storage, media capture and
computation capabilities. App discovery, installation and compatibility problems on the J2ME
platform – the only way for featurephone apps to access device functionality – are at least partly
responsible. The exception is mobile data collection and reporting apps, which typically use all three.
However, in general they also supply reliable, tested devices to users.
With some exceptions, outgoing communication of information, rather than computation or
user-generated content creation, is the main role of mobile technology in most m4d apps and services.
Market research, survey and data collection apps rely on obtaining information from users, and
allowing users to comment and ask questions is a relatively narrow deployment of user-generated
content. Nevertheless, divergence from a user pays model, as well the need to consider hardware and
software limitations on lower-end phones, give rise to a distinct set of affordances and constraints for
m4d apps and services compared to apps and services for the general market.
5.5.6 Measures of success
The question of how m4d apps and services measure success is partly a characterization of
discursive framing, identifying the kinds of evidence used to make a case for an app or service as a
solution to a development problem. Not all evidence is made public, much less promoted publicly.
However it is reasonable to infer a public argument from what is. As a starting point, the finding that of
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 131
71 apps and services, only 46 (65%) made mention of any measure of success is notable. The
remainder are discussed solely in terms of innovation, with no evidence presented for outputs or
outcomes in relation to a development problem.
For those that do report measures of success, different platforms afford different ways of
measuring outputs. Mxit has a chat paradigm that makes it trivial to report unique users, since each is
logged in to a user account while on the platform. Page views are the basic unit of measurement for the
mobile web. Of the inventory 71 apps and services, 32 (45%) report either user numbers or page views,
and 20 (28%) report these exclusively. Of the basic phone -only services, only three (TxtAlert, Project
Masiluleke, MoMConnect) report user numbers as a measure of success, all in conjunction with other
measures.
Apps and services with a software-as-service model tend to report implementer numbers,
sometimes in conjunction with user numbers. The number or countries in which the app or service is
used is also sometimes reported. Like user numbers, these measures are presented as an indication of
scale and reach. They are also used to demonstrate diverse applications of an app or service. Dimagi,
Cell-Life, Mobenzi, the Praekelt Foundation and Vodacom mHealth all provide case studies on their
website as an extended demonstration of possible applications.
Evaluation research is present but not common, and, with the exception of Nokia MoMaths,
confined the health sector. Two Cell-Life projects – Just Tested for HIV and Medical Abortion Support
- obtained funding to conduct randomized controlled trials (RCTs) comparing users who received
informational and motivational SMSs with those who did not. The first, by de Tolly, K., Skinner,
Nembaware and Benjamin (2012), found that users who received 10 motivational SMSs were 1.7 times
more likely to get an HIV test than the control group. The second study (Constant, de Tolly, Harries &
Myer, 2014) found that medical abortion patients who received timed SMSs for 30 days experience
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 132
less anxiety and stress than the control group. WisePill cites a study by Haberer et al (2013) that
demonstrated a correlation between adherence to HIV medication as measured by the device and
suppression of viral load. Dimagi maintains an “evidence base” compendium of research involving
CommCare, including several studies that found process improvements among community health
workers using CommCare, and one that found improved access to antenatal care among their pregnant
women clients (Chatfield, Javetski, Fletcher & Lesh, 2014). Finally, Nokia MoMaths cites (but has not
distributed) an evaluation study conducted by a consultancy company, in which a small improvement
in results was found for learners who completed all exercises.
Ten of the inventory apps and services mention receiving awards as a measure of success.
Awards for innovation or social innovation, given by both local and international bodies, are an
opportunity to put out a press release and a way for early-stage apps and services to gain visibility and
funding. WisePill and SIMPill, neither of which ultimately made it to commercial production, mention
awards prominently. Competitive awards for early-stage ideas are part of the culture of technology
startups. Their presence in m4d is a reminder of the alliances that constitute the field, with industry
bodies such as the GSMA and the Mobile Marketing Association using awards to promote a socially
inclusive image of mobile technology. In the best case, awards are a way to identify promising early-
stage apps and services, providing them with publicity opportunities and perhaps some funding.
However it is also worth considering distortions introduced to the field and to stakeholder perceptions
of m4d apps and services.
Overall, it is difficult to make claims about the success of m4d apps and services based on the
measures of success they make available. Lack of impact evaluation studies for m4d apps and services
is part of a larger incentive problem in conducting and publishing impact evaluation for development
interventions. For Mxit and mobile web apps, output measures such as user numbers are reported
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 133
because they are impressive and readily available. A few projects, such as Yoza Cellphone Stories and
YoungAfricaLive, report additional use data such as comments and length and frequency of visits,
which, although they are still output rather than outcome measures, make a stronger case. This should
be encouraged, as it is also readily available through web analytics tracking services such as Google
Analytics. Awards, which reward novelty rather than encouraging reflexivity, are appropriate for early-
stage ideas but do not appear to be good indicators of potential impact.
The inventory exercise found that published evidence on the impact of m4d apps and services
in South Africa is limited. It is worth considering the discursive, structural and technological factors
that have created a perception of impact – or at least of potential - in the absence of reliable evidence.
The practice of reporting user numbers and user behaviour statistics, which are readily obtainable
through secondary data analysis of operational metadata, is prevalent and worth further investigation.
This issue is raised again in the interview findings in 6.2.6, and is the main subject of the quantitative
analysis of operational metadata conducted as Part 4 of this research.
5.6 Discussion: South African m4d apps and services landscape
The inventory research component considered m4d apps and services as “forms of
computational expression” (Montfort & Bogost 2009, p.3), identifying points of concentration along a
set of dimensions in order to understand the forces shaping the field. The findings variously challenge
the idea of “serving the underserved through mobile” (GSMA, 2013). M4d apps and services are both
functionally concentrated, and constrained in terms of user base, funding model, and intervention
design. Without a field-level inventory exercise, it is easy for research on individual projects or
collections of case studies to produce a false perception of diversity.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 134
The main finding of the inventory exercise, in addition to breakdowns by technology platform,
sector, intended users and development theory of change, was that m4d apps and services are mostly
variations on a small number of archetypes. There is some diversity in content, but much less in form.
Fourteen archetypes were identified, namely info services, constant contact, curricula & practice,
surveys/market research, counselling, data collection / remote workers, general education & games,
reading/literacy materials, civic participation, further education & training, job search, pill containers,
crowdsourcing, mass reach. Info services were the most common, making up around a quarter of all
apps and services.
Archetypes appear in relation to the affordances of particular technology platforms –
counselling delivered via Mxit chat, info services on Mxit and the mobile web, constant contact
accomplished via SMS, and mobile data collection / remote worker support using J2ME and
smartphone apps. The association of mobile data collection / remote worker support apps with
software-as-service offerings and implementation assistance from specialist technology companies
suggests archetypes may also align with funding models and organizational forms. On the one hand,
concentration around archetypes provides a way to identify promising models for m4d apps and
services. On the other, the lack of diversity suggests a narrowing of the problem space for m4d
interventions.
The finding that youth are the intended users of nearly half of m4d apps and services in the
inventory provides further support for the idea of a narrowed problem space. In interviews, the design
process described by several interviewees involved matching the affordances – both technical and
social in the form of the user base – of platforms to an appropriate development intervention. Rather
than beginning with a predefined development problem and a proposed intervention, the design
process involved considering the platform and possible interventions together. To connect this again to
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 135
archetypes, there are a large number of mobile web/Mxit info services in the inventory, concentrated in
the health sector and premised on achieving development outcomes through behaviour change. This
reflects the affordances of the mobile web and Mxit in terms of functionality and user base, as well as
the prevalence of behaviour change communication as development intervention targeting youth.
Mxit, described by one interviewee as an “aberrational success” created by the artificially high
cost of SMS, is a unique feature of the South African m4d apps and services landscape. It is a popular
platform for m4d apps and services, with nearly half the inventory apps and services available on Mxit.
From the perspective of designers of m4d apps and services, two features are relevant. First, Mxit apps
use very little data due to design constraints, which specify they must be mostly text-based with a few
images. Data costs are thus kept low. Second, Mxit has a dedicated menu item, Mxit Reach, that
groups m4d apps and services. This provides discoverability to a large, content-hungry audience. As
Mxit loses market share to instant messaging apps such as WhatsApp and social networks such as
Facebook, the field loses its easily accessible youth user base. The info service and live counselling
archetypes, which rely most on discoverabily and the chat paradigm introduced by Mxit, are likely to
suffer.
Along with platform affordances, intended users, and a hybrid design process, sustainability
and funding models contribute to the narrowing of the m4d apps and services problem space. None of
the inventory apps and services earned revenue from users, and interviews confirmed that a free-to-
user model (with the exception of small data charges) is strongly preferred. The ability to provide zero-
rated services was the main reason for seeking operator partnerships. It was also a key point of non-
congruence between the design logic of platforms for m4d apps and services and the needs of low-
income users in the developing world.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 136
In the absence of end-user revenue streams, apps and services in the inventory were evenly
split between those that were self-funded, including promotional/social responsibility projects by
commercial organizations as well as software-as-service models, and funded projects with support
from development funders. The interview findings identify a relationship between funding sources and
sector concentration, noting that health and education projects are more attractive to funders because
they are seen as a social service and not expected to become self-sustaining.
Nearly half the apps and services in the inventory operate in the health sector, which, in
addition to funding, benefits from a long history of computerization projects. Health sector apps and
services support healthcare workers and chronic patients as well as reaching a youth audience directly
with behaviour change communication. Moves to incorporate m-health within a national e-health
framework, also described in the interview findings, indicate increasing maturity in the sector.
Conversely, projects addressing (youth) unemployment, largely absent from the inventory, were
identified by interviewees as particularly in demand. This may indicate a mismatch in m4d priorities.
Despite positive signs in the health sector, the evidence base for m4d apps and services as
development interventions is not well established. As noted by Rashid and Elder (2010), cost-benefit
analyses are lacking in the field generally. With a few exceptions, again mostly in the health sector,
apps and services in the inventory also did not report impact measures. Instead, many cited measures
of use, such as user numbers and engagement statistics (for example duration or frequency of use)
derived from secondary analysis of operational data. Mentions of innovation awards, often awarded by
mobile industry bodies, were also notable. The secondary data analysis conducted in Chapter 7, which
analyses operational data from 300 projects using CommCare for mobile data collection and remote
worker support, further investigates the role of data-driven research in m4d interventions.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 137
CHAPTER 6: Interview findings
As with the inventory research component, findings from the key informant interviews
demonstrate how the promise of m4d apps and services has been translated into practice in South
Africa. The sections that follow report project-level experiences from developers and implementers of
m4d apps and services, complementing the field-level counting exercise undertaken in the inventory.
General commentary covering trends in the field, from these interviewees as well as several South
African m4d researchers, is presented at the end of the section. To maintain the anonymity offered to
interviewees, app/service and organization names are redacted in reporting findings.
The overall picture of m4d apps and services that emerged from the interviews identifies them
as both promising and multiply constrained. Put another way, there is encouraging evidence that m4d
apps and services are used and useful, but both the potential user base and the possibilities for m4d
apps and services are narrower and more contingent than expansive narratives of their potential would
suggest. One set of constraints arises from the technology platforms and underlying infrastructures of
the mobile network. Another set, with clear links to previous versions of the ICT4D consensus,
foregrounds the motives of stakeholders, including funders and mobile operators, and the structure of
the external environment for m4d apps and services.
These are not mutually exclusive – for example, operator control of the mobile ecosystem is
defined in the GSM standard and reinforced through technologies and institutional arrangements
layered over it. Taken together, they suggest that m4d apps and services have a lot in common with
other generations of ICT4D intervention. Some innovations, such as mobile money in Kenya, are
broadly used and transform a set of social and economic activities. Many more are useful in focused
ways, such as providing health information to youth or supporting remote workers. However, as with
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 138
the inventory, the interview findings do not support the idea that marginalized individuals' access to a
personal communication device will necessarily be followed by access to a broad range of relevant
apps and services. As one interviewee put it, “people just generically say people are using mobile
phones in all kinds of great ways and that’s most certainly not the case.”
6.1 Technology
One of the early pieces of feedback I got on choosing an infrastructure studies/platform studies
approach to evaluating m4d apps and services was that a critical stance might not be the most
productive in terms of normative insight. The reviewer suggested a more pragmatic framework based
on matching the priorities of m4d interventions to a set of affordances and constraints of each platform.
Interventions prioritizing basic phone support, for example, would be directed to consider an SMS or
USSD solution, while interventions targeting urban youth might expect users to have featurephones
and be able to access a mobile web site (G. Marsden, personal communication, 2 December 2013).
Perhaps unsurprisingly, this framework was widely shared among interviewees involved in
building m4d apps and services. The trade-offs involved in working around platform constraints give
rise to a limited set of possibilities, and interviewees tended to concentrate on how well their model fit
within them. Following a general discussion of interviewee's decision to pursue of a mobile solution,
this section discusses constraints and considerations for the various platforms.
The origin story of interviewees' projects generally involved an idea for an app or service
rather than a needs assessment around a broad development issue. Here, an interviewee describes the
origin of an m-novel project:
“I was at [the foundation] as a fellow, and you had to come up with a project that basically
used technology in an innovative way to solve some of the education challenges facing the
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 139
country. And what I kept getting was that, low literacy levels, there's not enough books,
there's not a culture of reading, but there are phones, and especially in teenagers' hands.
And so the thinking was how can we actually use that reality of this wide uptake of phones
and relatively cheap data”
Another project, aiming to improve smallholder agriculture by providing price information,
arose from a combination of research on the detrimental effects of information asymmetry and the
availability of SMS tools through third-party service providers.
“I started to read a bunch of articles about the wholesale market for agricultural produce
and how it bad it was in Africa especially. So, the problem was basically that the middle
men are able to capitalize on price information and asymmetry, asymmetrical information
basically. And they capitalized on that to the detriment of the farmers well that was the idea
a couple of years ago. That's not 100% correct now. There has been a lot of discussion and
papers and stuff. But that's like one part of a bigger problem, the wholesale market. And so,
regardless, a couple of years later, so I had that idea for a long time. And a couple of years
later, I decided that it was actually pretty easy to start getting that information out. So, I
noticed that it was pretty easy to hook in to SMS systems. There were a lot of third party
SMS systems out there that had really easy APIs.”
A project that did not originate as a mobile app provided an interesting example of the
challenges involved in taking this path. Here, the interviewee describes the process of getting general
software development contractors to build a mobile version of an existing web site, after a South
African foundation approached her to offer funding:
“My goodness it was a painful process, it really was, oh my goodness I promise you with
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 140
all the grey hair I got now I mainly because of [this project]. We started with the – you
know the thing was as well we didn’t know about it ourselves. We had a website we knew
about mobi, I certainly knew nothing about building a mobi site. So it really was a lot of
trial and error, so we interviewed a number of different companies and even then there
weren’t that many companies doing mobi development and the ones that were, were doing
big developments for big companies and big money, so the smaller companies who we
could afford were not that interested in doing – they didn’t have the expertise that was the
main problem they didn’t have the expertise to do it, so eventually we did find a company
that we started with and it was really, really a hard slog, we ended up and they were so over
deadline and they didn’t produce what we wanted and it was very difficult and we
eventually we had to fire them and get somebody else, and that wasted a lot of time and lot
of money, made us not look great in our funders eyes, because we hadn’t kept to their
deadlines. So then we took on another company that completely developed it from scratch
and then, well they started developing from scratch and then they went into liquidation,
yeah they phoned me on a Friday night at 9 o’clock and said we’re going into liquidation
as of right now we will not be doing any more work on your site, and it was – they didn’t
do it in any normal site building languages, they did it their own bespoke design so it was
almost impossible for anybody to take over. So we did try, we got a company that tried to
take up what they had done and they actually eventually just said listen we can’t work with
this, it’s a mess, and then we started again with a brand new company from scratch and so
it was very painful and then it took about another year and we got to a place where we
were able to launch, we launched only towards the end of 2012, and since then it's been
fine, we’re still sitting with legacy issues though in that we should’ve built it in a platform
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 141
that everybody was comfortable and anybody could just takeover, even the last one was
written in php, but it was custom built hardcoded, all kinds of stuff...”
Taken together, these responses suggest that a mobile solution is often pre-selected based on a
general understanding of a development problem. Two interviewees addressed this issue directly,
cautioning against a technology-centric approach.
“I like telling people start with a problem, what are you solving and then find all possible
solutions that you can solve that problem with. And part of it is the app. The other part is
actually human beings. So, it's 20% app and 80% human. So, you need to first of all
identify the problem. Once you've identified the problem, just know how to approach it [..]
I will say let technology be the last thing you think about. Think about the solution and if
technology is the right solution then use it. And I know this is for somebody who is either
computer scientist, it's not the best thing to say but it is true.”
“But I really think each project should really do in depth user analysis, like who is your
target audience, and really understanding them and then from that will flow, okay, if this is
the target audience, what are their needs? How do they access technology?”
In practice it appears the risk of not having technologists involved early in the project may
outweigh concerns about the appropriateness of a technological solution, or at least a mobile solution.
The relatively obscure nature of platforms for m4d apps and services – both in general during the
initial period of their availability and in relation to the specific accommodations required by m4d apps
and services – means that most projects that get off the ground are technology-centric by virtue of the
specialist technical knowledge required. Interviewee comments on the relationship between m4d apps
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 142
and services and platforms, presented in the next section, illustrate this tension between novel
possibilities and development needs.
6.1.1 Platforms
Interviewees' description of their experience with platforms for m4d apps and services were
often phrased as opportunities, based on affordances and usage. For the former, interviewees
commonly mentioned low-cost access and accessibility on basic or feature phones as reasons for
choosing particular platforms. For the latter, Mxit's large user base of content-hungry youth was a
major attraction, as was the audience of operators' mobile web portals. Criticism of platform
limitations was also readily offered, particularly for J2ME apps.
Providing a service on multiple platforms was common, with combination Mxit / mobile web
services mentioned most frequently. A Mxit plugin for the Wordpress, developed by South African
company Kazazoom, bridges the divide between Mxit's sometimes esoteric technical requirements and
a well-known and well-supported web publishing system. The advantage of this, according to one
interviewee, was that although Mxit was their major driver of traffic, it was preferable not to be
dependent on it. At the time interviews were conducted, Mxit's active user base was understood to be
shrinking, a trend that appears to have continued as WhatsApp and Facebook erode its market share.
An interviewee with professional involvement in Mxit was keen to promote the privacy
afforded by Mxit's chat paradigm, with user actions, such as subscribing to particular content or
chatting to counselling-type services, handled privately as opposed to publicly on Facebook. Citing an
education app that makes use of the chat paradigm to present small amounts of content interspersed
with questions, the same interviewee suggested that chat-based content is more appealing to youth
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 143
audience. However, other interviewees were interested in Mxit mostly for its user base, and avenues
for rendering their app or service discoverable to them.
SMS and USSD were seen as a last resort for users without data-enabled feature phones. A
project with both mobile web and USSD interfaces to web content noted that while the USSD interface
had been included for people with less access and technical experience, it was not well used.
“we built a USSD portal with those very untechnological minded people, especially elderly
people who are only really used to working with USSD, we built them in mind and that’s
not really something that’s been hugely taken up”
A secondary role for SMS is offline data collection from users' own phones, as in this example
from an interviewee working on mobile data collection with J2ME.
“In some solutions, we have let’s say 5,000 users, they would probably use the SMS
channel and somewhere collecting data and not hand out 5000 Android phones. So it
depends on the quantities.”
SMS and USSD were also the focus of several complaints about platforms, particularly in
relation to country differences. Operator control and zero-rating were also mentioned as constraints,
and are discussed further in section 6.2.4).
“I think that the trickiest one was availability of service especially in Africa. If you look at
SMS and USSD...We’ve had a lot of trouble, delivering SMS is through the aggregators [..]
from our point of view, we are delivering the SMS to the aggregator, we're getting a receipt
that it’s being delivered but the end-user didn't receive it due to a lot of problems within the
local distribution of SMSs. ”
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 144
“It’s got the limited amount of time that you can be in the session. In Tanzania, for
instance, on certain networks they use, this session is like 60 seconds and South Africa is
maybe 180 seconds so you can capture a lot more data and then even we get signal
problems where this session would drop after 20 seconds and you need to do more and
more sessions to capture a certain amount of data. So we’ve worked out a bit of that
problems with the way the USSD session is designed. So for instance, when you drop the
session, you come back in, it will continue on the same spot where you left, but that adds
cost to the solution to use three or four sessions where you should actually just one to
capture a certain amount of data.”
Another interviewee discusses SMS integration through Fortumo, an aggregator with extensive
international coverage outside Africa:
“And Africa is actually not that great coverage but still it's probably the best. I know that
they're setting up. A lot of people are trying to get into Kenya. I think Tanzania is going to
be next on Fortumo's list, Rwanda is going to be there someday. But a lot of that, they've
talked about it a lot but it's very difficult because there's a lot of monopolies in these
countries or at least oligopolies. And they have to negotiate with them. And so, yeah, so
when you're entering the Nigerian market, that's actually relatively well-regulated, the
telecomm market. So, it's pretty easy to get to. South Africa. Not the best but it's a little bit
more functional. But the easiest to integrate with with Fortumo has been Latin America
was very easy”
During the GSMA's ICTD conference session, a participant also reported problems scaling
with USSD:
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 145
“Wikipedia text in Kenya is breaking things - load. USSD doesn't scale well. It scales well
for coverage but not for traffic. Towers have a limited number of sessions they can handle
at a particular time.”
Despite being part of an existing commercial ecosystem with the possibility for airtime
payments, SMS and USSD were not seen as revenue-generating options for m4d apps and services.
Premium SMS is an established revenue channel for commercial services selling content such as
ringtones and wallpapers, but an interviewee who had tried to use it noted that in addition to operator
challenges in different countries, even a small payment was a significant barrier to uptake given the
limited value of information that could be made available over SMS.
Interviewees who had been involved with J2ME apps and services were particularly critical of
the platform. On the one hand, J2ME is the only non-smartphone option for apps that use phone
features like GPS, memory, and on-device processing. On the other, interviewees felt it was neither
designed with their needs in mind, nor managed reliably by equipment manufacturers and software
vendors.
“My impression is that it was a framework created to let cellphones have really simple
games. And that was just like a little selling item like look at our cellphone you can play
Tetris or something. And like literally a lot of documentation you read about it was like oh
you can use this to store like a high scores field or something. Because you know for high
scores clearly you never want to store more than 10. And that’s like 200 bytes and no one
will ever need more than that. So to actually do anything serious with it was a constant
uphill battle. I basically feel like we rewrote half of Android in J2ME. We just kind of
ended up making a lot of fundamental utility frameworks to fix broken APIs and to work
around bugs in the APIs. Like things that seemed like they should work didn’t work so we
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 146
had to work around them. And even then it was only because we limited ourselves to Nokia
phones which was the best supported JVM, that actually even enabled us to do anything.
Like, despite these workarounds, if we had supported other platforms we would not be able
to do anything at all.”
“I don't know if any mobile app runs well on these phones, we still to this day have RMS
[storage] issues.”
With the exception of Mxit, all the J2ME apps with which interviewees were involved were
used in programmes that gave out phones to users, with phone model support defined by the
developers. Even Mxit, a J2ME app with motivated users, limited functional requirements and a large
development budget, reported problems with J2ME app support in feature phone operating systems:
“In the featurephone space from behavioral point of view, first you have to get some to
switch their data and figure out how to download Mxit. Often it ends up in bizarre places
on your phone so then people have got to find Mxit on their phone. It's not really a unique
thing to Mxit is primarily a featured phone thing, like the operating systems are all over
bizarre.”
Even Nokia, whose feature phones were identified as having the best support for J2ME apps,
didn't seem to be able to manage the rapid evolution of their own platform. Here, and interviewee who
worked on a research project using an app developed by Nokia's research division describes a bespoke
graphical user interface (GUI) toolkit developed seemingly in isolation despite several generations of
publicly available Nokia GUI toolkits (the most recent one being released for the Nokia ASHA range
of featurephones).
“I've got the actual code, they coded everything themselves. They didn't use any ASHA
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 147
built in any things. All the things these people made themselves, everything these particular
people made themselves. I think some of the comments on the codes are I think started as
far back as 2002 on this project. So they probably started on a way different model than
I've ported to get to this phone. But it does look like, yeah. That things have evolved. And
they just kept evolving their own code rather than going to like, the existing stuff. Or stuff
that's already built in.”
Interviewees' experiences building m4d apps on J2ME align with the findings in section 4.2.6,
which discussed the origin of J2ME and identified shortcomings arising form its design and
management. Problems with app distribution, device fragmentation, and standardization identified in
interviews reflect commercial priorities that demanded rapid hardware and software evolution at the
expense of third-party app support. Smartphone app platforms have different stakeholders and
priorities, and a different set of affordances as a result. They are far more favourable to third party
developers, and interviewees building J2ME apps were uniformly enthusiastic about increasing
Android penetration. For projects that give out phones – a model mostly found in mobile data
collection – the importance of feature phone support is decreasing. In Chapter 7: Findings from
CommCare metadata analysis, a quantitative comparison between Nokia an Android users of Dimagi's
CommCare mobile data collections / remote worker system takes a deeper look at differences between
these two.
At this point it is useful to revisit two ideas from the start of this section. Findings from the
interviews suggest that while the ideal model of an m4d intervention from a development perspective
is technology agnostic, most projects start with a particular technical solution in mind. Even where
services are offered on multiple platforms, tradeoffs for each platform combine produce a limited set
off possibilities for m4d apps and services in general. The intersection of this set of possibilities with a
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 148
set of development interventions – behaviour change communication, community health work and
remote worker models, provision of educational content – defines the problem space for m4d apps and
services.
6.2 Projects
This section documents interviewees' descriptions of the structure and trajectory of the m4d
apps and services. I use the word 'projects' to denote the focal point of activities, partnerships and
reflections. Although these are presented separately from the technical aspects discussed in the
previous section, they are intertwined at multiple levels. This is particularly evident in the discussion of
challenges experienced building m4d apps and services, and in the many comments on operator
partnerships.
6.2.1 Process
As discussed above, most interviewees' accounts of how projects came about involved an idea
about using mobile technology to address a particular development problem. Project initiation often
involved developing a working proof-of-concept for the app or service. With the exception of one
funder-initiated project, funders were not generally involved at this stage. Two interviewees reported
covering initial costs themselves from savings, three had research funding, and four were part of
projects that arose within or in association with commercial IT organizations. Prototyping was used,
often simultaneously, to explore the problem space and to mobilize resources for further work. One
interviewee justified the process as “we learn through trial and error”. While this might seem obvious,
it can also be difficult to reconcile with rigid agreements and timelines that come with development
funding. The same interviewee, who did not have mobile-specific expertise, was relying on software
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 149
contractors, and did not prototype a mobile solution, identified difficulties meeting obligations to
funders as a major source of stress in the project.
In a related issue, several interviewees described a trajectory that departed quite significantly
from the initial idea. One project shifted from expecting users to pay for information through premium
SMS to targeting government departments tasked with improving conditions in the sector. A similar
shift occurred in a project previously concerned with crowdsourcing infrastructure fault information
from communities, which switched to supporting local government in delivering infrastructure services
instead. A project that started helping jobseekers to develop CVs became a job search portal. Another
project, having initially taken on extensive custom work for clients in the health sector, moved to a
platform-based software-as-service approach. Financial sustainability, user feedback, and changing
partnerships were among the reasons discussed for these shifts.
Several interviewees had been involved in projects that were no longer actively developed or
maintained, having failed to obtain funding or become self-sustaining. Content on mobi sites and Mxit
remained available but was not updated. Web hosting for these projects were prepaid for several years
out of project funding, or hosting was shared with active projects. One project using SMS shut down
completely, highlighting the difference in cost structure. Although it was difficult to obtain reliable
information on project status during the inventory exercise, the same pattern seemed to hold, with
many dormant Mxit apps and mobi sites.
6.2.2 Users and non-users
Across both the inventory and the interviews, the main intended users of m4d apps and services
were youth. Some interviewees explicitly designed projects to take advantage of mobile penetration
among youth. This included designing content with a youth audience in mind, or choosing a platform,
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 150
such as Mxit or operator mobile web portals, where the demographic profile of users was known to
skew young.
“And we know that this, that the target audience or the users that's using this platform is
youth, you know. It's people aged 18 to, let's say, 40. They spend hours on these platforms
because it doesn't cost them any money. So, why can't we put something useful and around
sexual and reproductive health and around HIV on these platforms that would engage
users? So, that's how the idea came about.”
An interviewee whose project had developed a USSD interface and mobi site found that
residents of rural areas had been slower to adopt the technology, but were increasingly using feature
phones to access the mobile web version of the site. The USSD version was not well used. The
interviewee suggests that it had been superseded by the mobile web version, although noting that older
people were the original target audience for USSD.
“At that stage as well a lot of people didn’t have phones, certainly not as many as have
now, and they were much older phones as well, so what you could do with them was much
more limited. Whereas now with our current user base, if I look at our analytics the bulk of
them are using smart phones, even the poorest of the poor kids are getting hold of smart
phones, or at least feature phones but the fancier ones, but there’s nobody using – we built
a USSD portal with those very untechnological minded people, especially elderly people
who are only really used to working with USSD, we built them in mind and that’s not
really something that’s been hugely taken up.”
Another interviewee described difficulties experienced by older people contacting a call centre
for physical infrastructure fault reporting.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 151
“An older generation don't understand the concept of a call centers. If you phone in and ask
to speak to Thando and you say Thando I don't have water or the water is running that's
Thando. So Thando then says to you do you want your reference number? And now the
lady doesn't know what the reference number is, so she wouldn't know what that mean. So
she just says no that fine sweetheart I will phone again tomorrow. Tomorrow phones again
and want to speak to Thando now Thando, I don't know who Thando is because Thando
might be sitting in a different call center in another part of the world. Then say where your
reference number is? What is a reference number? You know this notion of how do
different population groups engage.”
In the face of access and use challenges, however, two interviewees who had been involved in
income generation and employment projects suggested that motivation and relevance might be the
most critical factor in determining adoption.
“When you talk about bottom of the pyramid, people who don't speak English, the first
thing you hear from people is how would they use? Will they know how to navigate all
these thing? But when you're telling somebody if you go to this place, you're going to find
a job, they will find a way of getting there.”
“I suppose the lesson that could be extracted from that is that people will learn what needs
to be learned if there's sufficient incentive to learn it. So if the technology you are giving
them is going to help them make money, help them find a job or that's going to help them
get their job done better and make more money. They're quite happy to learn and in fact
they're excited.”
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 152
The above quotes suggest that the predominance of youth-focused m4d apps and services
should be understood in terms of both drivers and barriers. Interviewees attributed non-use among
older persons to less access and less experience with mobile technologies. However, no apps or
services in the inventory targeted older persons specifically, and needs analyses covering this group
specifically seem scarce. By contrast, apps and services on Mxit, with low data costs, as well as free-
to-user operator portals, benefit from a large, content-hungry audience of youth, for whom there are
few alternatives. Where user numbers are reported as a measure of success, this should be understood
in the context of a heterogeneous user profile.
6.2.3 Funding
Interviewees uniformly confirmed the inventory findings that development funding was and
would remain the primary funding source for m4d apps and services. In particular, user-pays models
were rejected – an important finding of this research given the discursive linking of mobile apps and
services with development on the one hand, and the 'fortune at the bottom of the pyramid' on the other.
As demonstrated in the quotes below, user-pays isn't impossible – people, including people with very
little income, pay for mobile communications and may even pay for entertainment services – but for
m4d apps and services it was seen as unlikely.
One interviewee pointed to high mobile communication costs as a barrier to paying for other
services, while another suggested that given the choice, entertainment content simply won out.
“You know in urban areas, poor people, they want what rich people want on their phones.
So they want to do games, and they want to download music and you know, and if you
have a health app, that’s free or its got a nice easy icon or you know, people will use it but
people are not going to rush out and buy phones to get health apps or voting information or
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 153
any of those things. So but you know what we did find especially with Mxit [..] was that
people were using those free apps quite extensively but they’re paying for the most
extraordinary things like this flush or whatever its called [..] paying, you know, not large
amounts of money but paying to do it, and paying for biblical quotes, happily.”
“I think there is a host of services that can be delivered at the bottom of the pyramid via
mobiles. I think that is true. I think they will be much more successful services if they were
cheaper, if access were cheaper, you'd see a huge growth in uptake if they were really
really affordable and people didn't have to wonder whether it was worth spending the next
two rand on that. ”
In addition, revenue sharing sees operators and intermediaries such as Mxit and SMS
aggregators take the lion's share of the amount charged. Here an interviewee describes sustainability
challenges in e-Book sales.
“it's kind of low-cost, high-volume model, but the volume just isn't there [..] and the other
thing is, I mean even if people do buy, Mxit will take 50 percent or 70 percent or
whatever.”
Although user-pays was judged infeasible, not all projects were directly grant-funded.
Interviewees involved in organizations with a software-as-service model – primarily mobile data
collection systems – reported some government clients, but also relied on development funding
indirectly through grants to clients. By maintaining a common core system and charging for
customization and support, the cost and risk of technology development could be spread among
multiple clients. This model was common among specialist m4d organizations, many of whom started
doing custom projects but moved to focus on developing a customizable platform for particular kinds
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 154
of m4d apps and services. Of these, mobile data collection was most common. Systems that allowed
clients to publish content such that it could be accessed on multiple mobile platforms – SMS, USSD,
mobile web, Mxit - were also mentioned.
Taking this one step further, one interviewee described a platform-as-service model, in which
their customizable m-health system would be sold to operator partners who would use it to develop an
m-health software-as-service offering. At the time of the interview this was mostly theoretical, with no
actual sales to operators. Other interviewees confirmed that operators appeared to be showing interest
in m4d apps and services, however ultimately the source of funding – some government work,
development grants whether to clients or direct – seems unlikely to change.
6.2.4 Operator relationships
Nearly every interviewee working on an m4d app or service had either partnered with or
attempted to partner with an operator. Operator partnerships were pursued either as funding source
(largely unsuccessful) or to obtain zero-rated access to services (mixed success).
“You need MNOs [mobile network operators] to zero-rate services. That's always the
number one reason to talk to an MNO. In practice that seldom happens.”
In the quotes below, an interviewee highlights the role of personal relationships in obtaining
access to operator partnerships, and others describe long-running negotiation processes that have yet to
bear fruit.
“Obviously, we wanted to partner with one of the network operators [..] and we started off
with MTN because we had a good relationship [..] Unfortunately, just before we could sign
off the deal, this lady resigned and she left and we couldn't get another foot in the door
with MTN. So, we gave up and we went to Vodacom.”
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 155
“We have [approached operators], and we really haven’t had a lot of luck which is, it was
we haven’t in the past say, past year and a half, prior to that we were chasing MTN,
Vodacom Cell C, all of them constantly, and I think at the time they didn’t really see, they
didn’t understand what we were trying to do.”
“You know, they're all... All the different stages of being interested or involved. MTN,
we've been in discussion with them for many, many, many years. And about this, they're
very keen. [..] Cell C, we've had interesting meetings with them but they have other
priorities on their plate at the moment. And so, similar to Telkom. So, we work with
Telkom on our for-profit side of our business. And MTN, we've also been working for a
number of years. So, we do have relationships with all of them. It's just, you know, getting
the timing right so that your priorities line up with their priorities. And it's sometimes
tricky.”
There was also one instance of an operator with shares in, and later full ownership of, an m-
health company. Here access to operator infrastructure (the 'stack') was presented as a competitive
advantage, with the interviewee noting the ability to zero-rate. In-country hosting for applications with
legal requirements around the storage of health data was also mentioned.
Interviewees also commented on the role of operators in relation to m4d apps and services in
general. Operator interest in m4d apps and services – as evidenced by the Vodacom ownership example
above – was perceived as arising from a desire to “own more of the value chain.” For some this was
simply evidence of the success of the model, while others were concerned about the implications of
operator control for innovation in apps and services generally. As the quotes below demonstrate, the
issue is both technical and strategic.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 156
“In some places it's absolutely pernicious. In Kenya, the extent to which Safaricom
controls, as a vertical market, mobile apps, it's unhealthy. I haven't seen signs of that
improving particularly, which is why you see so many people trying to work around it.”
“What gives them the competitive advantage in these circumstances is that it is all locked
down and they do their own things and if you come in you come in on their terms and I
mean the few people we spoke to said as much. That you know, basically it was selling
their souls. ”
“Some of the MNOs in Africa are really hard to deal with because technically they are
years and years behind South Africa. To set up a USSD code takes a day. Nigeria and
Ghana - USSD can take 2 years because they've never had to deal with it outside their own
system.”
“I was a bit blown away for the SMS project the Zambia results one, we were actually able
to get a telco to add new functionality for us which was a reverse billed short code. A short
code people could use for free and our project would be charged for. I think the
complicating factor was when it crossed networks. And it took several weeks, maybe
months even, but they actually did it and made it happen.”
The GSMA, which during the study period was increasingly focused on promoting m4d apps
and services, was criticised for misrepresenting their intentions and membership. Interviewees, as well
as participants at the ICTD conference panel discussion hosted by the GSMA, wanted GSMA operator
members to be encouraged to open up their networks to third parties. The GSMA's claiming to
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 157
represent a “mix of philanthropic and commercial stakeholders” was perceived as a veiling of
intentions.
“You need to lift the veil on what you actually want - it's not better health, it's more
subscribers. Design the programme that speaks to those true business interests. If you
pretend otherwise I would say it's your fault not theirs.”
“I think the GSMA are doing exactly what they are supposed to be doing, which is to be a
voice for their constituency, which are the mobile network operators. I think where they irk
me is when they pretend that they're speaking with a neutral voice. And this is sort of a
mobile narrative in general - "we are the face of development". You're not the face of
development you're the face of your shareholders. And it's kind of self-evident. I don't
mind them doing what they're doing, but I mind them pretending that it's something else.”
“But their jobs are – just never forget what their jobs are. If you don’t forget what their job
is they are doing their job very well.”
Asked to consider whether policy interventions, whether in conjunction with or outside of the
GSMA, might work towards opening up the mobile ecosystem to third parties, the interviewee quoted
below points out that unlike in the developed world, where mobile is a “complimentary” service,
developing world mobile networks are the only access option for many people.
“You know over of the world particularly in the North because mobile is so much a
complimentary service [..] I think Africa’s really leading a lot of that stuff in terms of
acknowledging that you know, getting access to the fixed network if you’re corporate or if
you’re a company supplying corporates but for ordinary people, you know, getting that
wholesale access network, getting access to platforms and things like is actually quite
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 158
critical.”
In this context, resentment towards operators is understandable. For several interviewees, the
move towards a smartphone ecosystem, in which operators have considerably less power, was
identified as a positive development and the most likely solution to their own access problems.
Discussion of the smartphone ecosystem continues in 6.3.3.
6.2.5 Government
Interviewees represented partnerships with government as both desirable and fraught.
Integration into government services in areas such as health and education was seen as a way to reduce
dependence on donor funding. However, attempts to work with government had also been frustrated by
staff turnover, lack of clarity in policy implementation, short timelines and communication difficulties.
“Every time that you're trying to get a contact in South Africa, the government, we maybe
got one response but after that, it was they would completely fall off the face of the Earth.
So, there was never any clarity about what frameworks we could fit into within South
Africa [..] Once we got leads on things and it would just fall apart. Nobody was really able
to follow through on it.”
“We didn't go through schools because it would have taken too long [..] We needed an
answer quickly.”
Most projects that reported working with government did so with grants from development
funders rather than government funding. One interviewee pointed to evidence of a transition:
“That’s happening in conjunction with Department of Health. The Northwest [province]
was quite unique about that is that the department is actually paying but for the services,
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 159
instead of a donor which is kind of the transition that we're trying to do justify. Started off
donor funded mainly the MRC and HST had funding and now the government's take note
of maintenance in running all the system. How long that will last we don't know but we
live in hope.”
In the health sector, government had begun to take on a co-ordinating role, although this was
described a “industry-driven.” South Africa is home to a number of organizations whose main focus is
m-health or e-health, including one with a specific focus on national-level architectures and
interoperability. A government-lead task team on standards for e-health, including members of these
organizations as well as the GSMA, has been involved in developing a high-level e-health strategy.
There are future plans to mandate technical standards for interoperability.
Moves to formalize m-health as part of a government-supported national architecture are
indicative of the relatively maturity of the sector and the strength of the evidence base. On the one
hand, government involvement in m4d projects is important for sustainability, legitimacy and avoiding
duplication of effort. On the other, with the possible exception of m-health, evidence of impact, and by
extension cost-benefit, is scarce. In the face of competing priorities, it is perhaps understandable that
government has not taken a more active role.
6.2.6 Evaluation and impact
As with the inventory apps and services, interviewees who claimed their projects were
successful often did so by citing user numbers. Additional evidence was sometimes presented,
including engagement measures based on tracking visits on web-based platforms. Interviewees cited
the number of return visitors to a mobi site or Mxit app, the number of users who read an entire story,
and users who completed an entire set of lessons. Several interviewees described the ability to track
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 160
user activity as a very positive attribute of the platform. In the quote below, an interviewee with a
J2ME mobile data project notes the potential for user tracking but also acknowledges that further work
is needed to effectively analyse user data in terms of performance metrics.
“Tracking and monitoring of how the system is actually used [is] kind of the hardest part of
the feedback loop. We did try and provide a lot of base logging. Not like spying but just
like you know how long was this thing open. Try things that would kind of let you
reconstruct how things are actually being used. And a lot of that was to kind of try to figure
out bugs because we would get bug reports and we would have no idea of what’s going on.
So there was a huge trove of data but then we would have to put the data into useful
metrics.”
One interviewee, with close ties to an academic institution, had tried to go beyond user
numbers and conduct an impact evaluation. They did this by surveying users of the app - a job search
site - about whether that had been contacted through the site and whether the site had helped them get a
job. The interviewee stated that 21% of respondents had got a job through the site. Despite
encouraging findings, the method – posting a question on the site and counting responses – limits the
strength of the evidence. Perhaps for this reason, it has not been published beyond a single mention in
a doctoral dissertation.
An interviewee working in academia brought up general difficulties communicating findings
about m4d apps and services, particularly in relation to academic publishing. Considering evaluation
in the general sense – lessons learnt being cited here rather than evidence of impact – the interviewee
questions whether journal publications would be read by people building m4d apps and services, and
wonders how to reconcile context-specific findings with the research quality markers of journal
publication.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 161
“The challenge with this is where to publish it and how to publish it. With lessons learnt it
is always a difficult one. You can publish it easily in a conference, but to get it into journals
you have to either have a research methodology that analyses how you learnt your lessons,
or you have to have some statistical analysis, or you have to have something, and so I think
the journals that lend themselves to publishing that wouldn't be the journals read by that
community. So I can publish a paper in Community Engagement, but who reads
Community Engagement, not people in the ICT sector [..] You see the lessons learnt are
very context-specific.”
Outside of academia, knowledge sharing was perceived to be hampered by commercial
priorities. One interviewee recounted how Nokia, whose for many years had an Africa-based research
group working on m4d issues, had run extensive studies on apps they had developed but published
nothing. As a counterexample, however, the GSMA's involvement in m4d, which was co-ordinated
under the banner of Mobiles for Development Intelligence during ICTD 2013, is an attempt to co-
ordinate knowledge-sharing in the sector. By 2015, the initiative had been renamed Mobiles for
Development Impact. The site at www.m4dimpact.com houses an extensive collection of whitepapers
and case studies written by or prepared for the GSMA. Interviewee criticisms about the GSMA's veiled
intentions urge caution when dealing with the data and analysis they produce. Nevertheless, the
initiative represents an acknowledgement of the need for knowledge sharing in m4d, which is shared
by researchers working in academia.
6.2.7 Sharing and re-use
Duplication of effort is a common theme in m4d, particularly in the health sector. Interviewees
were generally familiar with each other's work, and that of other organizations in the sector.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 162
Conferences and mailing lists supported knowledge sharing around technology development and
implementation practices. Several interviewees also mentioned attempts to share or re-use code,
employing a variety of models.
“It’s almost the moral obligation to reuse tech as much as possible. So you know there is a
lot of developers who will licensed their tech and they’re trying - and obviously, they are in
it to try to make money as an NGO or a PBO as we are Public Benefit Organization, our
philosophy is to try and leverage tech work we’ve already done. So if someone gave us
funding to do something, we’ll always look at the next iteration being used for something
else and then all we had really - there will only be a tiny amount of development work to
just tweak it a bit for that audience. Yeah. So it becomes very cost effective way of
delivering this thing.”
Maintaining an internal shared code base, which also satisfies commercial priorities, was the
most commonly implemented model. One interviewee described several other initiatives, including
sharing code through open source licenses. Interestingly, funder preferences are cited as a reason for
choosing an open source license, and platform limitations cited as a reason for the project's failure to
attract outside contributors..
“It was already open sourced by the time I got there. I think it was probably partly dictated
by the grants. Like we did get grants that may have dictated that the code be freely
available or at least [organization] does not retain the right to the copyright of the code.
Then there was other grants more open ended. They were like “make this thing better”.
And you can’t do that to a proprietary product. It has to be an open product. So as far as
being open sourced it kind of made an interesting showcase. African devs that wanted to
get their hands dirty with a real thing could play with it. But it never really was an open
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 163
source project in a sense of like a broad community contributing things. Instead people
who were paid to be familiar with the code were the core contributors. Possibly sole
contributors. Partly because there was a big learning curve to the code. Part of that was
dictated by J2ME being so awful.”
The other initiative mentioned by this interviewee (and with which I was also involved) aimed
to encourage standardization and co-ordination between groups working in the mobile data collection
through collaborative development of shared components. The quote below describes difficulties
establishing collaborative relationships among organizations that, despite working in development and
often on a non-profit basis, operate as part of a competitive field.
“It was nice to kind of meet with everyone to know what was going on. But as a productive
effort it really didn’t seem that worthwhile. Like everyone was doing similar things but no
one wanted to compromise enough to actually share everything. So they always wanted to
like - they always had some customization they wanted to do and that customization was
probably dictated by some external client or contract. And they couldn’t compromise on it.
So there was a lot of like we will have a big meeting and everyone come together and
everyone would drift off in their own direction until the next meeting and you know but
there was no big effort to come back together. So it was like very forced. Otherwise people
just drifted aimlessly. ”
The ideas above suggest that reducing duplication of effort through shared code might benefit
from looking beyond technical efforts to intervene at organization level. In particular, co-ordination by
funders, rather than grantees, seems worth pursuing (if challenging). The South African national e-
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 164
health strategy mentioned previously sees government taking a standards-based approach to co-
ordination, which may turn out to be the most appropriate level for government intervention.
6.3 General observations on m4d
To conclude the interview findings, this section considers responses of general relevance to the
m4d apps and services landscape. This includes comments on the concentration of m4d apps and
services in the health sector and consideration of the role of commercial stakeholders in constituting
m4d as a field. The final subsection considers the future of m4d apps and services in the context of
increasing smartphone penetration.
6.3.1 Sector concentration and lack of diversity
Interview responses about sector concentration of m4d apps and services on the one hand, and
user preferences on the other, identify a situation that is at least partially mismatched. While both
interviews and inventory demonstrated that the largest concentration, best evidence and greatest
diversity of archetypes for m4d apps and services is in the health sector, users were seen as being most
strongly interested in access to employment.
“It's clear to me that what I observed from the uptake of Mxit that [users at the bottom of
the pyramid] are looking for ways to communicate with one another. They're looking for
entertainment. I mean those obvious things. We really did see that they were keen to look
for ways of connecting to employment or ways of connecting to revenue.”
Interviewees suggested several reasons for concentration in the health sector. Below, one
interviewee describes how the construction of healthcare as a societal responsibility rather than an
individual one influences expectations for sustainability and funding. Another distinguishes between
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 165
apps and services intended for the general public and those that are implemented within an
organizational context, with the latter particularly strong in the health sector.
“ I think that when you're doing health, I think that there is an expectation that you're going
to provide a public service and you don't need to be profitable because everybody
understands the conflict of interest, the profit health and public services of--but when
people talk about agriculture, they see it as a private enterprise. They don't see it as like--
well, agriculture is actually a public good because, number 1, food security, number 2,
employment. But, people see it as a business. And if you're talking about a business, then
people are talking about profit, so it's very hard to get through to different people and talk
about this is actually a public good like that you're providing Africa.”
“There is a difference between apps that are deployed that require take up by the certain
general public or lay public verses specialize things that help, specialized people working
in the development fields to help them do their jobs better. There’s a ton of support and
momentum behind the latter category right now, particularly in health. There is almost too
many apps that help community health workers do their job better and you have no
problem demonstrating these things are successful [..] I think the distinction remains
important and I think that gets lost sometimes I think the general and hype literature about
the stuff.”
Another interviewee addressed the issue of lack of diversity in m4d apps and services in terms
of problems to design against.
“We still need more intuition and best practices about how we frame problems that we can
actually design against. And then what are the tools we can bring against those. We are all
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 166
sort of pushing on it but by no means got it figured out.”
The relatively long history of ICTs in healthcare, including e-health and telemedicine in the
developing world, provides a more stable set of “problems to design against”. Monitoring and
reporting for community health workers is perhaps the most established problem, with software-as-
service providing a relatively well established business model. During the GSMA session at ICTD,
participants working with operators also identified mobile money as having gained favour among
operators based on a strong business model. The result is clustering of m4d apps and services around a
relatively narrow range of archetypes, as demonstrated in the inventory research.
“You see the same kinds of projects, multiple different phases, multiple different people
doing it. Once the model works, like the community health workers, a lot of people are
doing it .”
Some interviewees mentioned innovations in apps and services that might address access to
employment. In one project, in which remote workers monitored a school feeding scheme using a data
collection app, microwork tasks were set to be added to the app to provide extra income for workers. In
another, a site with a job search component found users tended to be most interested in learnerships,
which are paid apprenticeships managed by government-supported industry bodies, and recommended
focusing on learnerships over jobs for school leavers without other qualifications. In a conversation
outside the interview process, an academic with extensive experience in the field was also relatively
unimpressed by the potential of mobile apps for job search, noting that “people want a job, they don't
want a social network where they can talk about getting a job” (G. Marsden, personal communication,
8 December 2013).
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 167
Others took a broader view and considered regulatory strategies to encourage access and
innovation. Here, one interviewee notes that Facebook is used by microentrepreneurs but generally not
considered an m4d app. Another suggests that development goals should be secondary to affordability
and openness at platform level.
“What constitutes an m4d app or service is really hard. I’m very interested in how small
businesses use Facebook [..] it’s not an m4d app. It’s just a functionality that like every
functionality when it makes sense these people will find it and they’re putting up little you
know Facebook pages for their bar or their clothing shop or something like that.”
“Perhaps rather than money going into apps development, specifically, what it should be
focused on is supporting [..] research but research and policy development in order to
create these environments where things are affordable and ubiquitous etc., and then
supporting developers, communities etc. [..] I think that’s more likely to create a kind of
digital inclusion then sort of very narrowly like you do this development thing.”
6.3.2 Stakeholders in defining m4d
In section 6.2.4, the GSMA's role in m4d was question by several interviewees, who saw
attempts to claim a dual development and commercial purpose as misrepresenting the profit motive of
the body's mobile operator shareholders. The additional point made by the interviewees quoted in this
section is that role of the GSMA, as well as multinational telecommunications companies (Vodaphone)
and equipment manufacturers (Nokia), includes participating in the definition of m4d as a field.
Aligning mobile apps and services with development is a discursive strategy with material
implications, as the interviewee below points out in relation to spectrum licenses and regulatory
treatment of mobile networks.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 168
“It’s absurd right now how the discourse serves a number of interest very well [..] there’s
meaning making going on that sets up how much licenses are worth and, again the GSMA.
It’s a bit problematic the way the GSMA is so prominent as arbiters of what constitutes
telephony.”
Later, the same interviewee notes that an association with development also benefits public
perception of operators and other mobile industry players. Another interviewee observes that mobile
operators in South Africa use the positive image of m-health to excuse closed platforms and
proprietary software.
“they get a lot of goodwill generated from every picture of a resource-constrained person
using your technologies. It's good for shareholder value. ”
“I mean a lot of their stuff when they spoke to me they were like, oh yes we are proprietary
about this but its actually because it’s a big social responsibility project around health or
whatever it was so we're not really that bad.”
With most interviewees involved in building m4d apps and services rather than producing
general analyses of the field, concerns about operator involvement were more widely discussed than
the ideas above. However, making the link is important particularly in light of general consideration of
technology solutions to development problems. While the players change in the smartphone era and
some of the practical aspects of interest to developers of m4d apps and services improve, definitional
issues remain.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 169
6.3.3 Smartphones
Increasing smartphone penetration, particularly among youth, was perceived as largely positive
by interviewees. Smartphone platforms – particularly Android – were already beginning to replace
J2ME for apps and services whose model involves providing phones to users as part of implementation
within an organization.
“I mean [organization] also tried to have this philosophy of target any phone, but then
realised that any project worth its salt will be buying - I mean it’s like a fleet management
kind of style of equipment so phones will be purchased like it’s one of the lowest items of
the budget. So basically if like a good Android is available for at the right price point just
choose that.”
In addition to being designed for third party access to a full range of hardware functions, the
Android app environment is commercially relevant beyond the developing world, and actively
managed as a result. One interviewee pointed out that Android apps were also likely to become
increasingly relevant to education, not necessarily for smartphones but in the context of school tablet
rollouts.
Interviewees whose projects targeted a general audience were hesitant to assume smartphone
ownership, preferring the flexibility of featurephone-accessible mobile web sites. While
acknowledging the existence “trickle-down” smartphone ownership, the quote below notes that a lack
of durability may limit access to cheap smartphones via the second-hand market.
“ I think the actual smartphones is getting out to rural areas in Africa [..] I think that's how
these 5, 10 years down the line before you start seeing a lot of that going there. And
smartphones, they're just not as durable. We don't know if they're going to be able to have a
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 170
third generation cellphone be sold to some guy out on the bush when the whole Nokia as in
the feature phones, those things can't break. You can resell them five or six times, but I just
don't know if the smartphones are going to be able to have that long of a lifetime.”
In addition to providing an open platform designed to encourage third party apps, the
smartphone ecosystem differs from the non-smartphone ecosystem in terms of stakeholders. Where
mobile operators (with close ties to hardware companies) previously controlled both the
communication network and platforms for apps and services, Internet companies such as Google have
intervened as the main provider of an app platform. Interviewees were generally positive about the
implications for m4d apps and services. To some extent this reflects strong feelings about the current
status quo rather than a particular enthusiasm for its evolution. Nevertheless, in the quotes below
interviewees note both structural and discursive changes accompanying the move towards the
smartphone ecosystem.
“I think it's clear that the smartphone is achieving layer separation for mobile phones, and
that's huge. It's huge because, you only have to look at something like WhatsApp to realize
that ok, we have broken SMS now. And that's massive. And I think that that will only
continue. I think that it's still an issue in terms of, you're very limited in choice in terms of
who you get your access from, but yeah, the revolution is arriving and it's a very good
one.”
“Yeah, yeah, well you know at one level it’s sort of just you know, replacing one
monopolist with another. But obviously then their approaches and attitudes are quite
different [..] so I do think that you Google controlling certain things is different too to a
national operator, mainly because they do have these claims, their own claims and claims
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 171
made upon them around you know, transparency and accountability and things like that [..]
So, you know I’m not naïve about what they are. But I do think that at least they’ve kind of
come under that pressure and whatever. At least they you know, feel the need to squirm.”
“The direct response to your question about what’s a lesson to learn from the feature phone
world as we enter the smart phone world is - I will take Vodaphone as an example. The
most prominent example of a technology firm accidentally being cast as a development
agency and deciding to wear that coat. I think that Google and Facebook, in particular,
Microsoft too, and everybody else too, but particularly those two for different self-serving
narrative reasons also find themselves in a place where there is this alignment between
their rhetoric around how the world works and the development narrative that is not
unproblematic and is worth very significant unpacking.”
This last quote is interesting because, despite identifying new players in the smartphone
ecosystem, the interviewee indicates a general need for critical assessment of the association of
technology companies with development. This is relevant to designed-in limitations of platforms for
m4d apps and services, and related criticism presented in this chapter of the role of operators in the
pre-smartphone mobile ecosystem. The conclusion takes up the idea of 'inter-generational' ICT4D
continuities in more detail.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 172
CHAPTER 7: Findings from CommCare metadata analysis
Following on from issues raised in the interview and inventory findings, this chapter takes up
data-driven evaluation of m4d projects, and particularly the idea that metadata from m4d projects –
such as user numbers, web analytics measures and software logs – can be used to make claims about
both user behaviour and intervention success. To explore how data-driven evaluation might work, I
conducted a statistical analysis of operational metadata generated by users of CommCare, a mobile
data collection and case management system for remote workers in the developing world. The
discussion covers findings from the analysis as well as from the process itself, from which I draw
implications for data-driven research in m4d.
CommCare is developed by Dimagi and available for Android and J2ME through a software-
as-service model, which may or may not include paid implementation support. The quantitative
analysis considers the effect of these two project design variables – choice of technology platform
(Android or J2ME on Nokia featurephones, referred to in this chapter as device type) and project
support structure (whether self-started or supported by Dimagi) – on user productivity and user
retention.
First, given the various limitations of J2ME as a platform, it is hypothesized that users of the
Android app will be both more productive and more likely to be retained over time than users of the
J2ME app, which runs on various Nokia featurephones. Second, users in projects with implementation
support from Dimagi are hypothesized to be more productive and more likely to be retained than users
in 'self-started' projects without such support. The data makes it possible to test these because
CommCare has both Android and Nokia (J2ME) users in relatively large numbers, as well as a
combination of projects that receive implementation support and those that do not.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 173
A full description of the data and methods, including the data acquisition and cleaning process,
hypotheses and justification for choosing mixed effects models, is presented in the sections that follow,
followed by the findings and discussion.
7.1 Data and methods
7.1.1 Data
To introduce the dataset, a basic description of CommCare functionality is necessary. In
common with other mobile data collection applications, it consists of a mobile app (Android or J2ME)
that communicates with a database server accessible to users through a web interface. Data is collected
by remote workers using mobile phones, and transmitted over the mobile network. The web interface
provides various reporting, monitoring and management functions. Both the mobile app and the server
code are open source, but nearly all CommCare projects use the hosted service provided by Dimagi,
which is free for up to 50 users and relatively low-cost thereafter. Other paid services offered by
Dimagi include implementation support, configuration, training and feature development, with
packages of these available in the form of service contracts.
The dataset contains metadata from all hosted CommCare projects. From a conceptual
perspective, metadata is generated when users (mobile workers, although the web interface can also be
used) fill out data collection forms, which may pertain to cases. For example, a community health
worker (user) fills out an antenatal risk assessment (form) when they visit a pregnant woman (case).
The contents of the forms and the attributes of the cases are configurable and project-specific, such that
a farmer being visited by an agricultural extension worker or a water point being monitored for water
quality might equally be considered a case. For this analysis, only attributes common across all
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 174
projects and case types were considered. The code book (Appendix D) contains additional information
about the variables included in the analysis and used to clean the data.
Fifteen projects opted out of making their metadata available for research, and were excluded.
No sensitive or protected data items were involved, and projects and organizations are not identified by
name.
7.1.2 Data gathering
Although metadata is collected as part of normal operations, accessing the data in a form
suitable for research required several intermediate steps. A data 'pull', which takes place approximately
once a month, runs through each of these steps in turn to keep the research data up to date through the
end of the previous month.
1. CommCare provides public-facing application programming interfaces (APIs) for data access.
Data is downloaded periodically from the APIs, with incremental downloads occurring after the
initial full download. A Python extractor script downloads the data.
2. A Python loader script store the data in a database designed as a conceptual representation of a
research/process improvement perspective on the data. For example, multiple forms about the
same case, completed consecutively by a single user, are linked to a single visit in the research
database model. The visit concept makes it possible to calculate measures of mobile worker
productivity independent of how many forms are typically involved in each interaction with a
client. The concept of a visit is less relevant to an operational perspective where the focus is on
the data collected as part of each form, and CommCare itself doesn't store data on visits.
3. Aggregation by project, user and month is performed, calculating out indicators such as number
of forms, number of cases visited, time between visits etc. Because it is computationally
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 175
intensive, aggregation is performed as part of the monthly data pull rathe than ad-hoc.
Aggregated data tables are stored in the database.
At this point, data is available for cleaning and analysis, including aggregate data by month,
project and user as well as project attributes.
7.1.3 Cleaning
In addition to the conceptual translation of operational to research data, further cleaning was
necessary to obtain a reliable and consistent dataset. Data cleaning is a well-documented challenge in
using secondary data for research. Bailey et al (2012, p.80) identify five challenges: “missing data,
erroneous data, uninterpretable data, inconsistencies among providers and over time, and data stored in
noncoded text notes”. With the exception of non-coded text, all are relevant to CommCare metadata.
Additional challenges are introduced by user-provided data, since anyone can create a CommCare
project.
The data cleaning steps were as follows:
1. During extraction and loading to the database, forms with missing or obviously erroneous start
and end times – the start and end time being determined by the time set on the phone, and
subject to misconfiguration – were adjusted to the time the form was received at the server.
Forms with erroneous start and end times were identified based on a deviation of more than 60
days between the form end time and the time it was received at the server.
2. Also during extraction and loading, forms for which no case could be found to match the case
identifier in the form were excluded.
3. Projects flagged as test projects (a flag set most often by Dimagi staff) were excluded.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 176
4. All data for web users (who fill our forms on the CommCare website rather than a mobile
phone, and as such do not correspond to a remote worker model – in most cases these are test
users) were excluded, as well as data for test users, identified by username.
5. Missing values on the variables of interest caused data to be excluded when testing the various
hypotheses. The data used to test each hypothesis is described further in Table 5 on page 180.
7.1.4 Hypotheses
The hypotheses tested using CommCare metadata are based on two main ideas. First, given the
various limitations of J2ME as a platform (discussed in 4.2.6), it is hypothesized that users of the
Android app will be both more productive and more likely to be retained over time than users of the
J2ME app, nearly all of whom use various Nokia featurephones. Second, in interviews and in
assembling the inventory, it was suggested that rather than technology always directly
enabling/'empowering' previously marginalized users, the most established archetype of m4d apps and
services – mobile data collection and remote work, often in the health sector - operates within an
organizational context. These apps and services are generally the core product of a specialist m4d
technology company, with implementing organizations accessing the technology, frequently
accompanied by paid implementation support, through a software-as-service model. The role of
implementation support is explored in response to the suggestion that the functions of specialist m4d
technology companies extend beyond technology development.
While it is not possible to measure outcome measures directly, retention and productivity
measures are considered reasonable choices as dependent variables. Specifically, productivity is
operationalized as the number of visits a user completes in a month, remembering that a visit can
include filling out multiple forms about a case or a set of related cases. Retention is operationalized as
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 177
short-term retention - users submitting data in their second month on CommCare who are still
submitting data three months later i.e. in their fifth month - and long-term retention - users submitting
data in their sixth month on CommCare who are still submitting data six months later i.e. in their
twelfth month.
The specific hypotheses are as follows:
H 1. Users submitting data with the Android app will complete more visits than Nokia device users
submitting data with the J2ME app, at 6, 12 and 18 months (H1.1, H1.2, H1.3 respectively).
H 2. Users in Dimagi-assisted projects complete more visits than users in self-started projects, at 6,
12 and 18 months (H2.1, H2.2, H2.3 respectively).
H 3. Users submitting data with the Android app are more likely to be short-term retained
(submitting data in both month two and month five) than users submitting data with the J2ME
app.
H 4. Users in Dimagi-assisted projects are more likely to be short-term retained (submitting data in
both month two and month five) than users in self-started projects.
H 5. Users submitting data with the Android app are more likely to be long-term retained
(submitting data in both month six and month twelve) than users submitting data with the
J2ME app.
H 6. Users in Dimagi-assisted projects are more likely to be long-term retained (submitting data in
both month six and month twelve) than users in self-started projects.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 178
7.1.5 Analysis and descriptives
A key issue in analysis is that there are sources of variance at at least two different levels – or,
put another way, each user's behaviour is not an independent observation. Retention and productivity
are measured at user level. Users are organized into projects. Sources of project-level variance include
various issues related to programme design and management, including whether the project is Dimagi-
assisted or a self-starter. Furthermore, although a project is the level at which deployments are
managed, some organizations have multiple CommCare projects. For example, one organization with
multiple project is running a randomly controlled trial (RCT) with sites in several different countries
but the same study protocol at each site.
Figure 21: Median visits per project in month six, showing projects that are part of
organizations that have multiple projects
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 179
Figure 21 shows the median number of visits in month six for projects that are part of
organizations with multiple projects, with organization membership indicated by colour. Visual
inspection suggests organization-level effects may be present as well. Organization membership is
therefore considered as an additional grouping level, with projects nested within organizations.
To accommodate nesting in the dataset, mixed effects models were estimated, with random
effects for project membership and organization membership. To assess significance, the full model,
including both self-starter and device type fixed effects as well as the two random effects, was
compared in turn to a reduced model with each of the fixed effects removed. For H1 and H2
(productivity as DV), the models for the productivity measure (nvisits) are shown in Table 4. The
analysis was repeated at 6, 12 and 18 months.
Null model nvisits ~ 1+ (1|project) + (1|organization)
Full model nvisits ~ 1+ device_type + self_started + (1|project) + (1|organization)
Device type fixed effect
only
nvisits ~ 1+ device_type + (1|project) + (1|organization)
Self-started fixed effect
only
nvisits ~ 1+ self_started + (1|project) + (1|organization)
Table 3: Mixed-effects model for productivity measure, R notation.
For H3 to H6 (retention as DV), the models are shown in Table 4. Both short-term and long-
term retention models were estimated.
Null model retained ~ 1+ (1|project) + (1|organization)
Full model retained ~ 1+ device_type + self_started + (1|project) + (1|organization)
Device type fixed
effect only
retained ~ 1+ device_type + (1|project) + (1|organization)
Self-started fixed
effect only
retained ~ 1+ self_started + (1|project) + (1|organization)
Table 4: Mixed-effects model for retention measures, R notation.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 180
Table 5 describes the cleaned data used in the analysis, organized by DV of interest for each
model. The breakdowns by IVs device type and self-started (i.e. not dimagi-assisted at project
initiation) reflect the fact that the variables are measured at different levels, with device type measured
for each user and self-started for each project.
Dependent
variable
Users Users by device type Projects Projects by self-
started
Organizations
Visits in
month 6
3066 Android =1861 (61%)
Nokia = 1205 (39%)
151 True = 50 (33%),
False =101 (67%)
130
Visits in
month 12
1814 Android =1074 (59%),
Nokia = 740 (41%)
91 True = 34 (37%),
False =57 (63%)
76
Visits in
month 18
1037 Android =582 (56%),
Nokia = 455 (44%)
51 True = 20 (39%),
False =31 (61%)
42
Short-term
retention
4309 Android =2631 (61%),
Nokia = 1681 (39%)
200 True = 78 (39%),
False =122 (61%)
172
Long-term
retention
1443 Android =811 (56%),
Nokia = 632 (44%)
77 True = 28 (36%),
False =49 (64%)
63
Table 5: Independent variables and group counts, by model
Descriptive statistics on the DV for H1 and H2, visits at each of 6, 12 and 18 months, are
difficult to present because of the multi-level structure of the data. In Figure 22, Figure 23 and Figure
24, I show the mean and standard error for the visits variable in each month, by project. Retention, the
DV for H3 – H5, is shown in Figure 25 (short-term retention) and Figure 26 (long-term retention) as
percent of users retained per project. In both sets of figures, projects from the same organization have
the same colour. Both horizontal and vertical scales have been chosen to fit the range of the data; visual
comparison from month to month should take this into account.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 181
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 182
Figure 22: Mean and standard error, visits by project, users' sixth month on CommCare
Figure 23: Mean and standard error, visits by project, users' twelfth month on CommCare
Figure 24: Mean and standard error, visits by project, users' eighteenth month on CommCare
Figure 25: Short-term retention percentage by project
Figure 26: Long-term retention percentage by project
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 183
7.2 Findings
7.2.1 Predicting user productivity (number of visits)
A set of three linear mixed-effects models were estimated relating the fixed effects self-started
and device type, as well as nested random effects for project and organization membership, to the
number of visits in the user’s sixth, twelfth and eighteenth months using CommCare. The statistical
software R 3.1.2 (R Core Team, 2014 ) and the package lme4 4.1.1 (Bates, Maechler, Bolker & Walker,
2014) were used to estimate the models. The dependent variable (number of visits) was log-
transformed to mitigate heteroskedasticity. The full model is represented in R as follows:
log_nvisits ~ self_started + device_type + (1|project) + (1|organization)
The full model was compared in turn to two reduced models with each of the fixed effects
removed, using likelihood ratio tests to determine the preferred model. Confidence intervals were
computed based on a likelihood profile, again using lme4. The model was estimated using restricted
maximum likelihood (REML), although comparison using likelihood ratios was done on models using
a maximum likelihood (ML) estimator.
A test for influential observations (in this case, influential projects) was performed using the R
package influence.ME (Nieuwenhuis, Grotenhuis & Pelzer, 2012). No projects exceeded the threshold
value of (Belsley et al, 1980, cited in Nieuwenhuis, Grotenhuis & Pelzer, 2012, p.40) for
DFBETAS, therefore no data was excluded from the model.
At months 6 and 12, the device type only model is preferred over the full model. The presence
of the self-started fixed effect does not significantly improve model fit at the p < .05 cutoff, so it is
excluded. At month 18, neither the self-started nor device type fixed effects significantly improve
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 184
model fit at the p < .05 cutoff, and the null model, with organization and project random effects only, is
preferred. Results of the likelihood ratio tests are shown in Table 6.
Month Fixed effect dropped in model
comparison
p
6 device type 5.99 0.0144*
self-started 0.22 0.5413
12 device type 4.00 0.0454*
self-started 0.04 0.8487
18 device type 3.15 0.0761
self-started 1.02 0.3131
Table 6: Results of likelihood ratio tests for fixed effects in productivity models.
Parameter estimates for the device type fixed effect at months 6 and 12 are shown in Table 7.
Values for have been exponentiated to make them easier to interpret as a ratio of the effect with ϐ
respect to the reference category (in this case, Android users).
IV Fixed effect 2.5% 97.5% t
Visits in users’ sixth
month on
CommCare
device type =
nokia
0.72 0.55 0.94 -2.43
Visits in users’
twelfth month on
CommCare
device type =
nokia
0.72 0.52 0.99 -2.09
Table 7: Parameter estimates for significant fixed effects in productivity models.
The models predict that Nokia users complete less visits than Android users at both month 6
and month 12. Predictions based on the fixed effects term only are given here; however note that these
should be interpreted as applying to users in the same project, or in projects with the same value for
both project and organization random effects. At month six, the model predicts that Nokia users
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 185
complete 28% less visits than Android users, 95% CI [45% less visits, 6% less visits]. At month
twelve, Nokia users are again predicted to complete 28% less visits than Android users, 95% CI [48%
less visits, 1% less visits].
The R package MuMIn1.12.1 (Bartoń 2014) was used to calculate the proportion of variance
explained by the fixed effects only (marginal R
2
) and by the whole model (conditional R
2
) ( Nakagawa
and Schielzeth, 2013). Results are shown in Table 8. The low marginal R
2
for months 6 and 12 (at
month 18 R
2
is zero because no fixed effects were included in the model) show that the device type
fixed effect explains only around 1% of the variance.
Model for visits at
month
Marginal R
2
Conditional R
2
6 0.0107 0.5846
12 0.0119 0.6147
18 0.000 0.6548
Table 8: Marginal and conditional R
2
for productivity models.
Figure 27, which shows visits per user in month 12 (each user is a single dot), confirms that
even within projects, user productivity varies considerable. Neither the self-started fixed effect at
project level, nor the device type fixed effect, which although it is set at user level is nearly always
chosen at project level and the same for all users in a project, is able to explain this, and further
consideration should be given to uncovering user-level variables that might predict productivity.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 186
7.2.2 Predicting user retention
Two mixed effects logistic regression models relating the self-started and device type fixed
effects, as well as random effect terms for project and organization membership, to retention were
estimated using lme4. A short-term retention model, for the probability of retention for three months
following a user’s second month on CommCare, and a long-term retention model, for the probability of
retention for six months following a user’s sixth month on CommCare, were estimated. The retention
variable was based on calendar months, so any user who was active in the month three months after
their second month, or six months after their sixth month, was considered retained.
Figure 27: Visits per user in their twelfth month on CommCare, by device type
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 187
The models were fit by maximum likelihood (ME) estimation based on Laplace approximation.
The full model was compared in turn to two reduced models with each of the fixed effects removed,
using likelihood ratio tests to determine the preferred model. Results are shown in Table 9.
For short-term retention, likelihood ratio tests lead to the full model with both device type and
self-started fixed effects being preferred. Comparing the full model with a reduced model dropping
each fixed effect in turn showed the inclusion of both fixed effects significantly improved model fit.
For long-term retention, likelihood ratio tests lead to the null model being preferred.
Comparing the full model with a reduced model dropping each fixed effect in turn showed the
inclusion of neither fixed effect significantly improved model fit.
Fixed effect dropped in model
comparison
p
short-term retention device type 25.51 <.0001***
self-started 4.27 0.0387*
long-term retention device type 0.06 0.8022
self-started 0.69 0.4060
Table 9: Results of likelihood ratio tests for fixed effects in retention models.
Confidence intervals were computed based on a parametric bootstrapping, again using lme4
(for details see Bates, Mächler, Bolker & Walker, 2014, p.27). The preferred model predicts a greater
probability of short-term retention - for three months following the user’s second month on CommCare
- for Nokia users over Android users, and greater probability of short-term retention for users in
Dimagi-assisted projects over users in self-started projects. For long-term retention, neither
independent variable significantly improves the model. Model coefficients for the short-term retention
model are shown in Table 10.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 188
IV Fixed effect 2.5% 97.5% z
short-term
retention
device type =
nokia
3.38 2.01 5.50 4.91
self-started =
true
0.51 0.27 1.00 -2.10
Table 10: Parameter estimates for significant fixed effects in retention models.
The short-term retention model predicts that Nokia users are 3.38 times (238%) more likely to
be retained for three months following their sixth month on CommCare than Android users, 95% CI
[2.01 times, 5.5 times]. Users in self-started projects are around half (51%) as likely to be retained
compared to users in Dimagi-assisted projects, 95% CI [27%, 100%]. From the plot in Figure 28, it
appears the device type effect is at least partly driven by a large number of Android projects with low
or no retained users. Visual inspection of the corresponding plot, shown in Figure 29, confirms the
model finding this pattern does not hold for long-term retention.
Figure 28: Percent of users short-term retained per project, by device type
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 189
The marginal and conditional R
2
values for the short-term model are 0.0640 and 0.5551
respectively. The fixed effects explain 6.40% of the variance in short-term retention, while the model
as a whole explains 55.51%.
7.3 Discussion
Table 11 summarizes the findings in relation to hypotheses 1-6.
Figure 29: Percent of users long-term retained per project, by device type
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 190
H1 Users submitting data with the Android app will complete more visits
than Nokia device users submitting data with the J2ME app, at 6
(H1.1), 12 (H1.2) and 18 (H1.3) months.
H1.1 and H1.2
supported (months 6 and
12), H1.3 not supported
(month 18)
H2 Users in Dimagi-assisted projects complete more visits than users in
self-started projects, at 6 (H2.1), 12 (H2.2) and 18 (H2.3) months.
Not supported
H3 Users submitting data with the Android app are more likely to be
short-term retained (submitting data in both month two and month
five) than users submitting data with the J2ME app.
Not supported
H4 Users in Dimagi-assisted projects are more likely to be short-term
retained (submitting data in both month two and month five) than
users in self-started projects.
Supported
H5 Users submitting data with the Android app are more likely to be long-
term retained (submitting data in both month six and month twelve)
than users submitting data with the J2ME app.
Not supported
H6 Users in Dimagi-assisted projects are more likely to be long-term
retained (submitting data in both month six and month twelve) than
users in self-started projects.
Not supported
Table 11: Summary of CommCare data analysis findings
Android users were more productive than Nokia users at months 6 and 12. The relationship
was not significant at month 18 at the p < .05 level, although given the smaller number of users who
had been submitting data for 18 months or more, this may be partly due to the small sample size
relative to the effect size. H3 and H5 were not supported. The findings suggest that while Android
users are more productive, they are not more likely (in fact less likely) to be short-term retained, and
not more likely to be long-term retained.
Since the analysis was not set up to be able to claim causality, it is not possible to attribute
increased productivity (or decreased short-term retention) to the choice of Android phones. Indeed,
discussions with Dimagi staff suggested that Nokia users might be less productive because projects
that serve rural areas, with greater travel time between sites, are more likely to choose robust, battery-
saving Nokia phones. For short-term retention, the relatively straightforward process of installing and
testing the CommCare app on Andoid phones may lend itself to greater experimental use. Projects that
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 191
choose Nokia phones, by contrast, generally need to choose and acquire specific phone models in
advance, and perform installation of the app via one of the somewhat convoluted pathways for
installing third-party J2ME apps. Choosing Nokia phones may indicate a higher level of upfront
commitment. Greater short-term retention in Dimagi-assisted projects over self-started projects may be
similarly related to upfront commitment.
Although several hypotheses were supported, the proportion of the variance explained by the
fixed effects is very low. The effects of project and organization membership (the random effects) are
much greater. This is unsurprising given that the analysis was limited to considering independent
variables that could be derived from operational metadata. The larger point about these kinds of
analyses is that unless they are planned in advance – which secondary data analysis / data-driven
research is generally not – or otherwise augmented by additional analyses, they are almost certainly
subject to omitted variable bias.
In terms of using metadata from m4d projects to understand user behaviour, omitted variable
bias may translate to biases in how users and behaviour are conceptualized. Measuring success by user
numbers, for example, may obscure the demographic profile of users and hide demographic biases.
Impressive engagement, for example as measured through time spent on a particular page, may be at
least as much about lack of alternatives as a genuine interest in the content. In the example described
here, Android users are more productive, but it is possible that Nokia featurephones, being more
durable, enable users to serve remote areas more effectively. Secondary use of metadata from m4d
projects undoubtedly has a place in evaluation and impact assessment. However, the temptation to
replace formal evaluation research with data-driven research involving secondary data may be
misleading. This applies both to individual projects and the field as a whole.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 192
Given the prevalence of measures of use in claims of success made by m4d apps and services,
and the increasing availability of secondary data generated through use, it is important to understand
general opportunities and constraints for this kind of analysis. On this level, the main finding from the
data analysis exercise is that any attempt to establish correlation from secondary data inevitably suffers
from omitted variable bias. To some extent it may be possible to mitigate this with advance
consideration of how operational data might be used for secondary research. The multi-level structure
of the dataset, with users grouped into projects and organizations and sources of variance at user,
project and organization level, is also relevant to many analyses involving secondary data from m4d
apps and services. This is most obvious in the software-as-service model, but also applies to (for
example) education interventions where grouping by school or classroom is relevant. Failure to
account for grouping may cause unevenly distributed effects be missed. Overall, results from the data
analysis exercise indicate caution when planning and assessing data-driven research as a measure of
success for m4d apps and services.
In addition to demonstrating the limits of the evidence base, the nature of the evidence used to
make claims about the success of m4d apps and services is relevant to the discursive construction of
the field. Apps and services that achieve high user numbers, or whose 'innovative' deployment of
mobile technology against a pressing development 'problem' matches the preferred discursive frame of
industry bodies, gain prominence relative to those that do not. Without disaggregated demographic
information, it is possible to make a claim of “serving the underserved through mobile” (GSMA, 2013)
where in fact the audience for m4d apps and services in South Africa is strongly biased towards youth.
Similarly, research that describes the field by presenting example apps and services from different
sectors feeds a claim to greater diversity than is really present. Neither bias excludes the possibility that
m4d apps and services can be successful as development interventions. Rather, they remind us that
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 193
evidence about individual, group or organizational gains from m4d apps and services is deployed as
part of a discursive strategy of expansive rather than specific claims.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 194
CHAPTER 8: Conclusion
In this final chapter, I consider what, looking back at the non-smartphone era in ICT4D, we
might learn from the promise of m4d apps and services and its translation into practice in the South
African context. I begin by briefly summarising the findings and locating m4d as part of a broader
theme in which ICT4D promotes a succession of (new) technology solutions to development problems.
I then suggest analytical and practical responses, centring on recognizing historical and technical
contingencies in platforms for ICT4D interventions. Finally, returning to ideas about technology,
development and user-driven innovation, I propose broadening the problem space of ICT4D
interventions to take into account the wider communication ecology of mobile users in the developing
world.
8.1 The realisation of the m4d apps and services promise
Both the macro-level discursive realisation of the field of m4d and the micro-level
technological specificities of platforms have shaped the South African m4d apps and services
landscape, giving rise to particular concentrations and exclusions. The multi-level, multi-method
design of this study makes it possible to develop an argument about causal relationships across levels
of analysis. With reference to the unified conceptual framework developed in Chapter 2, the findings
from the four research components concentrate into the following main points:
1. Claims about the potential for m4d apps and services should be understood in the context of
embodied infrastructural limitations, arising from the gap between design context and the use
contexts in the developing world.
The m4d apps and services consensus claims that, since previously disconnected mobile users
in the developing world now have access to a computing and communications device, mobile apps and
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 195
services can be developed to serve their particular needs. This includes ameliorating lack of access (or
substandard access) to social services such as healthcare and education, as well as supporting
livelihood activities and providing access to information that would otherwise not be easily accessible
for reasons of connectivity, cultural norms or physical location. Like other ICT4D interventions,
interventions involving building and operating m4d apps and services are premised on the idea that
new technologies can be deployed to support development activities.
The inventory research identifies multiple established apps and services designed to support
development activities, particularly in health and education. However, both the challenges described by
interviewees operating established apps and services, and functional concentrations of apps and
services that exploit affordances and avoid constraints of particular platforms, demonstrate slippage
between the way the mobile technology platforms were designed to be used and the way designers of
third party apps and services want to use them. Free-to-user (or very low cost) services are perhaps the
best example. Examining the design history of the GSM network confirms differing design priorities
between the original design context – predominantly European business users – and third-party
services working in the context of development interventions.
ICT4D works with technologies that are not, in general, designed with the intended users of
ICT4D interventions in mind, and furthermore exist within an infrastructural legacy of the same.
Understanding how this affects the kinds of apps and services that are likely to be developed and be
successful, as well as identifying mitigating strategies for designers and developers, and technical and
operational interventions at platform level, is a practical necessity for ICT4D. It is also part of a critical
engagement with claims about the potential for successive new technologies to contribute to
development.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 196
2. Partly in response the limitations of the platform, development problems and potential m4d
interventions are often identified concurrently, with problems matched to intervention opportunities.
The result is a field with limited diversity, with designs for m4d apps and services reducible to a
relatively small number of archetypes.
Standard wisdom in ICT4D holds that technology solutions should only be developed when
they are the most appropriate response to development problems identified a priori. However,
interview respondents tended to describe the project initiation process as a co-evolution, in which a
development intervention and possible technology solution are identified and developed concurrently.
In the inventory, multiple apps and services with essentially the same functionality were found in every
sector, and grouped into archetypes to formally describe the functional clustering of m4d apps and
services. Archetypes were linked to the affordances and constraints of particular technology platforms,
such as youth-oriented information services developed to take advantage of the large, content-hungry
youth audience on Mxit. The claim of limited diversity arises from comparing the problem space of
development in general – based on context-specific needs, with many possible solutions only some of
which involve technology – with the solution space for m4d apps and services, which is defined,
among other factors, by the affordances and constraints of platforms for m4d apps and services.
One of the questions this raises is whether, as platforms like Mxit are replaced by new
platforms with different sets of affordances and constraints, the archetypes associated with them will
survive in the same form. In the case of mobile data collection apps and services moving from J2ME to
smartphone app platforms, interviewees saw the evolution as largely a positive one, with new features
like video capture supplementing older text-based data collection forms. The discoverability provided
by Mxit, however, is not readily replicated in any of the emerging social media platforms. The idea that
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 197
platforms influence the kinds of apps and services that are developed as m4d interventions also
requires us to consider what is lost or gained when old platforms are replaced by new ones.
3. Alignment with actants in the global network for m4d apps and services, including funders, mobile
network operators and “circuits of truth” (Roy 2010, location 123) such as the GSMA's Mobiles for
Development Impact programme produces additional incentives for concentration by archetype,
sector, intended users and geographic reach.
In addition to technology platforms, the findings suggested several other factors influencing
concentration by sector, archetype, intended users and geographic reach. Funder priorities, particularly
the focus on HIV/AIDS in the South African case, were suggested as a factor driving concentration in
the health and education sectors. Livelihood support apps and services, particularly as interventions
that would assist users in gaining employment, were identified by interviewees as a possible gap.
Funder priorities in turn were seen as subject to a particular set of ideas about sustainability and
development, with livelihoods projects expected to become financially self-sustaining where health
and education projects were not. This expectation is part of what Roy means by “circuits of truth” in
the production of ideas about development.
The geographic reach of apps and services in the inventory, with the exception of software-as-
service models in mobile data collection and a few large well-funded projects, was limited to a single
country. Interviewees who had worked with operators to obtain access to platform features noted that,
because operators are licensed by country and organize their infrastructure on a per-country basis,
working in multiple countries required multiple separate agreements with operators even if the parent
company was the same. Returning to the idea of co-evolution with platform affordances, intended
users for the majority of apps and services in the inventory were youth, with platforms such as Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 198
providing ready access to a youth audience, and youth generally seen to be more receptive to mobile
content.
Taken together the findings do not support idea that a highly diverse ecosystem of apps and
services will arise to (paraphrasing the GSMA) serve the underserved through mobile. Rather, existing
development priorities combined with technological affordances favour some kinds of apps and
services over others. These correspond to particular sectors, intended users, and technology platforms.
Identifying these, and highlighting areas that are underdeveloped, is a more realistic response to the
possibility of m4d apps and services than the narrowing of the problem space implicit in promoting an
expansive vision.
4. The discursive construction of the m4d consensus as it relates to apps and services is not neutral.
Whatever benefits accrue to users in the developing world, frames that position technologies as
offering solutions to development problems can be designed to achieve favourable public
perception, and by extension favourable regulatory treatment, for mobile industry players.
While not negating their potential and actual contributions, the development promise of m4d
apps and services functions as “discursive work” (Gillespie, 2010, p. 347) congruent with the interests
of major players in the mobile ecosystem. Interviewees identified discursive work in a sustained
programme of research and dissemination by the GSMA, as well as research lead or funded by
equipment manufacturers such as Nokia. They also noted that there are material advantages to a frame
that aligns mobile networks with development, for both shareholder value and regulatory treatment.
The discursive construction of mobile technology, and specifically apps and services, as an aid
to solving to development problems obscures the historically and socially contingent nature of the
platform. A more modest, more specific version of the promise of m4d apps and services, taking into
account constraints on the variety, sustainability, scalability and potential audience of m4d apps and
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 199
services, might more usefully direct efforts in ICT4D. The next section considers how ICTD research
might contribute to moderating the “discursive work” that aligns new technology paradigms with
development.
8.2 New technology paradigms and ICTD research
The relationships identified in the findings illuminate the discursive limitations of m4d as a
technology paradigm in ICT4D. One issue with this is that, across successive generations of ICT4D,
the knowledge production capacity of commercial players has not been matched by equivalent public
interest research (Gurumurthy, 2010, p.62). Also, it is widely recognized that ICT regulation struggles
to keep pace with technology innovation, a problem that is particularly acute in resource-constrained
settings. The nature and intent of “discursive work” by technology companies linking technology to
development is persistent enough to warrant routine consideration, and is a key subject for critical
ICTD research.
Looking ahead to the smartphone ecosystem, for example, there is a clear need for a systematic
critical analysis of initiatives such as Internet.org. Providing zero-rated data for Facebook and selected
health and education web content, the service has recently been the subject of popular criticism in India
on net neutrality grounds. Morozov (2014) points out that in addition to enticing users to Facebook,
third parties are incentivized to provide services within Facebook's platform. By dystopian extension,
he sees the decentralized open Internet returning to the centralized corporate control of the pre-
smartphone era. Although Morozov's particular dystopia seems difficult to reconcile with user demand
for this and other services – critically, affordable services – a critical perspective can illuminate issues
for further research as well as policy consideration.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 200
The more moderate lesson from the m4d consensus is that the interests of stakeholders the
discursive construction of the field should be matched with a strong focus on evaluation and context
specificity. Given the concentration of m4d apps and services – particularly info and content services –
in the health and education sectors, it seems appropriate to advocate for the means to make this content
widely available at no cost. Evaluation is critical, particularly in the form of neglected cost-benefit
studies. Good evaluation provides a strong basis to claim that, for example, mobile operators should be
encouraged to provide a service at no cost to users. Cost-benefit arguments would allow ICT4D
interventions to be judged against other development interventions, moving towards a mature field in
which technology interventions are normalized rather than considered separately as early-stage
innovation.
Concurrently with evaluation it is important to consider who is excluded from services on a
particular platform, or excluded in ways that mobile technology might be able to address. The youth
focus of m4d apps and services seems likely to be replicated in the smartphone era, and there may be
other specific exclusions. In Pakistan, for example, research among social media users found that
gender norms around public and private interactions prevented women from using Facebook
(Schoemaker, 2015). Understanding patterns of exclusion again opens the way for specific
interventions. The incentives leading technology companies to link technology to development further
encourage universalizing the claimed benefits. Development research is uniquely placed to consider
specific populations, historical evidence, and unintended consequences that may advance or diminish
the interests of vulnerable groups.
8.3 Platforms and design considerations for m4d apps and services
Consensus around mobile technology as an aid to solving to development problems has, partly
as a result of justifiable enthusiasm at the expansion of mobile communication, obscured the
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 201
historically and socially contingent nature of the platform. The findings of this research demonstrate
that limitations resulting from design decisions taken long ago and far away affect the variety,
sustainability, scalability and potential audience of m4d apps and services. Real gains notwithstanding,
technology solutions to development problems operate within an infrastructural legacy. For m4d apps
and services, the GSM network and platforms for apps and services are the most direct source of
affordances and constraints. Older infrastructures such as the electricity and road networks in turn
constrain where towers are placed, where hardware and airtime credit are sold, and where people can
use power-hungry devices and platforms. Technology diffusion to the developing world is always
uneven, both in terms of functionality and distribution.
On a practical level, considering the design logic on the one hand and the specific requirements
of m4d apps and services on the other is a good starting point for identifying 'pain points' in designing
for particular platforms. This is partly a case of bringing together different bodies of knowledge.
Interviewees who had built m4d apps and services had a good sense of what they and others had
struggled. They had less to say about where their problems arose in design of the mobile network and
platforms for apps and services, and what it would take for mobile operators to address them.
Operators, by contrast, were perceived as interested in the needs of m4d app and service users but not
well informed about specifics.
Looking ahead to smartphones, the design logic has two clear (related ) advantages: “layer
separation” resulting in less centralized control of the platform, and third-party access by design in the
form of app marketplaces and support for increasingly feature-rich mobile web sites. Yet concerns
specific to low-income or rural developing world users remain, among them hardware durability,
battery life and data costs. For models where users are provided with devices by an employer or and
organization of which they are a member, such as mobile data collection or school tablet programmes,
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 202
these are relatively easier to circumvent with technical and support process design. Apps and services
that rely on users owning and maintaining a smartphone have a smaller potential audience until these
concerns are resolved. Increasingly capable feature phones, particularly for apps and services on the
mobile web, are the more likely user-managed platform for the near future.
The finding that the design process for m4d apps and services is best represented as a co-
evolution of design based on affordances of the platform, the audience, and possible development
interventions is perhaps more relevant to development funders than designers themselves. Interviewees
expressed a sense that technology solutions should be developed in response to a thorough examination
of the problem space indicating a particular need. Yet the specialized nature of technology design for
development often lead to technologists – and by extension technology solutions – being included in
project initiation. Thoroughly examining the problem space prior to choosing a technology solution
might be best conducted as a separate task, undertaken independently. Alternatively, a technology
solution might be chosen in advance, but with an advance understanding of the relative cost-benefit
compared to other interventions.
Finally, the data and analysis challenges demonstrated in Chapter 7 might lead designers of
m4d apps and services to consider in advance how to facilitate secondary data use. It is widely
accepted that insights into user behaviour can be obtained from secondary analysis of operational data.
However m4d apps and services wanting to make claims about development impact, or otherwise take
advantage of operational data in evaluation, have particular data needs. These might be addressed in
technology design – for example by requiring demographic data that is critical for evaluation but might
otherwise be missed – as well as in evaluation design, for example by adding smaller validation studies
to confirm an inference made from a data set with omitted variables. The interview and inventory data
demonstrated that secondary analysis of operational data was already a significant basis for claims
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 203
about the success of m4d apps and services. Designing with evaluation in mind would make these
claims stronger, and improve the evidence base for m4d apps and services.
8.4 Broadening the scope of ICT4D interventions
Having considered a design perspective in the previous section, this section addresses the
lessons from the pre-smartphone m4d apps and services landscape from the perspective of
development interventions. In particular, I consider three possible (and non-exclusive) responses to
concentration by sector and archetype for m4d apps and services, most of which address either health
or education.
The first response derives from the finding that ICT4D intervention models using technology
(including mobile) are already well established in health and education sectors, with established
intervention models and donor preferences for social services having lead to sector and archetype
concentration. One response is to attempt to identify opportunities in other sectors, perhaps with
different funding models and a less well defined problem space. Interviewees suggested employment
and livelihood support as possible areas of unmet demand. It may be that, having established that
reaching (urban) youth through mobile is now mainstream development communication rather than
innovation, m4d apps and services are also able to focus on the more diverse unmet needs of specific
populations.
The second is to take advantage of the relatively more established body of evidence in the
health and education sectors – with reduction to archetypes providing a typology for intervention
design - to routinise the use of mobile apps and services in the delivery of social services. This includes
involving government and other established service providers, such as large health non-profits, in cost-
benefit analyses supported by good evaluation data. Regardless of whether donors continue to fund
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 204
m4d apps and services, established intervention models now exist and are mature enough to move
beyond pilots and consider the development of sustainable implementation ecosystems. Mobile data
collection, which is a core business model for multiple specialist ICT4D technology companies, is the
most routinised archetype. Further study of the sustainability model of this ecosystem and how it does
or does not contribute to better access, lower cost and better quality in developing world social
services, is a useful avenue for research on bringing archetypes and sectors – rather than individual
apps and services – to scale.
A third response is to consider how, having seen the vast expansion of access to mobile
communication, development interventions might look beyond a model where the development
industry builds apps and services for beneficiaries. In user research conducted by Research ICT Africa
in 2012, they found that users at the bottom of the pyramid were interested in the same apps and
services as the rest of the pyramid – they just needed them to be cheaper. The charge of
“developmentalism” (Liang 2010) in ICT4D should not lead us to abandon to effective m4d
interventions in the form of apps and services. However, there is also a need for policy, research and,
conceivably, technical interventions that address affordable access to a broad-based communication
ecosystem for low-income users.
Going one step further, we might also consider how developing world users are using mobile
platforms to produce and publish their own content, and develop their own formalized interactions
analogous to m4d apps and services. Research among small Facebook users in Nairobi (Wyche, Forte
& Schoenebeck, 2013) found significant use of Facebook for livelihood activities, including promoting
small businesses online, accessing remittance flows from contacts overseas, and job search. The
question for development interventions then becomes how best to support user-driven innovation. This
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 205
may include formalizing services but may equally return to the idea of supporting affordable access to
a broad range of communication tools, technologies and platforms.
Despite the mixed track record of m4d apps and services in the pre-smartphone era, there is
reason to be optimistic about what is to come. Platforms designed on a logic of openness, including the
mobile web and smartphone app environments, are replacing limited platforms premised on mobile
operator control. Vast amounts of data about user interactions with apps and services within and
beyond m4d are increasingly available for analysis, insights from which could help m4d apps and
services demonstrate impact and improve intervention design. Particularly in health and education,
there are now mature intervention models that are ready to move beyond the pilot phase to become
powerful yet routine tools for teachers, students, healthcare workers and administrators. The ideas
discussed in this chapter intervene at various levels to suggest how ICT4D might respond to new
technology paradigms, based on findings from the pre-smartphone era. In the face of exaggerated
claims about the potential of new technologies for development, it is still possible to make moderate,
nuanced and contingent claims, both in terms of narrowing the scope of individual apps and services
for particular audiences, and broadening the problem space for ICT4D interventions.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 206
References
“A cheap new tablet that can speak Zulu”. (2014, October 24). Retrieved from
http://www.balancingact-africa.com/news/en/issue-no-730/computing/a-cheap-new-tablet-t/en
“Halfway There: How to promote the spread of mobile phones among the world's poorest.” (2008,
May 29). The Economist. Retrieved from http://www.economist.com/node/11465558
“M-Pesa boosts Safaricom revenue as the company posts historic Sh25b pretax profit.” (2013, May
15). Standard Digital News - Kenya. Nairobi, Kenya. Retrieved from
http://www.standardmedia.co.ke/?articleID=2000083601
“Please Call Me” plaintiff got R5m boost. (2013, August 12). Mail and Guardian. Retrieved from
http://mg.co.za/article/2013-08-12-please-call-me-plaintiff-got-r5m-boost/
Acker, A. (2014). The Short Message Service: Standards, infrastructure and innovation. Telematics and
Informatics, 31.4 (2014). 559-568.
Agar, J. (2013). Constant Touch: A Global History of the Mobile Phone. London, UK: Icon Books.
Aker, J., & Mbiti, I. (2010). Mobile phones and economic development in Africa. Center for Global
Development Working Paper. Retrieved from
http://papers.ssrn.com.libproxy.usc.edu/sol3/papers.cfm?abstract_id=1693963
Anand, S. A. (2005). Development of a Cellphone Based Monitoring and Management-Support
System for Anti-Retroviral Therapy (MSc dissertation). University of Cape Town, South Africa.
Ancker, J. S., Shih, S., Singh, M. P., Snyder, A., Edwards, A., Kaushal, R., & the HITEC Investigators.
(2011). Root Causes Underlying Challenges to Secondary Use of Data. AMIA Annual
Symposium Proceedings, 2011, 57.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 207
Andrade, A. D., & Urquhart, C. (2010). The affordances of actor network theory in ICT for
development research. Information Technology & People, 23(4). 352-374.
Armstrong, P., Lekezwa, B., & Siebrits, K. (2008). Poverty in South Africa: A profile based on recent
household surveys. Stellenbosch Economic Working Paper, 4.
Avgerou, C. (2010) Discourses on ICT and development. Information Technologies and International
Development, 6 (3). 1-18.
Bailey, K. D. (1994). Typologies and taxonomies in social sciences. In Bailey, K. D. (Ed.). (1994).
Typologies and taxonomies: an introduction to classification techniques. Thousand Oaks,
California: Sage Publications. 1-17.
Bar, F., & Riis, A. M. (2000). Tapping user-driven innovation: A new rationale for universal service.
The Information Society, 16(2), 99-108.
Bar, F., Pisani, F., & Weber, M. (2007). Mobile technology appropriation in a distant mirror: baroque
infiltration, creolization and cannibalism . Available at
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.184.5809&rep=rep1&type=pdf
Barrantes, R., & Galperin, H. (2008). Can the poor afford mobile telephony? Evidence from Latin
America. Telecommunications Policy, 32(8), 521-530.
Bartoń, K. (2014). MuMIn: Multi-Model Inference. R package version 1.12.1. http://CRAN.R-
project.org/package=MuMIn
Bates, D., Mächler, M., Bolker, B.M. and Walker, S. (2014). “lme4: Linear mixed-effects models using
Eigen and S4.” ArXiv e-print; submitted to Journal of Statistical Software. Retrieved from
http://arxiv.org/abs/1406.5823.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 208
Bayley, K. B., Belnap, T., Savitz, L., Masica, A. L., Shah, N., & Fleming, N. S. (2013). Challenges in
Using Electronic Health Record Data for CER: Experience of 4 Learning Organizations and
Solutions Applied. Medical Care, 51, S80–S86. doi:10.1097/MLR.0b013e31829b1d48
Bebbington, A. (2000). Reencountering development: Livelihood transitions and place transformations
in the Andes. Annals of the Association of American Geographers, 90(3), 495-520.
Benbasat, I., Goldstein, D. K., & Mead, M. (1987). The case research strategy in studies of information
systems. MIS quarterly, 369-386.
Blaschke, S., Bokenkamp, K., Cosmaciuc, R., Denby, M., Hailu, B., & Short, R. (2009). Using Mobile
Phones to Improve Child Nutrition Surveillance in Malawi. UNICEF Malawi/UNICEF
Innovations. Retrieved from
http://new.sipa.columbia.edu/sites/default/files/UNICEFFinalReport_2009_0.pdf
Bogost, I., & Montfort, N. (2007). New media as material constraint: An introduction to platform
studies. Paper presented at the 1st International HASTAC Conference, Duke University,
Durham NC. Retrieved from http://bogost.com/downloads/Bogost%20Montfort
%20HASTAC.pdf
Botsis, T., Hartvigsen, G., Chen, F., & Weng, C. (2010). Secondary Use of EHR: Data Quality Issues
and Informatics Opportunities. Summit on Translational Bioinformatics, 2010, 1.
Bowker, G. C., Baker, K., Millerand, F., & Ribes, D. (2010). Toward information infrastructure studies:
Ways of knowing in a networked environment. In Hunsinger, J., Klastrup, L., & Allen, M.
(Eds.) International handbook of internet research. 97-117. Dordrecht, Netherlands: Springer.
Brett, E. (2009). Reconstructing development theory: International inequality, institutional reform and
social emancipation. Basingstoke, UK: Palgrave Macmillan.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 209
Brown, J., Shipman, B., & Vetter, R. (2007). SMS: The short message service. Computer, 40(12), 106-
11
Burrell, J. (2012). Invisible Users: Youth in the Internet Cafes of Urban Ghana. Cambridge, MA, USA:
MIT Press.
Burrell, J., & Toyama, K. (2009). What constitutes good ICTD research?. Information Technologies &
International Development, 5(3), 82-94.
Callon, M. (1986). Some elements of a sociology of translation: domestication of the scallops and the
fishermen of St Brieuc Bay. In Law, J. (Ed.). Power, action and belief: A new sociology of
knowledge, 32, 196-233.
Castells, M. (1996). The rise of the network society: The information age: Economy, society, and
culture (Vol. 1). Oxford: Wiley-Blackwell.
Castells, M. (2012). Networks of Outrage and Hope. Social Movements in the Internet age. Cambridge,
UK: Polity.
Castells, M., Fernandez-Ardevol, M., Qiu, J. L., & Sey, A. (2007). Mobile Communication and Society
- A Global Perspective. Cambridge, MA: MIT Press.
Chatfield, A., Javetski, G., Fletcher, A., & Lesh, N. (2014). CommCare Evidence Base. Dimagi, Inc.
Retrieved from https://help.commcarehq.org/download/attachments/12226140/CommCare
%20Evidence%20Base%20April%202014.pdf
Constant, D., de Tolly, K., Harries, J., & Myer, L. (2014). Mobile phone messages to provide support
to women during the home phase of medical abortion in South Africa: a randomised controlled
trial. Contraception, 90(3), 226–233. http://doi.org/10.1016/j.contraception.2014.04.009
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 210
Corker, J. (2010). "Ligne Verte" Toll-Free Hotline: Using Cell Phones to Increase Access to Family
Planning Information in the Democratic Republic of Congo. Cases in Public Health
Communication & Marketing, 4, 23-37.
de Tolly, K., Skinner, D., Nembaware, V., & Benjamin, P. (2012). Investigation into the use of short
message services to expand uptake of human immunodeficiency virus testing, and whether
content and dosage have impact. Telemedicine and e-Health, 18(1), 18-23.
Dedoose Version 5.0.11, web application for managing, analyzing, and presenting qualitative and
mixed method research data (2014). Los Angeles, CA: SocioCultural Research Consultants,
LLC. Available at http://www.dedoose.com
DeRenzi, B., Findlater, L., Payne, J., Birnbaum, B., Mangilima, J., Parikh, T., … Lesh, N. (2012).
Improving community health worker performance through automated SMS. In Proceedings of
the Fifth International Conference on Information and Communication Technologies and
Development (pp. 25–34). New York, NY, USA: ACM. Doi:10.1145/2160673.2160677
Diga, K. (2008). Mobile Cell Phones and Poverty Reduction: Technology Spending Patterns and
Poverty Level Change among Households in Uganda. Workshop on the Role of Mobile
Technologies in Fostering Social Development. Sao Paulo, Brazil.
Dodson, L., Sterling, R., & Bennet, J. (2013, December). Minding the Gaps: Cultural, Technical and
Gender-Based Barriers to Mobile Use in Oral-Language Berber Communities in Morocco. In
Proceedings of the Sixth International Conference on Information and Communication
Technologies and Development: Full Papers - Volume 1. 79-88. USA: ACM.
Donner, J. (2008). Shrinking fourth world? Mobiles, development, and inclusion. In Katz, J. (Ed).
Handbook of Mobile Communication Studies, 29-42. Cambridge, Massachusetts: MIT Press.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 211
Retrieved from http://www.ingedewaard.net/papers/mobile/2008_Handbook of Mobile
Communication Studiesby.doc
Donner, J. (2009). Mobile-based livelihood services in Africa: pilots and early deployments. In
Fernández-Ardèvol, M., & Ros, A. (Eds). Communication technologies in Latin America and
Africa: A multidisciplinary perspective. 37-58. Barcelona, Spain: Universitat Oberta de
Catalunya.
Donner, J., & Escobari, M. X. (2010). A review of evidence on mobile use by micro and small
enterprises in developing countries. Journal of International Development, 658(5), 641–658.
doi:10.1002/jid
Donner, J., Gitau, S., & Marsden, G. (2011). Exploring mobile-only internet use: results of a training
study in urban South Africa. International Journal of Communication, 5, 574–597.
Dourish, P. (2006). Implications for design. In Proceedings of the SIGCHI conference on Human
Factors in computing systems. 541-550. New York, NY, USA: ACM.
Dunnewijk, T., & Hultén, S. (2007). A Brief History of Mobile Communication in Europe. Telematics
and Informatics, 24(3), 164-179.
Elgin, B., & Silver, V. (2012, February). Syria Disrupts Text Messaging of Protesters With Made-in-
Dublin Equipment. Bloomberg. Retrieved from http://mobile.bloomberg.com/news/2012-02-
15/syria-blocks-texts-with-dublin-made-gear.html
Escobar, A. (1995). Encountering Development: The Making and Unmaking of the Third World.
Princeton, NJ, USA: Princeton University Press.
Estrin, D., & Sim, I. (2010). Open mHealth architecture: an engine for health care innovation. Science,
330(6005), 759-760.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 212
European Telecommunications Standards Institute (ETSI). (1996). Digital cellular telecommunications
system: Unstructured Supplementary Service Data (USSD) - Stage 2 (GSM 03.90). Valbonne,
France: ETSI Secretariat.
Faik, I., Thompson, M. and Walsham, G. (2013). Facing the Dilemmas of Development:
Understanding Development Action Through Actor-Network Theory. Actor-Network Theory
for Development Working Paper Series, University of Manchester. Available at
http://www.cdi.manchester.ac.uk/resources/ant4d/
Farley, T. (2005). Mobile telephone history. Telektronikk, 101(3/4), 22. Retrieved from
http://www.telecomwriting.com/archive/TelenorPage_022-034.pdf
Funk, J. L. (2007). Solving the startup problem in Western mobile Internet markets.
Telecommunications Policy, 31(1), 14–30. doi:10.1016/j.telpol.2006.11.003
Garud, R., Jain, S., & Kumaraswamy, A. (2002). Institutional Entrepreneurship in the Sponsorship of
Common Technological Standards: The case of Sun Microsystems and Java. Academy of
Management Journal, 45(1), 196-214.
Gillespie, T. (2010). "The politics of 'platforms'". New Media Society, 12 (347), 347-364,
doi:10.1177/1461444809342738
Gillwald, A. & Stork, C. (2009). Towards Evidence-Based ICT Policy and Regulation: ICT access and
usage in Africa. Vol 1 Policy Paper 2. Johannesburg: Research ICT Africa. Retrieved May 07,
2009 from http://www.researchictafrica.net/new/images/uploads/ria-policy-paper_ict-access-
and-usage-2008.pdf
Goodman, D. (2005). Linking mobile phone ownership and use to social capital in rural South Africa
and Tanzania. The Vodafone Policy Paper Series, 2(03), 53 - 65.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 213
GSM Association. (2013). Mobile for Development. Retrieved from
http://www.gsma.com/mobilefordevelopment/
Gurumurthy, A. (2010). From Social Enterprises to Mobiles: Seeking a Peg to Hang a Premeditated
ICTD Theory. Information Technologies & International Development, 6 (SE), 57-63.
Haberer, J. E., Kiwanuka, J., Nansera, D., Wilson, I. B., & Bangsberg, D. R. (2010). Challenges in
using mobile phones for collection of antiretroviral therapy adherence data in a resource-
limited setting. AIDS and Behavior, 14(6), 1294-1301.
Habib, A. (2013, August 8). Civil society is our best way to check misrule. Mail and Guardian.
Retrieved from http://mg.co.za/article/2013-08-08-00-civil-society-is-our-best-way-to-check-
misrule/
Hahn, H. P., & Kibora, L. (2008). The domestication of the mobile phone: oral society and new ICT in
Burkina Faso. The Journal of Modern African Studies, 46(01), 87-109.
Hammersley, M., & Atkinson, P. (2007). Ethnography: Principles in Practice (Third edition.). London,
UK; New York, NY, USA: Routledge.
Harris, I. (2010). The Evolution of SMS Features and Specifications from the beginning of 1997 to
Mid-2009. In F. Hillebrand (ed.) Short Message Service (SMS): The creation of personal global
text messaging. 111-134. West Sussex, UK: Wiley
Hatt, T., Wills, A., & Harris, M. (2013). Scaling Mobile for Development: A developing world
opportunity. GSM Association. Retrieved from
https://mobiledevelopmentintelligence.com/insight/Scaling_Mobile_for_Development:_A_dev
eloping_world_opportunity
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 214
Heeks, R. (2006). Health information systems: Failure, success and improvisation. International
Journal of Medical Informatics, 75(2), 125-137.
Heeks, R. (2008). ICT4D 2.0: The next phase of applying ICT for international development.
Computer, 41(6), 26-33.
Heeks, R. (2010). Do information and communication technologies (ICTs) contribute to development?.
Journal of International Development, 22(5), 625-640.
Heeks, R. (2013). Development Studies Research and Actor-Network Theory (Actor-Network Theory
for Development Working Paper 1). Centre for Development Informatics, University of
Manchester. Retrieved January 21, 2015 from http://www.cdi.manchester.ac.uk/resources/ant4d
Heeks, R., & Stanforth, C. (2007). Understanding e-Government project trajectories from an actor-
network perspective. European Journal of Information Systems, 16(2), 165-177. Rhodes, J.
(2009). Using Actor-Network Theory to Trace an ICT (Telecenter) Implementation Trajectory
in an African Women's Micro-Enterprise Development Organization. Information Technologies
& International Development, 5(3), 1-20.
Heikkinen, M. (2010, October). Knock, Knock, Nokia's Heavy Fall. Helsingin Sanomat, Kuukausiliite
monthly supplement for October 2010. Retrieved from
http://www.howardforums.com/showthread.php/1679550-Knock-Knock-Nokia-s-Heavy-Fall
Hellström, J. (2010). The innovative use of mobile applications in East Africa. Swedish International
Development Cooperation Agency (Sida). Retrieved from
http://www.texttochange.org/sites/default/files/newsfiles/sr2010-12_sida_hellstrom.pdf
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 215
Hersh, W. R., Weiner, M. G., Embi, P. J., Logan, J. R., Payne, P. R. O., Bernstam, E. V., … Saltz, J. H.
(2013). Caveats for the Use of Operational Electronic Health Record Data in Comparative
Effectiveness Research. Medical Care, 51, S30–S37. doi:10.1097/MLR.0b013e31829b1dbd
Hess, M., & Coe, N. M. (2006). Making connections: global production networks, standards, and
embeddedness in the mobile-telecommunications industry. Environment and Planning, 38(7),
1205-1227.
Hillebrand, F. (2010). The Creation of the SMS Concept from Mid-1984 to Early 1987? In F.
Hillebrand (ed.) Short Message Service (SMS): The creation of personal global text messaging.
23-44. West Sussex, UK: Wiley
Hillebrand, F., Rosenbrock, K. H., and Hauser, H. (2013). The Creation of Standards for Global
Mobile Communication: GSM and UMTS Standardization from 1982 to 2000. Wireless
Communications. Bonn, Germany: Hillebrand Consulting Engineers GmbH. Retrieved from
http://www.etsi.org/images/files/news/CreationOfStandardsForGlobalMobileCommunication.e
pub
Holley, K. (2010). The Evolution of SMS Features and Specifications from October 1990 to the end of
1996. In F. Hillebrand (ed.) Short Message Service (SMS): The creation of personal global text
messaging. 75-97. West Sussex, UK: Wiley.
Horst, H. A. (2013). The infrastructures of mobile media: Towards a future research agenda. Mobile
Media & Communication, 1(1), 147-152.
Horst, H., and Miller, D. (2005). From Kinship to Link-up: Cell Phones and Social Networking in
Jamaica. Current Anthropology, 46(5), 755-778.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 216
ict4chw. (2010). brainstorm! CHWs and airtime. [Electronic mailing list message thread]. Retrieved
from https://groups.google.com/forum/#!
searchin/ict4chw/managing$20airtime/ict4chw/dm1VFb5XQOE/9aKSAKp4qIIJ
International Telecommunications Union (ITU). (2014). Key ICT indicators for developed and
developing countries and the world (totals and penetration rates). Retrieved from
http://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx
International Telecommunications Union. (2004). Has Africa’s ICT renaissance begun? Retrieved
August 10, 2013, from http://www.itu.int/AFRICA2004/media/renaissance.html
Jensen, R. (2007). The digital provide: Information (technology), market performance, and welfare in
the South Indian fisheries sector. The Quarterly Journal of Economics, 122(3), 879-924.
Kuznetsova, A., Bruun Brockhoff, P., and Christensen, R. (2014). lmerTest: Tests in Linear Mixed
Effects Models. R package version 2.0-20. http://CRAN.R-project.org/package=lmerTest
Lacohée, H., Wakeford, N., & Pearson, I. (2003). A Social History of the Mobile Telephone with a
View of its Future. BT Technology Journal, 21(3), 203-211.
Lasar, M. (2008, December 16). FCC told that SMS short codes should be open to all. Ars Technica.
Retrieved October 26, 2013, from http://arstechnica.com/tech-policy/2008/12/fcc-told-that-
sms-short-codes-should-be-open-to-all/
Latour, B. (2005). Reassembling the social: An introduction to actor-network-theory. Oxford, UK:
Oxford University Press. http://site.ebrary.com.libproxy.usc.edu/lib/uscisd/docDetail.action?
docID=10233636
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 217
Law, J., & Callon, M. (1992). The life and death of an aircraft: a network analysis of technical change.
In W.E. Bijker & J. Law (Eds.), Shaping technology/building society: Studies in sociotechnical
change, 21-52. Cambridge, MA, USA: MIT Press
Le Bodic, G. (2005). Mobile Messaging technologies and services: SMS, EMS and MMS. Chichester,
UK: John Wiley & Sons.
Liang, L. (2010). Developmentalism : Technology and the Intellectual Life of the Poor. Information
Technologies & International Development, 6, 65-67.
MacKenzie, D. and Wajcman, J. 1999. The Social Shaping of Technology. 2nd edition. Milton Keynes,
UK: Open University Press
Marcus, G. E. (1995). Ethnography in/of the world system: The emergence of multi-sited ethnography.
Annual Review of Anthropology, 95-117.
Mason, P. (2012). Why it's kicking off everywhere: The new global revolutions. London, UK: Verso.
Maurer, B. (2012). Mobile money: Communication, consumption and change in the payments space.
Journal of Development Studies, 48(5), 589-604.
May, J., Waema, T. M., & Bjåstad, E. (2014). Introduction: The ICT/poverty nexus in Africa. ICT
Pathways to Poverty Reduction. Ottawa, Canada: International Development Research Center.
Retrieved from http://www.idrc.ca/en/resources/publications/openebooks/539-
7/index.html#ch01.
McLeod, D. (2013, April 17). Vodacom debuts SMS interconnect. Retrieved from
http://www.techcentral.co.za/vodacom-debuts-sms-interconnect/39657/
Mercer, C. (2006). Telecentres and transformations: Modernizing Tanzania through the Internet.
African Affairs, 105(419), 243-264.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 218
Montfort, N., & Bogost, I. (2009). Racing the Beam: The Atari Video Computer System. Cambridge,
MA, USA: MIT Press.
Morawczynski, O. (2009). Exploring the usage and impact of “transformational” mobile financial
services: the case of M-PESA in Kenya. Journal of Eastern African Studies, 3(3), 509-525.
Morozov, E. (2014, August 3). Facebook’s Gateway Drug. The New York Times, p. SR5. Retrieved
from http://www.nytimes.com/2014/08/03/opinion/sunday/evgeny-morozov-facebooks-
gateway-drug.html
Muller, R. (2012, May 24). Should you pay to stop SMS spam? Retrieved March 16, 2015, from
http://mybroadband.co.za/news/cellular/50617-should-you-pay-to-stop-sms-spam.html
Mxit. (2014, January). Mxit User Insights and Statistics. Retrieved January 11, 2015, from
http://advertise.mxit.com/files/Mxit_Stats_Pack_latest.pdf
Mxit. (n.d.). About. Retrieved January 11, 2015, from http://get.mxit.com/about/
National Communications System. SMS over SS7. Technical Report NCS TID 03-2, National
Communications System, December 2003. http://www.di.unisa.it/~ads/corso-
security/www/CORSO-0304/SMS/SMSoverSS7.htm
Opera Software. (2013, November). A closer look at local mobile-surfing habits globally - Opera
Software. Retrieved January 10, 2015, from http://www.operasoftware.com/smw/2013-11
Orlikowski, W. (2000). Using technology and constituting structures: A practice lens for studying
technology in organizations. Organization Science, 11(4), 404-428.
Palomäki, J. (2004). Case WAP: reasons for failure. In Luukkainen, S. (Ed.) Innovation Dynamics in
Mobile Communications: Proceedings of the Research Seminar on Telecommunications
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 219
Business. Helsinki, Finland: Helsinki University of Technology Publications in
Telecommunications Software and Multimedia . 981-101.
Patnaik, S., Brunskill, E., & Thies, W. (2009, April). Evaluating the accuracy of data collection on
mobile phones: A study of forms, SMS, and voice. In Proceedings of the International
Conference on Information and Communication Technologies and Development (ICTD), Doha,
Qatar, 2009.
Pottier, N. (2011, December 12). Re: virtual SIM hosting. [Electronic mailing list message] Retrieved
from https://groups.google.com/forum/#!searchin/rapidsms/two-
way/rapidsms/dRJyMZgWi6g/xa8lK_s-cbsJ
Praekelt Foundaton. (2013). Case Studies. Vumi. Retrieved October 29, 2013, from
http://vumi.org/case-studies/
Prahalad, C. K. (2006). The Fortune at the bottom of the pyramid: Eradicating poverty through profits.
Upper Saddle, NJ, USA: Prentice Hall.
Qiang, C. Z., Yamamichi, M., Hausman, V., Miller, R., & Altman, D. (2012). Mobile Applications for
the Health Sector. World Bank ICT Sector Unit. Retrieved from http://www-
wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2012/09/18/000333038_20
120918020413/Rendered/PDF/726040WP0Box370th0report00Apr020120.pdf
Qusay, H. (2004, January). J2ME Luminary Antero Taivalsaari. Retrieved from
http://www.oracle.com/technetwork/systems/taivalsaari-156729.html
Rahnema, M. (1997). Towards Post-Development: Searching for Signposts, a New Language and New
Paradigms. In Rahnema, M., & Bawtree, V. (Eds). The post-development reader. London, UK:
Zed Books. 304-331
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 220
Rashid, A. T., & Elder, L. (2009). Mobile Phones and Development: An Analysis of IDRC-supported
Projects. The Electronic Journal of Information Systems in Developing Countries, 36, 1-16.
Ray, B. (2011, October 6). Gemalto squeezes Facebook onto a SIM. The Register. Retrieved from
http://www.theregister.co.uk/2011/02/14/facebook_sim/
Redl, S., Weber, M., & Oliphant, M. (1998). GSM and Personal Communications Handbook. Boston,
MA: Artech House.
Research ICT Africa. (2012). Mobile Usage at the Bottom of the Pyramid in South Africa. InfoDev.
Retrieved from http://www.infodev.org/en/Publication.1193.html
Rhodes, J. (2009). Using actor-network theory to trace an ICT (telecenter) implementation trajectory in
an African women's micro-enterprise development organization. Information Technologies &
International Development, 5(3), 1-20.
Rovinen, P. & Yla¨-Antilla, P. (2004). Little Finland’s transformation to a wireless giant. In Dutta, S.,
Lanvin,B., & Paua, F. (Eds.). The Global Information Technology Report: Towards an
equitable information society. Oxford, UK: Oxford University Press
Roy, A. (2010). Poverty capital: Microfinance and the making of development. New York, NY, USA:
Routledge.
Safaricom. (2013). Celebrating 5 years of changing lives. Retrieved October 23, 2013, from
http://www.safaricom.co.ke/mpesa_timeline/timeline.html
Samuel, J., Shah, N., & Hadingham, W. (2005). Mobile communications in South Africa, Tanzania,
and Egypt: Results from community and business surveys. The Vodafone Policy Paper Series,
2(03), 44-52.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 221
Satchell, C., & Dourish, P. (2009). Beyond the User: Use and Non-use in HCI. In Proceedings of the
21st Annual Conference of the Australian Computer-Human Interaction Special Interest Group:
Design. 9-16. New York, NY, USA: ACM
Sauter, M. (2010). From GSM to LTE: An Introduction to Mobile Networks and Mobile Broadband.
Chichester, UK: John Wiley & Sons.
Schoemaker, E. (2015, July). Facebook Domestication. Tanqeed. Retrieved from
http://www.tanqeed.org/2015/07/facebook-domestication/
Schuurman, F. J. (2000). Paradigms lost, paradigms regained? Development studies in the twenty-first
century. Third World Quarterly, 21(1), 7-20.
Selwyn, N. (2003). Apart from technology: understanding people’s non-use of information and
communication technologies in everyday life. Technology in Society, 25(1), 99-116.
Selwyn, N. (2006). Digital division or digital decision? A study of non-users and low-users of
computers. Poetics, 34(4), 273-292.
Sen, A. K. (1999). Development as freedom. Oxford, UK: Oxford University Press.
Sey, A. (2008). Mobile communication and development: a study of mobile phone appropriation in
Ghana (Doctoral dissertation, University of Southern California).
Sey, A. (2011). 'We use it different, different': Making sense of trends in mobile phone use in Ghana.
New Media & Society, 13(3), 375-390.
Sigurdson, J. (2001). WAP OFF - Origin, Failure and Future. Working paper No. 135. East Asia
Science and Technology Programme, The Stockholm School of Economics. Stockholm,
Sweden. Retrieved from http://www2.hhs.se/eijswp/135.PDF
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 222
Skuse, A., & Cousins, T. (2008). Getting connected: the social dynamics of urban telecommunications
access and use in Khayelitsha, Cape Town. New Media & Society, 10(1), 9-26.
Smith, M. L., Spence, R., & Rashid, A. T. (2011). Mobile phones and expanding human capabilities.
Information Technologies & International Development, 7(3), 77-88.
Song, S. (2009). WGSDIA - Lobby For Cheaper SMS Charges. Many Possibilities. Retrieved October
26, 2013, from http://manypossibilities.net/2009/06/wgsdia-lobby-for-cheaper-sms-charges/
South Africa Audience Research Foundation. (2013). Cell Phone Trends. Retrieved from
http://saarf.co.za/amps/cellphone.asp
Spivak, G. C. (1999). A critique of postcolonial reason. Cambridge, MA, USA: Harvard University
Press.
Star, S. L. (2002). Infrastructure and ethnographic practice: Working on the fringes. Scandinavian
Journal of Information Systems, 14(2), 6.
Star, S.L. (1999). Ethnography of infrastructure. American Behavioral Scientist, 43(3), 377-391.
Star, S.L., & Bowker, G.C. (2006). How to infrastructure. In L. A. Lievrouw & S. M. Livingstone
(Eds.), Handbook of New Media: Social Shaping and Social Consequences of ICTs. London,
UK: Sage Publications. 151-162.
Star, S.L., & Griesemer, J.R. (1989). Institutional ecology, 'translations' and boundary objects:
Amateurs and professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39. Social
Studies of Science, 19, 387-420.
Statistics South Africa. (2012). Poverty Profile of South Africa: Application of poverty lines on the
Living Conditions Survey 2008/09. Pretoria: Statistics South Africa
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 223
Stein, D., & Valters, C. (2012). Understanding theory of change in international development. Justice
and Security Research Programme Paper 1. International Development Department, London
School of Economics. Retrieved from http://www.theoryofchange.org/wp-
content/uploads/toco_library/pdf/UNDERSTANDINGTHEORYOFChangeSteinValtersPN.pdf
Stiglitz, J. E. (2007). Making globalization work. New York, NY, USA: WW Norton.
Suchman, L., Blomberg, J., Orr, J. E., & Trigg, R. (1999). Reconstructing technologies as social
practice. American Behavioral Scientist, 43(3), 392-408.
Tarnacha, A., & Maitland, C. F. (2006,). Entrepreneurship in Mobile Application Development. In
Proceedings of the 8th international Conference on Electronic Commerce: The new e-
commerce: innovations for conquering current barriers, obstacles and limitations to conducting
successful business on the internet. 589-593. ACM.
Taylor, A. S., & Vincent, J. (2005). An SMS history. In Hamill, L. and Lasen, A. (Eds): Mobile World:
Past, Present and Future. 75-91. Springer: London
Telecom Regulatory Authority of India (TRAI). (2013, November 26th). Information Note to the Press
(Press Release No. 82/2013). Retrieved from
http://www.trai.gov.in/WriteReadData/WhatsNew/Documents/PR%20%20-%20USSD
%20based%20mobile%20banking%20services-New26112013.pdf
Thomas, S. (2015, February 27). Mxit: the rise and collapse of 'Africa’s largest social network'.
Memeburn. Retrieved from http://memeburn.com/2015/02/mxit-the-rise-and-collapse-of-
africas-largest-social-network/
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 224
Trosby, F. (2010). The Technical Design of SMS in DGMH from June 1987 to October 1990. In F.
Hillebrand (ed.) Short Message Service (SMS): The creation of personal global text messaging.
45-73. West Sussex, UK: Wiley
UNAIDS (2012). UNAIDS report on the global AIDS epidemic. Joint United Nations Programme on
HIV/AIDS (UNAIDS). Retrieved from
http://www.unaids.org/sites/default/files/media_asset/20121120_UNAIDS_Global_Report_201
2_with_annexes_en_1.pdf
Ureta, S. (2008). To move or not to move? Social exclusion, accessibility and daily mobility among the
low income population in Santiago, Chile. ‐ Mobilities, 3(2), 269-289.
Van der Boor, P., Oliveira, P., & Veloso, F. (2014). Users as innovators in developing countries: The
global sources of innovation and diffusion in mobile banking services. Research Policy, 43(9),
1594-1607.
Velge, F. (2013). Literacy Acquisition, Informal Learning and Mobile Phones in a South African
Township. In Proceedings of the Sixth International Conference on Information and
Communication Technologies and Development: Full Papers - Volume 1. 89-100. USA: ACM.
Vodafone India. (2013). Vodafone Vernacular SMS: Messaging in Hindi, Gujarati & Marathi | India.
Retrieved October 29, 2013, from
https://www.vodafone.in/vas/mailmessaging/pages/vernacularsms.aspx
Vodafone UK. (2013, June 21). JustTextGiving dissected: How does text donation work? Vodaphone
Social. Retrieved from http://blog.vodafone.co.uk/2013/06/21/justtextgiving-dissected-how-
does-text-donation-work/
von Hippel, E. (2006). Democratizing innovation. Cambridge, MA, USA: MIT Press.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 225
Walsham, G. (1997). Actor-network theory and IS research: Current status and future prospects. In
Lee, A., Liebenau, J., DeGross, J. (Eds.) Information Systems and Qualitative Research. 466-
480. New York, NY, USA: Springer US. 10.1007/978-0-387-35309-8_23
Walton, M. & Donner, J. (2012). Public access, private mobile: The interplay of shared access and the
mobile Internet for teenagers in Cape Town. Global Impact Study Research Report Series. Cape
Town: University of Cape Town. Retrieved November 21, 2012 from
http://library.globalimpactstudy.org/sites/default/files/docs/Public%20access,%20private
%20mobile%20final.pdf
Waverman, Leonard, Meschi, M., & Fuss, M. (2005). The impact of telecoms on economic growth in
developing countries. The Vodafone Policy Paper Series, 2(03), 10–24. Retrieved from
http://web.si.umich.edu/tprc/papers/2005/450/L Waverman- Telecoms Growth in Dev.
Countries.pdf
Williams, R., & Edge, D. (1996). The social shaping of technology. Research Policy, 25(6), 865-899.
Wilson, C., & Dunn, A. (2011). Digital media in the Egyptian revolution: descriptive analysis from the
Tahir data set. International Journal of Communication, 5(2011). 1248–1272.
World Bank (2012). Information and Communications for Development 2012: Maximizing Mobile.
Washington, DC, USA: World Bank Publications. Retrieved from
http://siteresources.worldbank.org/ACM.EXTINFORMATIONANDCOMMUNICATIONAND
TECHNOLOGIES/Resources/IC4D-2012-Report.pdf
World Wide Worx. (2012, May 10). Mobile pushes Internet to the masses. Retrieved from
http://www.worldwideworx.com/mobile-pushes-internet-to-the-masses/
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 226
Wyche, S. P., Forte, A., & Schoenebeck, S. (2013). Hustling online: understanding consolidated
facebook use in an informal settlement in Nairobi. In Proceedings of the SIGCHI conference on
human factors in computing systems. 2823-2832. ACM.
Zainudeen, A., & Ratnadiwakara, D. (2011). Are the Poor Stuck in Voice? Conditions for Adoption of
More-Than-Voice Mobile Services. Information Technologies & International Development,
7(3), 47-59.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 227
Appendices
Appendix A: Interview topic guide
Introduction
I'm Melissa Loudon, I'm a PhD student at the University of Southern California, and before that I spent
some time as a mobile developer working on projects in clinics and in rural communities in the Eastern
Cape [Cell-Life and UCT if appropriate].
My research is on mobile apps and services and development/the BoP. Thanks very much for agreeing
to do the interview.
How much time do you have? My questions should take less than 40 minutes, but I can work around
your schedule.
The interviews are confidential and I will only share anonymous de-identified quotes, except with your
permission. Also, of course you don't need to answer any questions you are not comfortable with; just
let me know and I'll skip them.
Finally, would it be ok if I use a voice recorder / record this interview? The recording is confidential
and will not be shared.
• Can you tell me a bit about what you do at [app/service], and how you came to be involved?
App/service overview
• What is [app/service]? how would you describe it?
• Who is involved? developers/operations/support?
• What is the geographic reach? Just South Africa, or other countries as well?
Users
• Who are your users?
• How do they benefit from the app of service?
• Are the users the same people you imagined when you were designing the app/service? Are
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 228
there groups you haven’t yet reached, or groups whose enthusiasm has been unexpected?
• Are they using it in ways you didn’t anticipate?
History
• How did the idea for [app/service] come about?
• How did you get to the point where you were ready to build something?
• What was the process like? did you do market research, run a pilot, speak to people doing
similar things?
• Who developed the app/service?
• What were the key specifications in the initial design? did it need to work on particular
hardware or accommodate particular kinds of users e.g. low-literacy users or users without a
formal address?
Technology choices
• Why did you choose [technology/platform]?
• Did you encounter any technology challenges? things that perhaps people don’t think about
when planning/designing?
• Were there different technology considerations in different countries?
• Did you use any existing tech components built by other people?
• Where did you find documentation about the platform? where did you find support?
Partnerships
• Have other organizations been involved in design, development, pilot/testing? Are the still
involved?
• How did [specific relationships] come about
• Do you have any operator involvement? if so, how did that come about? if not, did you seek
assistance from an operator?
Evaluation
• Have you done any form of evaluation for [app or service]? do you have any results you could
share?
• What would you do differently in hindsight?
• What went really well?
• What changes in the mobile ecosystem would make conditions more favorable for your
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 229
app/service, or BoP/development-related apps/services in general?
Future plans
• What’s next for [app or service]? are you working on new features or other projects?
• What do you think will happen as smartphone market share increases? will it change your
technology choices?
General questions about the field
• What do you think are the biggest challenges for mobile apps for the BoP/development?
• What kinds of apps/services are most likely to succeed (and why)? Is there ‘low-hanging fruit’
you’d like to see more people working on?
• What problems does the move towards smartphones potentially solve for your users? do they
introduce new problems?
• Do you think we’ll see regulators trying to take a more active role in the apps and services
space?
• Do you think development funders will continue to invest in BoP mobile apps? do you think
they should?
• What about government? is this something you think they should be taking on? have they
shown interest?
• Do you think private investors will continue to invest in BoP mobile apps and services? what
kinds of apps and services are most attractive to them?
• What do you think BoP users are most interested in?
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 230
Appendix B: Inventory dimensions
Dimension Categories Explanation Notes
App/service name free text
Description free text
Developers free text who designed
and/or built the
app or service?
Implementing
organization
free text who is operating
it?
Other partners free text
Intended users • Employees of a service
provider (e.g. community
health workers)
• Members of an organization
• Beneficiaries or customers
of an organization
• Particular marginalized
group
• Income- and asset- poor
individuals
• General public
Actual users free text What
demographic
information do
we have about
the app or
service’s current
users? For
example, are
they mostly in a
particular age
group, income
bracket, gender?
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 231
Development
'theory of change'
free text How is the app
or service
envisaged
(implicitly or
explicitly) to
contribute to
development?
Development
'theory of change'
category
• Livelihood support
• Access to formal services
• Social network maintenance
• Collective
action/’crowdsourcing'
Revenue model • Donor-supported
• Government service
• Advertising-supported
• User pays
• Organization pays on behalf
of user
Cost to user • Free
• Once-off charge
• Subscription
Languages • Afrikaans
• English
• Xhosa
• Zulu
• Swati
• Tswana
• Sotho
• Tsonga
• Northern Sotho
• Venda
• Southern Ndebele
• Other
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 232
Technologies
used - mobile
platforms
• SMS
• MMS
• USSD
• Voice
• Mobile Web
• Mxit
• Mobile java apps (non-
smartphone)
• Mobile apps (smartphone)
Of these, SMS,
USSD, mobile
web and non-
smartphone
mobile java apps
will be
extensively
described. MMS
is less common
and Mxit is
uncommon
outside South
Africa.
Smartphone
apps have a
short history and
smartphones are
not yet in wide
use among low-
income
consumers
(Research ICT
Africa 20102)
SMS platform
features used
• Zero-rated line
• Premium SMS
• Incoming SMS
• Outgoing SMS
• ‘Please call me’ messages
• Missed calls
• WASP involved
USSD platform
features used
• Zero-rated USSD band
• Premium USSD band
• Language selection step
• WASP involved
Mobile Web
platform features
used
• Accessed through an
operator-controlled portal
• Zero-rated on all operators
• Zero-rated on some
operators
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 233
Mobile java apps
platform features
used
• Available for download
online
• Requires specific phone
model (or limited list of
models)
• Signed
• Requires data connectivity
Mobile operator
involvement
• No
• One (dominant operator)
• One (non-dominant
operator)
• Multiple operators
• All operators
Mobile operator
involvement type
• Zero-rated services
• Technical
assistance/preferential access
• Funding only
Launch date month/year
End date (if
ended)
month/year
Status at end date
(if ended)
• Pre-pilot
• Pilot
• Live
Current status • Pre-pilot
• Pilot
• Live
• Ended
Available
measures of
success
• User numbers
• Revenue
• Awards and accolades
Successful? • No, failed
• Too early to say
• Sustainability issues
• Yes, successful
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 234
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 235
Appendix C: Apps and services in inventory
Name Description Sector Archetype
Mobile
platforms
Angel
Counselling on Mxit, mostly for stress/depression and substance
abuse. youth social issues counselling Mxit
Baby & Pregnancy
Info
Mxit app providing daily advice and information messages for
pregnant women or new mothers, tied to stage of
pregnancy/baby's age. health
info services,
constant contact Mxit
biNu
Combines an 'app platform' environment with airtime and content
rewards earned through surveys. App platform renders on server
and compresses response to user, so supports most J2ME phones
(and Android). Rewards can be 'cashed in' for SMS bundles,
airtime, music and other digital content income generation
surveys/market
research
J2ME app,
Smartphone
app
Bookly
App serves a collection of e-books and short pieces on mxit and a
mobi site. Some free books through Project Gutenberg, some
paid (per chapter up to around $3 for a whole book) through
publishers. Also a creative writing competition and user-
submitted content. education
reading/literacy
materials
Mobile Web,
Mxit
Brainwave Careers
Career and further education information, study tips, career quiz,
Mon - Thurs 12pm to 6pm live chat. employment
further education
& training Mxit
Bsmrt
Mobile learning mxit app and mobi site, including quizzes and
reference material. education
general
education &
games
Mxit, J2ME
app
Career Planet
Web and mobi site with extensive listings of jobs, bursary,
internship and training opportunities. Also accessible via USSD
menu. Career Planet is an NPO. employment
further education
& training
USSD,
Mobile Web
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 236
Cell-Life Capture
Mobile data collection. Based on OpenXData. Mostly in the
health sector, applications include M&E, home-based care. health
data collection /
remote workers J2ME app
Cell-Life
Communicate
SMS, USSD, please call me messages for healthcare, particular
emphasis on HIV/AIDS management applications. Used for
several research studies of SMS in healthcare, including medical
abortion coaching, encouraging HIV counselling and testing. health constant contact SMS, USSD
Childline Childline counsellors on Mxit. Part of Childline's crisis helpline,
based out of Durban office.
youth social issues counselling Mxit
Choma
Mobile magazine' targeted at young women 15-25. Mix of
entertainment and sexual health content. Mobi site, Facebook
page, Mxit portal. health info services
Mobile Web,
Mxit
CommCare
Mobile data collections, remote worker management, case
management. SaaS model with web-based admin/reporting
interface. Mostly used by mhealth apps. Android and J2ME, as
well as bespoke SMS options.
health
data collection /
remote workers
SMS, Mobile
Web, J2ME
app,
Smartphone
app
Corruption Watch
Mxit portal for Corruption Watch NGO. Users can read news and
get information about reporting corruption, and submit reports
through the site. youth social issues
civic
participation
Mxit
D-Siders Entertainment content - a Mxit/mobi site soap opera, text-based,
with user interaction via comments and voting
mass reach Mobile Web,
Mxit
EverEgg Paid eBooks on Mxit, sold by the chapter. education reading/literacy
materials
Mxit
Everything
Maths/Everything
Science
Grade 10-12 maths and science textbooks on mxit. Textbooks are
cc-licensed and curriculum-aligned. A paid web-based practice
service is the revenue source. education
curricula &
practice
Mxit
FunDza
Literacy non-profit publishing short stories, novels on Mxit, biNu
and mobi site. Also currently running a short story contest with
Mxit Reach. education
reading/literacy
materials
Mobile Web,
Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 237
Groovy Adventurers
Child development activities on Mxit. Activities help parents
assess a child's developmental milestones through play.
education
general
education &
games Mxit
H360
Mxit app and mobi site with HIV information, Q&A, stories and
quizzes. Every1Mobile is a for-profit company targeting African
youth. Commercial rather than development funding, also
supported by advertising.
health info services
Mobile Web,
Mxit, J2ME
app,
Smartphone
app
Harambee
Mobi site where youth 18-28 with a matric can apply to a
government-funded training and job placement programme.
Applications only via the mobi site or by phone. employment
further education
& training
Mobile Web
Health4Men Connect Mobi site with sexual health information targeted at MSM. Also
question and answer service with Health4Men doctors.
health info services SMS, Mobile
Web
HealthSmart Safer sex and HIV education messaging for MSM, connected to a
Columbia University ICAP project
health constant contact SMS
Hello Doctor Health information, articles and health-related group chat with
medical doctors
health info services Mxit
Hi4Life
Health information services for women and men 15-24, including
sexual health, pregnancy, HIV/TB.
health info services
SMS, USSD,
Mobile Web,
Mxit
ichoosewhen.org Mobi site and Mxit portal with family planning information health info services Mobile Web,
Mxit
ikapadata
Bi-weekly and bi-monthly township surveys. Data collectors are
trained, paid and provided with a phone with ODK installed. income generation
surveys/market
research
Smartphone
app
Imbizo men's health SMS testing and sexual health messaging for men in Imbizo, a
programme based in Soweto.
health info services SMS
Instant Africa
Mobi site that pays users - in airtime, shopping vouchers, or bank
account transfers - for conducting marketing research surveys
among their peers. income generation
surveys/market
research
Mobile Web
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 238
Just Tested for HIV
People who have just been tested for HIV can sign up for a series
of followup SMSs. health constant contact SMS
loveLifeMX
Text-based counselling as well as sexual health information,
monthly magazine and other 'edutainment' content. youth social issues
counselling, info
services Mxit
Lungisa
Users can report issues such as transportation failures, blocked
drains and broken pipes. The issue reports are forwarded to the
appropriate local authority. crowdsourcing
SMS, Mobile
Web, Mxit
MAMA South Africa
A mobi site with articles, stories and static content for women
who are pregnant or have an infant (<1 year). Users can also ask
questions. In addition, an SMS service provides subscribers with
timed informational SMSs for their pregnancy and the first year
of the baby's life. There are also weekly USSD quizzes. health
info services,
constant contact
SMS, USSD,
Mobile Web,
Mxit
mCent
Mobi site where users can earn airtime by taking market research
surveys and watching/sharing brand content. Also manages large-
scale airtime rewards programmes for brands. income generation
surveys/market
research Mobile Web
Medical Abortion
Support
Medical abortion information messages, and a three week
programme of supporting messages for women undergoing
medical abortion. Initially a WHO-funded research study, now
being included in IPAS's medical abortion rollout. health constant contact SMS
miGOx
Opinion polls and comments about government services, on
Mxit. youth social issues
civic
participation Mxit
Mindset Xtra Live
Access to Mindset Learn's Q&A service from Mxit. Users can
submit and browse questions. education
curricula &
practice Mxit
Mobenzi Intelligence
Mobile microwork platform built on Mobenzi Researcher. Tested
with sentiment analysis tasks, payment in airtime. income generation
surveys/market
research J2ME app
Mobenzi Outreach
Data collection and reporting for community health workers.
Includes facility appointment management, SMS reminders,
report generation, web interface. health
data collection /
remote workers,
constant contact
SMS, J2ME
app,
Smartphone
app
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 239
Mobenzi Researcher
Mobile data collection for J2ME and Android. Mobenzi provides
support and customization, including translating surveys to
phones, managing airtime, training survey workers. Web
interface and API, but no source code. health
data collection /
remote workers
J2ME app,
Smartphone
app
MobieG Counselling on Mxit and PC live chat youth social issues counselling Mxit
MobiSchool Short curriculum-aligned video lessons and tutoring on Mxit.
Grade 10-12, six subjects. Also exam-time live tutoring.
education curricula &
practice
Mobile Web,
Mxit
MoMConnect
Links antenatal and postnatal information/reminder SMSs to an
EMR that tracks pregnant women attending health facilities in
two districts: Umgungundlovu and Ethekwini, both in KwaZulu
Natal. health
info services,
constant contact SMS
Moraba
Smartphone (Android and iPhone) and Java game combining
Morabaraba with quiz questions on gender-based violence and
consent. youth social issues
general
education &
games
J2ME app,
Smartphone
app
mySaftey
Portal to safety-related mxit apps, including Rape Crisis, MAC
911, Pink Ladies, Angel, Childline. Intended for school-age
children - Department of Basic Education is a partner. youth social issues info services Mxit
Nal'ibali
Mxit and mobi site for an NGO that promotes childrens' reading
and reading aloud. Stories in several languages, reading aloud
tips, activities for reading clubs. education
reading/literacy
materials
Mobile Web,
Mxit
Nokia MoMaths Mobi site with maths exercises and theory for grades 10-12. education curricula &
practice
Mobile Web,
Mxit
Oxford Word-of-the-
Day Daily English word and definition education
general
education &
games Mxit
Periodic Table
Periodic table of the elements on Mxit. Appears to be a Mxit
Reach demo project. Content from Wikipedia.
education
general
education &
games Mxit
Pondering Panda Market research surveys on Mxit. income generation surveys/market
research
Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 240
Project Masiluleke
Used the advertising portion of please call me messages to
encourage people to call the National AIDS helpline. Message
content from iTeach, advertising space free from MTN (5% of all
daily please call me messages). health mass reach SMS
QuizMax
Multiple choice subject quizzes for Grade 12 learners in Maths
and Science. education
curricula &
practice
Mobile Web,
Mxit
Rape Crisis Information for people who have been raped. youth social issues info services Mxit
Rate my Clinic
Patients at health facilities answer a short survey on their
experience. health crowdsourcing USSD
Red HIV /
RedChatZone
Mobi site and Mxit portal for the National AIDS helpline.
Information and referrals to other helplines, as well as mxit
counselling chat service. health
info services,
counselling
Mobile Web,
Mxit
Rethink Education
High school curriculum content presented chat-style, with
exercises, as a Mxit app. 'Freemium' model - some schools pay
for the web version of the platform, along with learner tracking.
Mxit version is free. education
curricula &
practice Mxit
SIMmed
Chronic medication compliance. User dials a number on speed
dial when they take their medication each day. If the server
doesn't receive the user's checkin, reminder SMSs are sent to
user. SMS reminders may also be escalated to a designated
family member or friend. health pill containers SMS
SIMpill
Pill container with GSM module. Sends a checkin to server when
container is opened; reminders to patient and caregiver when
pills are not taken. health pill containers SMS
smartSex
Information, Q&A, forums and quizzes about sex and
relationships. health info services
Mobile Web,
Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 241
Student Village
Further education, job search and graduate recruitment
programme contact information and articles. education
further education
& training, job
search
Mobile Web,
Mxit
TB Free SMS Adherence support messaging for TB patients health constant contact SMS
Teachers' Corner
Mxit app for teachers. Subject-specific and general classroom
management information. education
curricula &
practice Mxit
TxtAlert
SMS appointment reminders and ability to initiate appointment
rescheduling (clinic calls back) via please call me messages. health constant contact SMS
Ummeli
Jobs and networking portal on YoungAfricaLive. "Linkedin for
the BoP". employment job search Mobile Web
Uusi
Mxit app and mobi site (uusi.me) - job search and job seeker
profiles. Job search seems to be populated from gumtree feed.
Mxit app also has a few articles on getting a job, and a link to
Instant Africa, which pays in airtime for surveys. employment job search Mxit
VIP:Voice
Election news and info targeted at young voters, as well as
comments/discussion and citizen monitoring on election day. Part
of a study run by Praekelt for Democracy International and UC
San Diego. youth social issues
civic
participation
USSD,
Mobile Web,
Mxit
Vodacom mHealth
Commmunity Care
Home-based care tracking - J2ME/Android app and associated
server-side data storage and reporting by Vodacom mHealth
(formerly Mezzanine, formerly GeoMed). health
data collection /
remote workers
J2ME app,
Smartphone
app
Vota Mzansi
Mxit app with information about the voting process for the 2014
elections. Also collects some data - intention to vote,
understanding of basic political process. youth social issues
civic
participation Mxit
WisePill
Pill container that triggers a checkin when the pillbox is opened.
If not opened within window, reminder SMSs to user and after 48
hours, caregiver. Online tracking of adherence data. health pill containers SMS
Young Mom Support
Information and live chat about teenage pregnancy. Linked to a
Cape Town -based support group for teen moms. youth social issues info services Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 242
YoungAfricaLive
Mobi site regularly updated with dynamic 'edutainment' and
entertainment content, and static content providing information
on HIV/AIDS and related topics. health info services Mobile Web
Yoza Cellphone
Stories
Short 'm-novels' accessible on a mobi site and through mxit.
Users can comment/discuss. Initial content was developed
specifically for Yoza project, deals with social issues. Now also
other content - school setworks, a small Afrikaans publisher. education
reading/literacy
materials
Mobile Web,
Mxit
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 243
Appendix D: CommCare data analysis code book
Variables used in models
Variable Values Definition Derivation
Number of
visits in
month
(nvisits)
Count of
visits by a
CommCare
user in a
single month
A visit is defined as a set of
forms completed
consecutively (within 12
hours) by the same
CommCare user, about a
single case or a set of related
cases, such as a mother and
child.
Form submissions
Short-term
retained
(active_diff
for input
values 3,2)
True/False A user is considered short-
term retained if they submit
data in their second month on
CommCare and also three
calendar months later i.e. in
their fifth month since starting
on CommCare
Form submissions
Long-term
retained
(active_diff
for input
values 6,6)
True/False A user is considered long-term
retained if they submit data in
their sixth month on
CommCare and also six
calendar months later i.e. in
their twelfth month since
starting on CommCare
Form submissions
Device type
(summary_de
vice_type)
Nokia/Androi
d
Version of CommCare app
used to submit forms during a
single month.
Form submissions. Most
users use either Nokia or
Android phones with the
corresponding app version;
users who submitted in other
ways (for e.g. via the web,
via SMS) or with multiple
different app versions in a
single month are rare and
were excluded.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 244
Self-started
(self_start)
True/False Self-started projects were
initiated without support from
Dimagi staff. Dimagi-assisted
projects have a self-started
value of False.
Project attributes. Dimagi-
assisted projects are flagged
on initiation by Dimagi staff.
Project
(project)
Project name
(anonymized)
A project is the main unit of
management for CommCare
deployments. Reporting,
billing and user management
happen at project level.
Project attributes, form
submissions.
Organization
(org)
Organization
prefix
(anonymized)
An organization groups
projects with the same project
owner. While most
organizations are involved
with only a single CommCare
project, several large
organizations have several
projects - for example, an
RCT with sites in different
countries.
Project attributes. Data
linking multiple projects to a
single owner is maintained
for billing purposes, but not
linked directly to operational
data. In the absence of a
direct link, I exploited an
unofficial naming convention
for projects to derive
organization membership,
such that projects with a
hyphenated name i.e. org-
project were assigned to an
organization named for the
part before the hyphen.
Variables used in data cleaning
Variable Values Definition Derivation
Test project
(test)
True/False True if project contains
training or test data and
should be excluded from
analysis
Project attribute. Flag set on
project initiation.
Data use
permitted
(can_use_dat
a)
True/False True if project owner has
consented to secondary use of
project metadata
Project attribute. Flag set on
project initiation.
running head: MOBILE APPS AND SERVICES FOR DEVELOPMENT 245
User ID
(user_id)
Alphanumeric Either a unique identifier for a
user, or a special value
indicating a test/demo user
with fake data
User attributes
User type
(user_type)
Web/Mobile Web user accounts can access
the web interface, while
mobile user accounts can
submit data through the
mobile app.
User attributes
Device type
(summary_de
vice_type)
Nokia/Androi
d
Version of CommCare app
used to submit forms during a
single month.
Form submissions. Most
users use either Nokia or
Android phones with the
corresponding app version;
users who submitted in other
ways (for e.g. via the web,
via SMS) or with multiple
different app versions in a
single month are rare and
were excluded.
Numeric
index of
month
(numeric_ind
ex)
Numeric The numeric index is defined
for each user, for each month.
It is the number of months
since the month of that user’s
first form submission.
Form submissions

Abstract (if available)

Abstract This project aims to describe and explain the field of mobile apps and services for development—'m4d' apps and services—primarily by studying its realisation in South Africa during the pre-smartphone era. Concepts from infrastructure studies (Star, 199) and platform studies (Montfort & Bogost, 2009) are employed to describe the relationship between technology platforms and the apps and services built on them. Research activities include producing a historical and technical account of the mobile network and platforms for m4d apps and services, taking inventory of m4d apps and services in South Africa and categorizing them according to various dimensions of interest, conducting and analysing unstructured interviews with developers and designers of m4d apps and services, and carrying out an exploratory quantitative analysis of secondary data generated through use of a mobile data collection app. The findings demonstrate that claims about the potential for m4d apps and services should be understood in the context of embodied infrastructural limitations, arising from the gap between design context and the use contexts in the developing world. The result is a field with limited diversity, with designs for m4d apps and services reducible to a relatively small number of archetypes. Alignment with actants in the global network for m4d apps and services, including funders, mobile network operators and ""circuits of truth"" (Roy 2010, location 123) such as the GSMA's Mobiles for Development Impact programme produces additional incentives for concentration. Finally, the discursive construction of the m4d consensus as it relates to apps and services is not neutral. Whatever benefits accrue to users in the developing world, frames that position technologies as offering solutions to development problems are designed to achieve favourable public perception, and by extension favourable regulatory treatment, for technology industry interests.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Mobile communication and development: a study of mobile phone appropriation in Ghana

PDF

Mapping out the transition toward information societies: social nature, growth, and policies

PDF

Voice, participation & technology in India: communication technology for development in the mobile era

PDF

Building a global information society: financing and regulating information and communications technologies for development (ICT4D) in sub-Saharan Africa

PDF

Participatory public culture and youth citizenship in the digital age: the Medellín model

PDF

The use of mobile technology and mobile applications as the next paradigm in development: can it be a game-changer in development for women in rural Afghanistan?

PDF

Philly Bike Report: a mobile app for mapping and sharing real-time reports of illegally blocked bike lanes in Philadelphia

PDF

Examining internal communication practices to improve teachers’ motivation to use ICTs in instruction: a case study of Peruvian secondary schools

PDF

The next-generation news application: Stories in context

PDF

Evaluating the utility of a geographic information systems-based mobility model in search and rescue operations

PDF

The role of GIS in asset management: County of Kauai Department of Parks and Recreation a need for an asset management program

PDF

Exploiting side information for link setup and maintenance in next generation wireless networks

PDF

Essays on information design for online retailers and social networks

Asset Metadata

Creator Loudon, Melissa (author)

Core Title Mobile apps and services for development: What can we learn from the non-smartphone era in ICT4D?

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 09/17/2015

Defense Date 08/25/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag ICT4D,information and communication technologies,infrastructure studies,international development,mobile apps,mobile phones,OAI-PMH Harvest,platform studies,South Africa

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Bar, François (committee chair), Annany, Mike (committee member), Castells, Manuel (committee member)

Creator Email melissa.loudon@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-184528

Unique identifier UC11275432

Identifier etd-LoudonMeli-3926.pdf (filename),usctheses-c40-184528 (legacy record id)

Legacy Identifier etd-LoudonMeli-3926.pdf

Dmrecord 184528

Document Type Dissertation

Format application/pdf (imt)

Rights Loudon, Melissa

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA