ArticlePDF Available

Consumption of Bottled Water at the Bottom of the Pyramid: Who Purchases First?


Abstract and Figures

While consumer and marketing research in developed markets is an established field, research on consumers in an Bottom of the Pyramid (BoP) setting is less established and mostly conceptual or qualitative. This paper examines the individual heterogeneity and the local context of BoP consumers with an empirical study on consumption of low cost bottled water on the Kenyan coast and the capitals of Uganda and Rwanda. The empirical analysis builds on existing research exploring consumer behavior, and it studies a database of 713 bottled water consumers in Kenya, Uganda and Rwanda. Consumers with a higher level of education were less likely to be late consumers. Additionally, early consumers were more likely to purchase due to a purposeful search for a bottled drinking water solution. Since we control for location specific effects we highlight the importance of supply driven consumption in the BoP market. Furthermore, the results suggest that the two water companies may not be reaching their targeted low-income consumers but rather middle class consumers. The research contributes to the larger BoP debate by presenting evidence that consumers in a BoP setting may purchase more on the basis of supply of a product rather than other socio-demographic factors such as income.
Content may be subject to copyright.
Consumption of Bottled Water at the Bottom
of the Pyramid: Who Purchases First?
Rachel Howell,
Kinsuk Mani Sinha,
Natascha Wagner,
Neelke Doorn,
and Cees van Beers
While consumer and marketing research in developed markets is an established field, research on consumers in an
Bottom of the Pyramid (BoP) setting is less established and mostly conceptual or qualitative. This paper examines the
individual heterogeneity and the local context of BoP consumers with an empirical study on consumption of low cost
bottled water on the Kenyan coast and the capitals of Uganda and Rwanda. The empirical analysis builds on existing
research exploring consumer behavior, and it studies a database of 713 bottled water consumers in Kenya, Uganda and
Rwanda. Consumers with a higher level of education were less likely to be late consumers. Additionally, early consumers
were more likely to purchase due to a purposeful search for a bottled drinking water solution. Since we control for
location specific effects we highlight the importance of supply driven consumption in the BoP market. Furthermore, the
results suggest that the two water companies may not be reaching their targeted low-income consumers but rather
middle class consumers. The research contributes to the larger BoP debate by presenting evidence that consumers in a
BoP setting may purchase more on the basis of supply of a product rather than other socio-demographic factors such
as income.
emerging markets, consumption, BoP consumers, bottled water, sub-Saharan Africa
While the last decade of research brought attention to the
potential for businesses reaching the large and untapped market
of low income consumers, it is debated whether businesses can
be both profitable and meet needs for low income consumers
(Agnihotri 2012; Karnani 2007; Meagher 2018; Prahalad 2005;
Ramani and Mukherjee 2014). In order for businesses offering
Bottom-of-the-Pyramid (BoP) innovations to be profitable and
self-sustainable their innovations need to be widely adopted
and consumed (Hall, Matos, and Martin 2014; Ramani and
Mukherjee 2014). Although research examining marketing and
consumption habits in developed markets is well established,
consumers and their consumption decisions in the BoP is a less
studied yet relevant focus as increasingly western firms target
BoP markets (Agnihotri 2012; Kotler et al. 2006; Sheth 2011;
Silvestre and Neto 2014). The seminal work of Prahalad and
others (Montgomery, Peredo, and Carlson 2012; Prahalad
2005) implies that consumers at the BoP are homogenous,
however more recent research points towards a more hetero-
geneous group of BoP consumers (Chikweche, Stanton, and
Fletcher 2012; Guarı´n and Knorringa 2014; Kotler et al.
2006; Marinakis, Walsh Prof dr, and Harms dr 2016). In par-
ticular, much of the BoP literature focuses on the Asian context
with less work looking at African markets and consumers (Mair
and Marti 2009; Prahalad 2012; Ramani, SadreGhazi, and Duy-
sters 2012). However, as a recent African Development Bank
report highlighted, GDP growth in sub-Saharan Africa has out-
paced other regions in the last 15 years, creating opportunities
for market development (AfDB 2016). This paper contributes
empirical evidence to the debate on BoP consumer hetero-
geneity with a focus on the African context, by investigating
characteristics of BoP customers who consume fresh drinking
water in Kenya, Uganda and Rwanda.
While marketing literature has increasingly looked at new
markets in Asia and Africa, focusing on the potential for busi-
ness, there is still limited understanding of how markets and
Faculty of Technology, Policy, and Management; Delft University of Technology,
Delft, Netherlands
School of Social Sciences; Wageningen University, Wageningen, Netherlands
Institute of Social Studies; Erasmus University Rotterdam, The Hague,
Corresponding Author:
Rachel Howell, Faculty of Technology, Policy, and Management, Delft
University of Technology, Jaffalaan 5, Delft, Netherlands.
Journal of Macromarketing
ªThe Author(s) 2019
Article reuse guidelines:
DOI: 10.1177/0276146719866890
consumers in these markets differ from developed markets
(Kotler et al. 2006; Sheth 2011). Authors have highlighted the
issue of poverty changing the way consumers make purchase
decisions (Bonsu and Polsa 2011; Chikweche and Fletcher
2011a). Additionally, other authors have criticized businesses
seeking to reach poor consumers, arguing that they are often
not being reached and may instead be exploited (Hahn 2009;
Karnani 2007, 2009; Meagher 2018) . Developing a better
understanding of these aspects requires a deeper look into how
consumers at the BoP are heterogeneous and how this influ-
ences the timing of their consumption decisions. Possible ways
to understand consumers’ consumption behavior is to analyze
individual characteristics, the local context in which the con-
sumer is embedded and finally, the cultural context that shapes
the local understanding (Ramani, SadreGhazi, and Duysters
2012; Zanello et al. 2016). Despite the seeming resemblance,
the various factors affecting the decision process of a BoP
consumer are considerably different than the traditional con-
sumers. One way to look at consumer heterogeneity is through
timing of purchase and how BoP consumers may differ based
on their time of purchase. Timing is relevant due to the severe
budget and income constraints BoP consumers face that affect
timing of purchase (Viswanathan et al. 2014; Yurdakul, Atik,
and Dholakia 2017).
The study at hand contributes to the scarce literature on BoP
consumer heterogeneity by assessing the characteristics of con-
sumers based on their time of purchase and whether there are
systematic differences between groups. Two types of charac-
teristics are analyzed – individual characteristics such as edu-
cation and income, and features of the local context such as
how the consumer received information about a product. The
following research question is addressed: What is the effect of
BoP consumer characteristics on their timing of consumption?
The research contributes to BoP marketing literature in two
ways: 1) it exposes the heterogeneity amongst BoP consumers
by focusing on timing of consumption, 2) quantitative study on
consumers in three countries. The context under study is the
consumption of bottled water – of two different socially
oriented companies (Dutch Water Limited and Jibu, LC3) in
urban East Africa (Kenya, Rwanda, Uganda). The first is a
Dutch founded water company on the Kenyan coast and the
second an American founded “social enterprise” in the capitals
of Rwanda and Uganda. We study 713 bottled water consumers
breaking them apart by their time of purchase. This research
aims to address the gap in empirical work at the BoP by
focusing on differences between consumers based on their
time of purchase and not at differences between consumers
versus non-consumers, thus exposing BoP consumer hetero-
geneity. The study shows that late consumers have a lower
level of education whereas earlier consumers are more likely
to have purposefully searched for bottled drinking water. A
novel feature of our analysis is that we can control for entry
into different sales areas/zones. Controlling for location spe-
cific effects we demonstrate that consumption at the BoP is
highly supply driven. Most importantly, the research suggests
that unlike traditional consumers in developed markets,
consumption of bottled water for BoP consumers does not
seem to be influenced by wealth.
The remainder of this paper is structured as follows:
Section 2 provides an overview of existing literature on
macromarketing at the BoP, BoP consumer studies and
bottled water consumption. Section 3 presents the conceptual
model. The study context is set up in section 4 and the vari-
ables and empirical model are detailed in Section 5. Results
are presented in section 6 and the last two sections provide the
discussion and conclusions.
Literature Review
Macromarketing at the BoP
The advent of three changes have brought increasing attention
to the importance of business in emerging markets and the BoP.
First, besides Prahalad’s initial opening of the possibilities at
the large and untapped BoP market, growth in emerging mar-
kets has occurred due to economic reforms in these countries.
Second, due to the limited demand in advanced economies
because of aging populations, BoP markets have been brought
to the forefront. Finally, the advent of free trade and a rising
new middle class have increased business opportunities at the
BoP (Sheth 2011). Therefore, understanding how marketing
and consumer behavior is different in BoP markets has become
a growing area of research (Bonsu and Polsa 2011; Davies and
Torrents 2017; Jagadale, Kadirov, and Chakraborty 2018; Jais-
wal and Gupta 2015; Kotler et al. 2006; Yurdakul, Atik, and
Dholakia 2017). Besides the business opportunities in these
new markets, there is the question of how marketing to these
new demographics plays a role in poverty alleviation (Kotler
et al. 2006). Therefore marketing literature on the BoP can be
divided into two streams: case studies on specific markets (Alur
and Schoormans 2013; Banbury, Herkenhoff, and Subrahman-
yan 2015; Chikweche and Fletcher 2011a; Davies and Torrents
2017; Jagadale, Kadirov, and Chakraborty 2018; Viswanathan
et al. 2014; Viswanathan and Rosa 2010), and how marketing
strategies and consumer behavior more generally might differ
in these markets (Agnihotri 2012; Bonsu and Polsa 2011;
Kotler et al. 2006; Piacentini and Hamilton 2013; Sheth
2011; Subrahmanyan and Tomas Gomez-Arias 2008; Yurda-
kul, Atik, and Dholakia 2017).
Focusing on the second stream of literature within market-
ing and consumer studies, highlights a few points about oper-
ating within BoP markets. First, authors have illuminated the
issue of heterogeneity at the BoP both in terms of markets and
consumers (Kotler et al. 2006; Sheth 2011). Prahalad’s work
overlooked the heterogeneity of this demographic (Agnihotri
2012; Kotler et al. 2006). Rather than being one homogenous
group with similar purchasing power, BoP consumers are dri-
ven to purchase products for different reasons and have differ-
ent socio-demographic characteristics that drive them to make
purchase decisions (Jagadale, Kadirov, and Chakraborty 2018;
Sheth 2011). In addition to consumer heterogeneity, markets
themselves are heterogeneous in the sense that one particular
2Journal of Macromarketing XX(X)
geographically demarcated market can be made up of consu-
mers that have means to purchase and those who do not (Sheth
2011). Second, marketing scholars emphasize that poverty lim-
its consumers from being part of a consumer society (Yurdakul,
Atik, and Dholakia 2017). Consumption decisions are made
less from the perspective of choices but more from immediate
needs which challenges Prahalad’s notion of freedom of choice
and individual empowerment through marketplace participa-
tion (Bonsu and Polsa 2011; Viswanathan et al. 2014). Related
to the issue of participation in a consumer society, other
research has exposed that purchase decisions may be driven
more by availability of products than specific desires (Chik-
weche and Fletcher 2011b). Finally, studies that have focused
on specific emerging/BoP markets have illustrated the lack of
branding in many BoP markets (Sheth 2011) and the problem
of distribution and the importance of social networks and local
entrepreneurs in overcoming this problem (Alur and Schoor-
mans 2013; Chikweche and Fletcher 2011b; Davies and
Torrents 2017).
Existing marketing literature on BoP markets exposes two
areas for future research: unpacking consumer heterogeneity in
BoP markets from an empirical perspective, and understanding
better what drives consumers to purchase particular products,
especially products that could be considered basic needs like
clean drinking water (Kotler et al. 2006; Sheth 2011; Yurdakul,
Atik, and Dholakia 2017). This paper will explore consumer
heterogeneity from a marketing perspective.
Consumer Studies at the BoP
BoP literature following Prahalad’s seminal work, emphasized
both the need to define the poor and acknowledge their hetero-
geneity (Agnihotri 2012; Karnani 2007; Kotler et al. 2006).
Investigating these two aspects requires a look at consumption
at the BoP and what drives BoP consumers to make purchasing
decisions. The literature on consumers and purchasing deci-
sions at the BoP is mostly dominated by qualitative work
(Chikweche, Stanton, and Fletcher 2012; Nakata and Weidner
2012; Ramani, SadreGhazi, and Duysters 2012; Subrahmanyan
and Tomas Gomez-Arias 2008) with fewer quantitative inves-
tigations (Ernst et al. 2015). While the focus of this research is
on consumption of goods not adoption of innovations, there is
some overlap in factors driving these decisions. Adoption
implies a time dimension and looking at the overall market,
whereas consumption is an instantaneous purchase decision.
Therefore, an overview from both types of literature is given.
First an overview of qualitative work on adoption of innova-
tions at the BoP, BoP consumers and their purchasing decisions
is given and second the limited quantitative work is described.
Zanello et al. (2016) conducted a systematic literature
review looking at factors influencing purchase decisions at the
BoP. The review highlights the importance of individual char-
acteristics like education and financial means in a consumers’
choice. The characteristics of consumers that allow them to use
a particular innovation to their advantage play a key role in
creating demand at BoP (Ramani, SadreGhazi, and Duysters
2012). An understanding of the nature and process of demand
creation is important to understand the needs of the customers
at BoP. The discrepancy in the perception of needs as expe-
rienced by the BoP and as perceived by a supplier is one of
the main obstacles for purchase at the BoP (Zanello et al.
2016). The needs mismatch can be observed in case of
technology design, product design and delivery design
(Ramani, SadreGhazi, and Duysters 2012). Chikweche and
Fletcher (2012) also stress the lack of contextualized and
local knowledge possessed by foreign firms seeking to oper-
ate in developing countries.
Nakata and Weidner (2012) developed a conceptual model
that looked at adoption of products at the BoP. The model
emphasized adoption as a two-stageprocess:symbolicand
material. Symbolic means a new product is “wanted/desired”,
while material is the actual purchasing of the product (Sen
According to the model, usually poor consumers
desire/want a product, but the product is not purchased due to
poverty. Therefore, both stages of adoption do not always occur
at the BoP. Nakata and Weidner’s (2012) model for new prod-
uct adoption considered the effect of social context, poverty,
marketing environment, and new product attributes.
Furthermore, a limited number of studies have looked at the
specificities of consumers in a BoP context, like the role played
by family in the purchase decision making process. Chikweche,
Stanton, and Fletcher (2012) highlight how BoP markets have
different characteristics through a case study about the role of
family decision making in purchasing in Zimbabwe. Unlike
developed markets, at the BoP the decision making for con-
sumption of goods is a more of a joint process, with possibly
different roles for different family members. Sometimes these
roles overlap between individuals but the frequent difference
between user and decision maker or even buyer plays a role for
marketing and design of products. Another important observa-
tion that likely applies in other BoP contexts is the different
types of purchasing decisions. Products are either bought
instantaneously as they are needed, when the product is actu-
ally available, or when the household can afford products
(Chikweche, Stanton, and Fletcher 2012; Viswanathan et al.
2014). Education and information available to the consumer
play a role in purchase decisions (Adkins and Ozanne 2005;
Nakata and Weidner 2012; Prahalad and Hammond 2002;
Ramani, SadreGhazi, and Duysters 2012). Those with higher
levels of formal education may make earlier purchase deci-
sions. Finally, linking consumption to the discussion of timing
of purchase also requires a look at the search process and the
role of information that consumers seek out or receive (Chik-
weche and Fletcher 2012; Lu
¨thje, Herstatt, and Von Hippel
2005; Ramani, SadreGhazi, and Duysters 2012; Subrahmanyan
and Tomas Gomez-Arias 2008). Authors such as Kotler et al.,
(2006) mentioned the importance of marketing to early adop-
ters of innovations at the BoP because this group can further
diffuse innovations through word of mouth referral.
Focusing on quantitative work, Ernst et al. (2015) studied
the effect of bricolage, local embeddedness and standardization
on product purchase. The quantitative findings from Ernst et al.
Howell et al. 3
(2015) support the qualitative research: bricolage and local
embeddedness are positively related to adoption highlighting the
importance of understanding aspects of local context like the
product design required by the local population, and the social
and cultural aspects of the local setup. Rahman, Hasan, and
Floyd (2013) focus on the role played by brand orientation in
creating a relative advantage at the BoP market of Bangladesh.
Examining consumer data from Grameenphone Community
Information Center, the authors show that brand orientation as
a strategy has a positive effect on adoption of the innovation.
Consumption and Purchase of Clean Water Technologies
At the BoP, where public access to clean drinking water is often
limited, bottled water or other clean water innovations are nec-
essary. From a macro perspective work has looked at the pri-
vatization of water resources and the rise of the bottled water
industry both in developed and emerging markets (Patsiaouras,
Saren, and Fitchett 2015). Specifically, there has been a rise in
multinational companies CSR initiatives in emerging markets
providing bottled drinking water (Brei and Bo¨ hm 2011). These
companies claim to be alleviating poverty through provision of
clean drinking water, but a more critical view suggests that the
companies are also trying to change their image in developed
markets (Brei and Bo¨hm 2011).
Cohen et al. (2017) conducted a quantitative study to inves-
tigate predictors of boiled water and bottled water consumption
in rural China analyzing data from 450 rural households in
Guangxi province. The results show that female-headed house-
holds were more likely to boil water whereas higher-income
households with younger, literate and male heads were more
likely to purchase bottled water. Ritter et al. (2017) tested the
effect of marketing strategies on consumers and suppliers for
the case of household chlorination products employing a ran-
domized controlled trial in rural Haiti. The results of the study
suggest that visits from sales agents may increase purchase of
chlorination products, however, the rise in sales does not cover
the costs of the visit. The authors argue that in developing
countries decisions related to promotion and pricing have long
lasting health implications.
Focusing only on factors driving bottled waterconsumption in
emerging markets, there are a limited number of bottled water
specific studies with mixed results. Research in Guatemala, the
Philippines and Ghana showed that generally smaller, higher
educated and higher income households were more likely to con-
sume bottled water (Francisco 2014; Quansah, Okoe, and Angenu
2015; Va´squez 2017). However in rural Guatemala, Va´squez
(2017) found that income was not a significant driver for smaller
purchases of bottled water. Income also had a minimal effect on
bottled water consumption in the Philippines (Francisco 2014).A
gender effect did not show up in any of the quantitative studies,
but qualitative research on bottled water consumers in Brazil
hinted that women may be important decision makers in the
choice of bottled water (de Queiroz et al. 2013).
Overall, it has been found that the limited uptake of decen-
tralized water solutions like bottled water were due to low
affordability of the products, high income variability of con-
sumers, and the companies’ challenges in communicating their
value proposition (Dahlberg 2017). In summary, the work look-
ing at bottled water consumption highlights the role of house-
hold head gender, education level, and wealth (Francisco 2014;
Quansah, Okoe, and Angenu 2015; Va´squez 2017).
Conceptual Model
Building on literature discussed above, we develop a concep-
tual model and three hypothesis regarding the timing of con-
sumption of bottled drinking water in a BoP context.
Education is considered to be a vital starting point for intro-
ducing change in consumption behavior of BoP consumers for
basic needs like clean drinking water, health care alternatives, to
mention few examples (Prahalad and Hammond 2002; UNDP
2008). Education provided to the consumer can be divided in
two main categories: traditional education and customer educa-
tion. Traditional education comprises of formal education pro-
vided at educational institutes, learning programs available on
online platforms. Usually, BoP consumers have access to a lower
level of education when compared with the main stream con-
sumers (Ramani, SadreGhazi, and Duysters 2012; Nakata and
Weidner 2012). One of the reasons for the low level of education
could be lack of access to up to date formal educational institu-
tions that provide basic education and skill training (Subrahman-
yan and Gomez-Arias 2008). In the presence of lack of basic
education skill set, a BoP consumer may not feel equipped with
sufficient information to make a consumption choice or lack the
required agency to make a decision, in other words a lack of
empowerment (Adkins and Ozanne 2005). Education is one of
the inputs which raises awareness, helps the poor consumer in
understanding different aspects of their decision process like
today’s gain in terms of need fulfillment as a compensation for
today’s financial loss. Education thereby increases the chances
of an informed choice being made by the poor consumer. Hence,
we argue that a high level of education leads to a consumption
choice being made earlier rather than later.
H1: Higher level of education increases the chances of ear-
lier consumption.
Receiving information is one of the first steps in gaining
awareness about a product, especially at the BoP. To under-
stand the structure of information received it is important to
analyze who is bringing the information, to whom the informa-
tion being transferred to and who is responsible for exchange of
the information in a local setup. BoP consumers receive infor-
mation via two means: personal network and in-person promo-
tions. Consumers at the BoP depend on their personal network
comprising of friends, relatives, neighbors while making con-
sumption choices (Nakata and Weidner 2012). Traditional
means of product promotion like advertisements via television
are less influential in the BoP setup, where consumers prefer in-
person promotion (Nakata and Weidner 2012; Chikweche and
Fletcher 2012). Information received through the personal
4Journal of Macromarketing XX(X)
network and in-person promotions helps the consumer in
understanding their need requirement. Both personal network
and in-person promotion give the opportunity to receive infor-
mation via personal interaction thereby, providing the possibil-
ity to see another consumer in their use environment.
Understanding of need information and intended use environ-
ment help the consumer in building his/her local information
¨thje, Herstatt, and Von Hippel 2005). One of the inputs
affecting accumulation of local information is personal net-
work and chances of early consumer’s personal network know-
ing about the product is less when compared with the network
of late consumer, even if all the consumers had a chance to
attend the product demonstration. Furthermore, given the col-
lectivist culture in a BoP setting, the local information about a
product possessed by late consumers will be greater compared
early consumers. Hence, we argue that receiving local infor-
mation increases the chances of late consumption.
H2: Local information increases the chances of late
External information search is one input in the consumer deci-
sion process. Smith and Beatty (1987, p. 85) define external
information search as “the degree of attention, perception and
effort directed toward obtaining environmental data or informa-
tion related to the specific purchase under consideration.” While
gathering information a poor consumerconsiders three main fac-
tors: cost benefit assessment of the purchase choice, factors
related to the local context, and household specific factors (Chik-
weche, Stanton, and Fletcher 2012; Ramani, SadreGhazi, and
Duysters 2012; Subrahmanyan and Tomas Gomez-Arias 2008).
Local factors could include cultural environment, social capital,
family systems and access to distribution systems. Household
specific factors could include tastes, needs and willingness to
explore different alternatives. Given the presence of collectivist
culture at the BoP and the dependence of other consumers on their
cultural environment while making choices, consumers who do
not engage in a search process follow the choices of the majority
BoP consumers or act to fulfill certain aspirational needs (Chik-
weche and Fletcher 2014; Jaiswal and Gupta 2015). Conse-
quently, consumers who engage in purposeful external
information searchto understand consumption options mightgain
exposure to additional sources of information not available in his/
her local context. The additional information may result in an
earlier purchase when compared with other consumers.
H3: External information search increases the chances of
earlier consumption.
Study Context and Sampling
In this section we describe the local context of the three study
countries: Kenya, Rwanda and Uganda.
Like many developing countries, Kenya faces the common
problem of delivering fresh water to the entire population.
Particularly due to urbanization, supply of quality water is a
challenge. Often the quality of piped water is mediocre or
unsafe for drinking. Although one of the Millennium Develop-
ment Goals was improving access to improved water sources
and progress was made in this regard it seems unclear whether
access to an improved water source also implies access to safe
drinking water (WHO et al. 2015). According to available data
from 2014, 67%of the Kenyan population (with higher levels
in urban areas) have access to an improved water source, where
an improved water source is classified as the ‘main source of
drinking water [being] a household connection (piped), public
tap or standpipe, tube well or borehole, protected dug well,
protected spring, rainwater collection, or bottled water’ (Kenya
National Bureau of Statistics et al. 2015). Yet, there is evidence
that many improved sources do have the bacteria e-coli and
other contaminants suggesting that a better definition of
improved water source is needed (Grady et al. 2015). Further-
more, a study for rural Kenya demonstrates that coping with
poor water quality makes households incur further costs (Cook,
Kimuyu, and Whittington 2016). Access to clean water for all
is further put under pressure since Kenya’s water supply was
privatized in 2002 through reforms that decentralized water
provision (Ministry of Water and Irrigation 2017; Water Ser-
vices Regulatory Board 2014).
Piped water delivery in Kampala, Uganda is slightly differ-
ent than in Kenya. The National Water and Sewage Corpora-
tion (NWSC) is responsible for water delivery in Kampala.
NWSC is a parastatal, owned by the government. The costs
of the water corporation are only partially covered by revenues
from consumers. Larger investments come from the govern-
ment through donor funds. NWSC is under the jurisdiction of
the Ministry of Water and the Environment.
The water quality
at the source is managed by the Ministry but tapped water
quality is managed by NWSC. Unlike at the Kenyan coast,
nearby Lake Victoria. While there are some supply and demand
issues, it is an infrequent occurrence compared to Kenya. Qual-
ity issues in Kampala are mainly due to old piping infrastruc-
ture and illegal tapping of pipes. Yet, water quality issues have
created a large market for bottled water. This market is not
properly regulated leading to cases where branded bottled
water containers are filled with unpurified water and resold.
Due to regulation problems, knowledge on certification in
Kampala seems to be high and consumers desire water that has
been certified by the Uganda Bureau of Standards (UNBS).
While Kigali, Rwanda faces some of the same water deliv-
ery issues as in Kenya and Uganda, the governance of water
delivery is again different. Water delivery is regulated by the
Rwanda Utilities Regulatory Agency (RURA). Companies
must pay for a license for a particular operating area in order
to delivery water to that region. Although competition is tech-
nically allowed, when a company acquires a license for an area,
no other companies can supply water in the same area. In Kigali
and the major cities of Rwanda, water is supplied by a govern-
ment owned, private company called Water and Sanitation
Corporation (WASAC), which has a monopoly there.
Howell et al. 5
is attempting to address existing shortages in water supply.
Two new purification plants are nearly complete that are
expected to fulfill the remaining supply needs of Kigali. Like
in the Kampala and Mombasa area, Kigali’s piping infrastruc-
ture is old and illegal tapping of water is widespread. More-
over, like Mombasa, Kigali has experienced rapid population
growth in particular in the period directly following the geno-
cide of 1994. The existing infrastructure is insufficient for the
current population.
In all three locations, the combination of insufficient
supply, and unsafe quality has created a market for bottled
water. While the underlying factors that have constituted
markets for innovations in bottled water are similar, location
specific factors are likely to have influenced how these
products are consumed.
Case Descriptions
Data for this study was collected from two bottled water com-
panies: Dutch Water Limited (DWL) and Jibu, LC3. While
there is a large market for bottled drinking water in East Africa
(Dahlberg, 2017), the two companies were specifically selected
for their unique business models and explicit social oriented
missions. The two selected companies highlight the ethical
complexities of fresh drinking water provision in an emerging
market setting. Bottled water is traditionally an expensive
product and companies like Coca Cola’s Dasani brand, have
often been criticized for creating markets for bottled drinking
water in areas plagued with water quality and supply issues
(Brei and Bo¨hm 2011). DWL and Jibu are both socially
oriented in their mission yet take a different delivery approach.
Unlike traditional bottled water companies, Jibu and DWL
seek to only use reusable bottles to minimize plastic waste.
DWL is a water company based in Mtwapa, Kenya
foundedin2006throughapartnership and investment from
two Dutch firms: Hatenboer and Reikon. DWL provides
bottled drinking water to three Kenyan counties (Kilifi,
Kwale, and Mombasa) in the region surrounding Mombasa.
DWL’s core low cost product is a 10L reusable jerry can.
DWL’s distribution model is to sell to small shops in the area
through water delivery on tuk-tuks.
Jibu, LC3 (Low profit liability company) is an American
founded “benefit corporation” selling low cost bottled drinking
water in Uganda, Rwanda and Kenya. Jibu was founded in 2012.
It started with simultaneous pilots in Uganda, Rwanda and the
Democratic Republic of Congo in 2013. Jibu’s declared goal was
to find a suitable model to provide low cost drinking water and
shift consumers from boiling to bottled water. In 2014, Jibu
began with a franchise model in Kigali and Kampala. Water
purification is decentralized at each franchise location. Purifica-
tion is done with ultra-filtration technology designed specifically
for the conditions in each city. Franchises purify city water and
package it into various sizes of reusable bottles.
While DWL has the social mission of providing affordable
drinking water, it also has the dual goal of running a financially
sustainable business. In the beginning of 2016, due to changes
in tax law, DWL had to adjust its business activities to accom-
modate a sharp increase in excise taxes on water. Although all
competitor manufacturers of bottled water were initially
affected by the excise tax increase, DWL is the only company
that fully complied, meaning that water prices for a 10L jerry
nearly doubled and DWL now offers a much higher priced
product than competitors.
Jibu provides a contrasting case study in the consumption of
bottled water. Due to Jibu’s social benefit corporation status,
the company aims to provide the lowest cost refill prices for
bottled water in their areas of operation. A social benefit cor-
poration is a relatively new business designation that allows
businesses to incorporate a social mission explicitly into their
strategy, allowing them to accept different types of investments
and zero margins
. Jibu consumers pay a relatively high bottle
deposit fee, then a low refill price. In some cases, because profit
is not the only motive of Jibu, they accept zero margin on their
refill prices to maintain the lowest cost water.
Sampling for the two companies was done using stratified ran-
dom sampling to account for demographic differences in the
geographic areas of operation for the two companies. In both
cases randomization occurred at the sales area (DWL) or zone
level (Jibu) where demographic characteristics would be sim-
ilar rather than randomly selecting consumers from the entire
population. This insured that the overall sample was represen-
tative of DWL and Jibu consumers.
Jibu conducted phone interviews with their consumers based
on lists of loyal customers collected from each franchise. Jibu
divides their cities into zones, which are based on population
density. Consumers were randomly selected from the given lists
according to the percentage of sales in each zone. Jibu conducted
data collection in December 2016 (Rwanda) and March 2017
(Uganda). The DWL questionnaire was based on Jibu’s question-
naire for consistency between the three countries. There were 19
and 20 zones in Rwanda and Uganda respectively. The majority
of sampling occurred in Jibu zones within Kigali and Kampala,
but some consumers were sampled from franchises in Busenyi
(North western Rwanda), and Entebbe (near Kampala-Uganda).
These areas are also mostly urban like Kigali and Kampala. In
addition to Jibu’s consumer survey, we collected similar data on
Jibu non-consumers in May 2017. Non-consumers were ran-
domly interviewed at Jibu micro-franchise locations.
Data collection in Kenya was done in March 2017. DWL’s
main factory is in Mtwapa in Kilifi county. From there water is
distributed to six distribution points located in the three coun-
ties of operation. At each of the six-distribution points and the
main factory there are sales areas. Sales areas represent the
routes the water is distributed along using auto rickshaws. In
each sales area there are DWL vendors, usually small kiosks
that sell DWL and other water brands, along with general prod-
ucts. According to DWL, there are roughly 5,000 vendors in
Mombasa, Kilifi, and Kwale counties. Out of the 5,000 ven-
dors, 112 were randomly selected for our study. At each vendor
6Journal of Macromarketing XX(X)
three consumers and four non-consumers were interviewed.
During test surveys it was discovered that it can be challenging
to locate consumers, therefore selection was based on proxim-
ity to the vendor.
Consumers were determined by their response to the ques-
tion: “When did you last purchase DWL/Jibu water?” Respon-
dents who purchased DWL/Jibu water in the last year were
considered consumers. Figure 1 gives an overview of the sur-
vey respondents and selection criteria. In Kenya, we have a
sample of 326 consumers, the Rwanda sample consists of
292 consumers, and the Uganda sample of 188 consumers.
Data and Empirical Model
Variable Description
The focus of the study is exposing consumer heterogeneity,
therefore consumers were divided into three categories based
on their time of first purchasing DWL or Jibu water. The
dependent variable is the consumer category. Jibu data was
based on a company survey and the questionnaire used for
DWL was based on Jibu’s survey. The categories of consu-
mers are as follows: (i) Early consumers (first purchase
greater than 2 years) (ii) Middle consumers (first purchase
between 1 and 2 years) and (iii) Late consumers (first pur-
chase from 1 year to present).
To ensure comparability across the empirical analyses of the
three datasets only variables that were measured in all three
datasets were included. In the first specification, only control
variables were included: the gender of household head which is
a binary variable that is equal to 1 for male and zero otherwise,
and the Progress out of Poverty Index (PPI). Chikweche, Stan-
ton, and Fletcher (2012) discuss the role of decision maker and
purchaser gender in African households. The PPI score does
not measure wealth directly but can be viewed as an asset
Panel A: DWL Sampling
Panel B: Jibu Rwanda Sampling
Panel C: Jibu Uganda Sampling
Vendors (N=112): Stratified random sample
Consumers, 3 at each vendor (N=326)
Time of
Purchase> 2
years ago
N= 121
Time of
Purchase> 1
2 years ago
N= 97
Time of
Purchase 1
year ago
N= 108
Consumers (N=292) : Random selection from consumer lists at each franchise
Time of
Purchase> 2
years ago
N= 68
Time of
Purchase> 1
2 years ago
N= 105
Time of
Purchase 1
year ago
N= 119
Consumers (N=188) : Random selection from consumer lists at each franchise
Time of
Purchase> 2
years ago
N= 16
Time of
Purchase> 1
2 years ago
N= 57
Time of
Purchase 1
year ago
N= 115
Figure 1. Sampling in the three countries.
Howell et al. 7
index. A score is assigned to every respondent based on his or
her response to a set of ten questions related to assets and
household demographics.
The PPI allows us to empirically
represent the local context of each country since the question-
naire differs from country to country. It should be noted that the
PPI score is updated regularly based on new household surveys
conducted in each country. For Uganda the 2012 survey was
used, for Kenya 2005, and for Rwanda 2005.
The next three specifications add the independent variables:
education, local information and external information. educa-
tion which consists of two binary variables, one for secondary
education and one for higher education. The excluded category
is made up of individuals with primary or no education. Local
information measures whether the individual had received per-
sonal information about the bottled water under study either
through a company representative or personal recommendation
from a friend or relative. The fourth specification adds the
independent variable of a purposeful search for clean water
which represents the external information search. A table
detailing the different variables, what they proxy, and their
literature motivation is presented in the appendix (Table A1).
All specifications contain controls for the location of the
sales points to capture differences in infrastructure and market
settings across the different sales points that are likely to affect
the decision and timing of uptake of bottled water. Note that
both companies entered different sales areas/zones at different
points in time. Therefore, it is necessary to control for these
regional differences. The location specific effects control for
differences in local infrastructure, local embeddedness, and
local context within each sub-market is represented by the
different sales areas/zones. This takes the work of Ernst et al.
(2015) on local embeddedness a step further. Table A2 in the
appendix shows the relative frequencies for consumer cate-
gories per sales area/zone, demonstrating that there is signifi-
cant variation across sales areas by consumer category. For
example, there is a higher relative frequency of respondents
in the late category for Kilifi as compared to early consumers
because Kilifi is an older market area for DWL. Similar relative
frequencies can be found for the Jibu zones depending on the
entry of the company into the respective zone.
Empirical Model
With the empirical analysis we want to assess the heterogeneity
across consumers by analyzing the determinants of the timing
of initial consumption across our three consumer groups.
Before studying within consumer heterogeneity, we start out
with a comparison between consumers and non-consumers to
set the stage. Standard innovation diffusion literature compares
consumers with non-consumers through a probit or logit model
(Geroski, 2000). We employ the logistic function to model a
binary dependent variable, i.e. the decision to consume bottled
water. For every individual iin location lwe estimate the
following model:
PðYil ¼1Þ¼1=ð1þexpððb0þb1XilÞÞ ð1Þ
where Y
is equal to 1 for consumers and X
contains the
control variables of interest, i.e. gender of the household head,
the education level, an asset index, and location specific
effects. We limit the analysis to this small set of predictors
since we are mainly interested in purchase decisions related
to education, economic wellbeing and easy access to the water,
i.e. being near to the bottled water supply. Standard errors are
clustered at the location level. Imposing the logistic transfor-
mation, we obtain a model that is linear in its predictors. The
model is fit by maximum likelihood.
lnðPðYil ¼1ÞÞ ¼ b0þb1Xil ð2Þ
The outcome of interest (dependent variable) is the category
in which the consumers fall based on their timing of initial
consumption: (i) early consumers, (ii) middle consumers, (iii)
late consumers.
Technically our outcome of interest can be viewed as an
ordinal variable. At the same time, the distance between cate-
gories is not a measure of actual distance. We cannot “rank” the
categories such that we could say late consumers are more
important than early ones. The different consumer groups
merely represent the different points in time when the consu-
mers started purchasing bottled water. Therefore, we decided to
employ a multinomial logistic regression model, which extends
the logit model to a multiclass analysis without taking the
ordering into account. For every class we estimate the effect
of the predictors on the probability of success in that class
compared to the reference class. Each class has its own inter-
cept and regression coefficients meaning that the predictors can
affect each category differently.
We predict the probability
that individual i from location lfalls into consumer class kwith
k¼1,2,3. Since we have 3 possible outcome classes, we run
two binary logistic regression models and treat the third class as
reference against which we regress the other classes. We can
write the probability for every consumer iin location lfalling
into one of the k consumer categories as a set of two indepen-
dent probabilities:
PðYil ¼jÞ¼expðbjXilÞ,1þX
;for j ¼1;2
Similarly to the basic logit model the above probability
employs a linear predictor function for every consumer iin
location lfalling into consumer category kwe can specify it
as follows:
fðk;i;lÞ¼bkXil ¼b0;kþb1;kmale hhil þb2;kPPIil
þb3;keducationil þb4;kinfoil þb5;kpurposeil
where the vector b
collects all regression coefficients asso-
ciated with the kth outcome; X
contains the control variables.
We control for gender of the household head (male_hhh
), an
asset index (PPI
), the education level (education
), receipt of
8Journal of Macromarketing XX(X)
local information about bottled water (info
), an external infor-
mation search for it (purpose
) and location specific effects
). Akin to the basic logit model we apply the logistic
transformation that allows us to model the logarithm of the
probability of seeing a given outcome using the linear predic-
tors. The model is estimated by maximum likelihood. Standard
errors are clustered at the sales point level to account for cor-
related residual variation.
Since the coefficient estimates of a multinomial logit model
cannot be directly interpreted, the marginal effects were com-
puted and are presented (Cameron and Trivedi 2009; Wool-
dridge 2016).
Descriptive Statistics
Descriptive statistics for all three countries are presented in
Table 1. DWL entered the market in 2006 so the early con-
sumer category makes up a larger percentage than for Jibu
Uganda and Rwanda. Almost 40%of DWL consumers are
early consumers with a fairly even split in the middle and late
categories. The opposite is true for Jibu due to their market
entry in 2014. For Jibu Rwanda 40%of consumers fall in the
late category, and for Jibu Uganda 61%. Jibu Uganda only has
9%of consumers in the early category whereas there’s a more
even split for Jibu Rwanda between the early and middle cate-
gories (24%and 36%). There is stiff bottled water competition
in Uganda, so it is possible that the customer lists in Uganda
represent more recent Jibu consumers.
Concerning the characteristics of the consumers there are
some demographic differences between the three countries.
The majority of DWL consumers have an education level of
secondary (55%). One quarter of the DWL sample is university
educated. Education levels are even higher for Jibu consumers,
with 40%(Rwanda) and 55%(Uganda) of consumers being
university educated. This finding indicates already that both
companies are not reaching the intended BoP consumers,
which are more likely to be those with no or limited education.
Turning to gender, a lower number of DWL consumers were
both male and the household head (38%)thanJibu(65%-
Rwanda, and 59%-Uganda). Looking at how consumers
received information about DWL and Jibu shows that most
(38%) consumers received information about DWL through
personal interactions (i.e. a friend or a DWL sales representa-
tive) with the remaining consumers either purchasing because
of a purposeful search for clean drinking water or an imperso-
nal source like an advertisement. For Jibu the source of infor-
mation varied between countries. In Rwanda 23%of
respondents received personal information about Jibu versus
65%for Uganda. A purposeful search for a new affordable
clean drinking water brand was highest for Rwanda consumers
(19%) compared to Kenya and Uganda (16%and 13%respec-
tively). Finally, looking at the asset index as a proxy for wealth
of the consumers across the three countries using the PPI score
shows that the average PPI score is at least ten points higher for
Rwandan and Ugandan consumers (70 and 68 respectively)
compared to Kenya (57). Note that higher PPI scores can be
used to indicate higher levels of wealth. The difference in
average PPI score could be due to the sampling differences (for
Jibu mostly loyal customers were interviewed) but also due to
geographic differences since Jibu is located in the capital cities
of Uganda and Rwanda where wealth is likely higher than on
the Kenyan coast.
The distribution of the PPI score is fairly spread for Kenya
and more skewed to the right for Uganda and Rwanda, i.e. the
scores fall toward the higher side of the scale and there are very
few low scores (Figure 2). Figure 2 shows that DWL has a
larger spread of income levels in spite of having the highest
water prices in the coastal region. Using a cut off of $2.50/day
(Prahalad 2005), 42.6%of DWL’s consumers could be classi-
fied as low income. However, a greater percentage would be
considered middle class ($2.50-$8.44/day) and above. In con-
trast to DWL consumers, both Jibu Uganda and Rwanda con-
sumers make up a higher income demographic. The majority of
both country’s consumers would be considered middle class or
higher. The wealth and education related findings are interest-
ing since Jibu is a social business that has an explicit mission
statement targeted at the BoP. The descriptive statistics already
suggest that they might not reach their targeted audience.
The first analysis is the estimation of a logit model of consu-
mers versus non-consumers of the two companies. The logit
model of consumers only is given in the Appendix (Table A3)
and highlights the reason for only focusing on consumers of the
two companies. Because not all non-consumer data was avail-
able for Uganda the logit was only conducted for Kenya and
Table 1. Descriptive Statistics.
Kenya Rwanda Uganda
Dev. Mean
Dev. Mean
Consumer categories
Early consumer 0.398 0.237 0.093
Middle consumer 0.285 0.360 0.292
Late consumer 0.318 0.403 0.615
Control variables
Male household
0.383 0.647 0.590
PPI poverty score 57.193 15.319 70.147 9.706 68.317 8.279
Education (Excluded category: No formal education
and primary education)
Secondary 0.551 0.338 0.280
University 0.252 0.406 0.534
Received personal
about water
0.38 0.219 0.652
Purposeful search 0.164 0.212 0.13
Observations 274 278 161
Note: The descriptive statistics were derived based on the regression models
including the full set of covariates.
Howell et al. 9
Rwanda. Additionally, there was no data available for the
household role for Rwanda, so instead of the male household
head variable simply respondent gender was used.
The logit results presented in Table A3 in the appendix show
two findings that help strengthen the argument to further break
apart the consumers of the two companies. First, the results
differ between the two countries. In Rwanda, male respondents
are more likely to purchase water, and those with a university
education are more likely to be Jibu consumers. Additionally,
for Jibu consumers, using the asset index, wealthier consumers
are more likely to purchase Jibu water. The results are different
for Kenya. Gender and wealth have no influence, rather DWL
consumers are merely more likely to have a higher level of
education (secondary or university). Therefore, understanding
the heterogeneity within consumers and between countries is
the focus.
The results of the main multivariate analysis are presented in
Table 2 for each category of consumer. The middle category
was used as the base category but the full regression results are
presented to highlight the differences between the three con-
sumers categories. Four specifications are presented for every
country. They all contain location fixed effects to account for
the fact that DWL and Jibu entered different areas at different
points in time.
Specification 1 focuses on the socio-demographic determi-
nants, ie: wealth and gender. For Jibu Rwanda consumers, early
consumers are about 15 percentage points more likely to be
male household heads. In Uganda, an opposite effect is shown
with middle consumers being less likely to be male by 11
percentage points and late consumers being more likely to be
male by about 9 percentage points. The result does not show up
for Kenya, which could be due to cultural differences between
the three countries. Additionally, note that overall Kenya had a
lower percentage of male household heads (38%) than Rwanda
and Uganda. However, the existing literature also suggests that
women household heads may be less likely to be risk taking in
consumption decisions (Byrnes, Miller, and Schafer 1999).
Finally, it is important to note that the Rwanda dataset had the
highest percentage of male household heads (65%).
Looking at wealth as measured by the asset index PPI score,
overall wealth does not determine consumption decisions in all
countries. For Kenya and Rwanda there is an economically
small wealth effect for late consumers. For both countries weal-
thier respondents are less than 1 percentage point less likely to
be late consumers. Since the wealth effect, if identified at all, is
very small, these results suggest that wealth cannot be consid-
ered an important driver in timing of purchase. Considering
that DWL nearly doubled their water prices during the period
of time for late consumers, the individuals who we identify as
late consumers do not have income as a binding factor to con-
sumption. Overall specification 1 neither supports nor dis-
proves that DWL or Jibu reaches poor households. But taken
with the descriptive statistics, the findings suggest that neither
company are reaching BoP consumers. This result is particu-
larly relevant for Jibu since Jibu income data suggests that Jibu
consumers are at a higher end of the income spectrum.
In the next step (Specification 2), the effect of education is
assessed. There is a small education effect for the middle and
late consumer categories of DWL. For two of the specifications
in the middle category of DWL consumers, having a university
education increases the chance of being a middle consumer by
almost 12 percentage points. However, the education effect
does not show up for Jibu consumers. But it should be noted
that Jibu consumers are on average even higher educated com-
pared to the DWL consumers. Due to a relative lack of varia-
tion in education among Jibu consumers we are unable to
identify an education channel for Rwanda and Uganda. This
could be due to sampling issues and demographic differences
between Jibu being located in capital cities and DWL in a less
educated and less urban region.
In specification 4 we added information related variables,
namely whether the consumer received local information about
Panel A: DWL Consumers
Panel B: Jibu Uganda Consumers
Panel C: Jibu Rwanda Consumers
< $1.25 $1.25 -
$2.50 -
$4.00 -
> $8.44
Percentage of DWL
Daily Household Income (2005 PPP)
1% 5% 7%
< $1.25 $1.25 -
$2.00 -
$2.50 -
$5.00 -
> $8.44
Percentage of Jibu Uganda
Daily Household Income (2005 PPP)
< $1.25 $1.25 -
$2.00 -
$2.50 -
$5.00 -
> $8.44
Percentage of Jibu Rwanda
Daily Household Income (2005 PPP)
Figure 2. Income distribution of consumers.
10 Journal of Macromarketing XX(X)
Table 2. Empirical Results of Determinants of Consumption of Bottled Water in Kenya, Rwanda and Uganda: Multinomial Logit Estimates.
Kenya Rwanda Uganda
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Early Consumers
Male household head 0.012 0.018 0.024 0.026 0.156*** 0.152*** 0.149*** 0.146*** 0.025 0.023 0.027 0.024
(0.060) (0.057) (0.053) (0.054) (0.037) (0.043) (0.045) (0.049) (0.037) (0.037) (0.032) (0.031)
PPI score -0.059 0.001 0.001 0.001 -0.002 -0.002 -0.002 -0.002 -0.001 -0.001 -0.001 -0.001
(0.067) (0.002) (0.002) (0.002) (0.003) (0.003) (0.003) (0.003) (0.002) (0.002) (0.002) (0.002)
Education (Excluded category: No formal education and primary education)
Secondary -0.017 -0.0133 -0.013 -0.025 -0.028 -0.035 0.008 0.026 0.027
(0.092) (0.097) (0.099) (0.072) (0.073) (0.072) (0.055) (0.054) (0.052)
University 0.060 0.080 0.083 0.003 0.004 0.002 -0.018 -0.007 0.001
(0.095) (0.101) (0.103) (0.046) (0.046) (0.044) (0.031) (0.030) (0.028)
Received personal information about water -0.103 -0.127* -0.023 0.003 0.054 0.086
(0.069) (0.071) (0.053) (0.064) (0.062) (0.064)
Purposeful search -0.067 0.1190.109***
(0.051) (0.076) (0.037)
Middle Consumers
Male household head 0.065 0.084** 0.084** 0.080** -0.092* -0.091-0.083 -0.089 -0.113* -0.110** -0.114** -0.114**
(0.043) (0.039) (0.039) (0.040) (0.054) (0.058) (0.063) (0.062) (0.059) (0.056) (0.055) (0.054)
PPI score -0.055 0.002 0.002 0.003 -0.006** -0.007** -0.007** -0.007** 0.004 0.004 0.004 0.004
(0.087) (0.002) (0.002) (0.002) (0.003) (0.003) (0.003) (0.003) (0.005) (0.005) (0.005) (0.005)
Education (Excluded category: No formal education and primary education)
Secondary 0.067 0.066 0.061 -0.032 -0.038 -0.041 -0.054 -0.056 -0.053
(0.074) (0.074) (0.075) (0.077) (0.076) (0.074) (0.120) (0.118) (0.125)
University 0.120* 0.116* 0.106 0.061 0.054 0.057 0.075 0.070 0.072
(0.070) (0.070) (0.069) (0.101) (0.102) (0.103) (0.087) (0.085) (0.092)
Received personal information about water 0.022 0.080 0.105 0.114 -0.037 -0.102*
(0.047) (0.055) (0.076) (0.086) (0.056) (0.061)
Purposeful search 0.186* 0.095* -0.199*
(0.096) (0.056) (0.115)
Table 2. (continued)
Kenya Rwanda Uganda
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Late Consumers
Male household head -0.089 -0.102* -0.108* -0.106* -0.064 -0.060 -0.066 -0.057 0.088* 0.087* 0.087* 0.091*
(0.055) (0.059) (0.057) (0.056) (0.042) (0.044) (0.044) (0.046) (0.052) (0.050) (0.048) (0.049)
PPI score -0.005** -0.004** -0.003** -0.004** 0.008*** 0.009*** 0.009*** 0.009*** -0.003 -0.003 -0.003 -0.003
(0.002) (0.002) (0.002) (0.002) (0.003) (0.003) (0.003) (0.003) (0.006) (0.006) (0.006) (0.006)
Education (Excluded category: No formal education and primary education)
Secondary -0.049 -0.053 -0.047 0.057 0.066 0.076 0.046 0.030 0.026
(0.103) (0.109) (0.104) (0.067) (0.067) (0.066) (0.115) (0.122) (0.129)
University -0.180 -0.196-0.189 -0.064 -0.058 -0.060 -0.057 -0.064 -0.073
(0.118) (0.125) (0.123) (0.116) (0.116) (0.116) (0.093) (0.096) (0.099)
Received personal information about water 0.081 0.047 -0.082 -0.117** -0.017 0.017
(0.065) (0.059) (0.064) (0.059) (0.056) (0.053)
Purposeful search -0.119 -0.214*** 0.090
(0.088) (0.070) (0.121)
Sales point/Location FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 274 274 274 274 278 278 278 278 161 161 161 161
Chi2 37.90 43 46.05 52.34 135.39 139.46 142.21 151.31 72.95 75.65 76.66 79.56
p-value 0.036 0.035 0.031 0.013 0.000 0.000 0.000 0.000 0.001 0.002 0.003 0.003
Pseudo R2 0.064 0.072 0.077 0.088 0.226 0.233 0.238 0.253 0.258 0.267 0.271 0.281
Note: Results from a multinomial logit regression. Standard errors are in parentheses. ***/**/* indicates statistical significance at the 1/5/10% level, respectively; p-value¼12%, p-value¼11%
DWL/Jibu and whether the consumer searched for bottled
water on purpose. Rwandan consumers receiving local infor-
mation are less likely to be late consumers . In Kenya early
consumers were about 13 percentage points less likely to have
received local information. Finally in Uganda, consumers in
the middle category were less likely to have received local
information. This result fits with the market situation in
Rwanda. Jibu entered the market when there was already a
well-established market for bottled water, suggesting that early
consumers would have purchased due to curiosity and not
through local information.
Looking at the consumption decision being made as a result
of a purposeful search for drinking water alternatives shows a
few interesting results. For Uganda, early consumers were 10
percentage points more likely to have conducted a purposeful
search. A nearly significant (p ¼12%) similar result is shown
in Rwanda with early consumers being 11 percentage points
more likely to conduct a purposeful search. For the middle
category of consumers in Uganda and Rwanda, opposite effects
are shown. Uganda middle consumers are less likely to have
been purposefully searching for water (20 percentage points)
and in Rwanda the middle consumers are slightly more likely to
conduct a purposeful search. Finally, Rwandan late consumers
were more than 20 percentage points less likely to have been
purposefully searching for water. In Kenya, purposeful search
only showed up significant for consumers in the middle group
with those purchasing due to a purposeful search being 19
percentage points more likely to be middle consumers. For
DWL consumers, in the last year before data collection water
prices doubled, which means particularly for late consumers
the reasons to purchase DWL water may have changed.
The most important finding is related to the supply side
dimension of consumption and is disguised in the tables pre-
sented. Across all empirical models we include location fixed
effects. The coefficients associated with the location fixed
effects tend to be large in absolute terms and are jointly statis-
tically significant (p-value < 0.000) suggesting that it is loca-
tion characteristics that are the most important determinants of
the consumption of bottled water. Since, the two companies
entered the different sales areas at different moments in time,
the location fixed effects also include market entry and suggest
that the strongest determinant of uptake is the timing of market
entry. This finding shows that at the BoP, due to the many
constraints the individuals and households are facing, the like-
lihood of demand driven consumption is limited.
Concerning the reliability of our results across specifica-
tions, Table 2 shows that there is stability of the coefficient
estimates across models. The stability of the coefficients sug-
gests that the identified pathways are meaningful and not dri-
ven by omitted variables that might appear in the specifications
with a limited number of control variables. Secondly, as a
further robustness check the PPI score was divided into two
categories by using the median PPI score for each country
dataset to construct a binary variable “asset poor” (1 ¼poorer
group, 0 ¼higher group). These results are presented in the
Appendix and show consistency with the main results in terms
of coefficient signs and values.
The regression results presented several insights regarding con-
sumer heterogeneity and some differences from the initial
hypotheses. The first hypothesis regarded the role of education
in the timing of purchase. It was hypothesized that those with a
higher level of education would be more likely to be earlier
consumers. Those with a higher level of education will have
higher knowledge of the health impacts of unsafe drinking
water. Because of the already high levels of education of Jibu
consumers there was no education effect for Jibu consumers in
Rwanda and Uganda. However, the consumer logit model did
show that higher educated individuals were more likely to pur-
chase Jibu water. Considering that in Rwanda and Uganda in
particular most individuals were already using either treated or
bottled water, this result suggests that Jibu water might be
considered higher quality by “higher” status individuals. In
Kenya, there was an education effect for middle and later con-
sumers. University education decreased the chance of being a
late consumer, suggesting that earlier consumers could be con-
sidered “higher” status individuals (in terms of education and
knowledge about clean drinking water). University education
also increased the chances of being a middle consumer com-
pared to earlier or later. Ramani, SadreGhazi, and Duysters
(2012) also found the importance of education purchase of
innovations related to health. Previous studies on bottled water
consumption also showed higher education levels increasing
the likelihood of purchasing bottled water (Francisco 2014;
Quansah, Okoe, and Angenu 2015; Va´squez 2017). Therefore,
the education hypothesis was partially corroborated through the
Kenya results.
The second hypothesis looked at the role of local informa-
tion in the timing of consumption. Local information com-
prised of two components: recommendation from friends and
family and being approached by a company representative.
Overall local information had minimal effect on the timing of
purchase. While prior literature suggests that particularly at the
BoP word of mouth information about a product can influence
purchase decision (Kotler et al. 2006; Nakata and Weidner
2012), our results show minimal effects from local information.
This could be possible because of two reasons. Firstly, for
consumer to receive information through word of mouth
enough consumers need to be aware of the product, which will
not be the case if it is a new product. So, word of mouth may
not lead to early purchase but rather to late purchase when
sufficient number of consumers know about it and can spread
the word. Secondly, the minimal and mixed effects may be due
to the features of the markets under study. The last category of
consumers for Kenya were those who first purchased when
DWL water was increased sharply and was the one of the
highest priced in the market. Therefore it is expected that these
consumers would be purchasing for different reasons than ear-
lier consumers. For Jibu consumers, most of them were already
Howell et al. 13
using treated or bottled water indicating that they were already
aware of the health benefits of clean drinking water.
The last hypothesis looked at the role of external informa-
tion search. It was expected that consumers who engaged in a
purposeful search for a clean drinking water alternative would
be earlier consumers. We were able to partially corroborate
this hypothesis. In both Uganda and Rwanda, early consumers
were more likely to have engaged in a purposeful search. For
the consumers who purchased later the results were mixed.
Because Jibu is new to the market in both Uganda and
Rwanda and therefore consumers in these markets are already
familiar with bottled water, it fits that earlier consumers are
those who are curious about drinking water alternatives and
are seeking out another water source. In Rwanda, we also saw
that late consumers were less likely to be purposefully search-
ing for water.
While wealth was not included in our conceptual model and
merely used as a control variable, the results also suggested that
both companies are not reaching their target demographic. The
results fit with literature criticizing western firms who seek to
achieve social missions like reaching low income consumers
and suggest that firms like DWL/Jibu are often reaching con-
sumers at the higher end of the BoP or middle class (Karnani
2007; Meagher 2018).
Finally, the results emphasized the supply driven side of
consumption as backed up by prior literature (Chikweche and
Fletcher 2011b; Sheth 2011; Viswanathan et al. 2014). The
strongest determinant of purchase was from the location
effects, indicating that consumers may purchase based primar-
ily on availability.
To provide clean water is one of the sustainable development
goals (SDGs) set out by the United Nations in 2015. Particu-
larly in the African context where infrastructure is still lacking,
clean drinking water is often an unfulfilled need that is now
creating markets for bottled water companies like DWL and
Jibu (Brei and Bo¨ hm 2011; Patsiaouras, Saren, and Fitchett
2015). A first step in reaching this goal is to understand the
characteristics of consumers living at BoP and the local context
they inhabit. Consumers are usually studied as a homogenous
source of demand (Kotler et al. 2006; Prahalad 2005; Sheth
2011). In this paper, we shed light on the heterogeneity
amongst BoP consumers and highlight features of their local
context by comparing three countries and two firms, with the
help of a multivariate analysis.
The study contributes to the larger picture of marketing at
the BoP by empirically assessing whether the characteristics of
early consumers are systematically different from consumers
who purchase at a later time. First, the role of formal education
on timing of consumption was explored. We have identified
that a higher level of education made it less likely to be a late
consumer. Higher education levels also likely play a role in
purchase of a product like clean drinking water, where educa-
tion would imply a higher level of knowledge about the health
impacts of poor drinking water quality (Ramani, SadreGhazi,
and Duysters 2012). By looking at consumers versus non con-
sumers we also saw that higher educated were more likely to
consume bottled water which fit with previous bottled water
studies. Furthermore, we identified that the information search
process plays a role in the decision to purchase.
In contrast, wealth related factors had little to no influence
on time of purchase. The consumers under study tended to be
wealthier than the average poor person in the countries under
study. BoP innovation literature often focuses on low price
points as a means to reach low income consumers, yet the
results from the DWL and Jibu study suggest that the price
might not be low enough for poor consumers to purchase.
Prior marketing literature also emphasized that poverty is an
inhibitor in participation in a market society (Bonsu and Polsa
2011; Yurdakul, Atik, and Dholakia 2017) which is also sug-
gested by DWL and Jibu’s inability to reach poor consumers.
More importantly the sales area effects are strongest in our
model, which implies that purchase of bottled water is driven
more by supply of bottled water rather than demand. This
result also fits with marketing and consumer literature that
highlighted how BoP consumers tend to make instantaneous
purchase decisions based more on availability of a product
than on demand (Chikweche and Fletcher 2011b; Sheth
2011; Viswanathan et al. 2014).
The BoP literature can be divided in two camps. One camp
focuses on market development and emphasizes the need to
reach the poor consumers (Prahalad 2005; Ramani and
Mukherjee 2014; Sheth 2011). The emphasis on poor consu-
mers is in line with our finding that consumption is driven by
supply. The other camp criticizes businesses seeking to reach
this demographic since they fear the poorest might not be
reached or if targeted successfully might be exploited (Karnani
2007, 2009; Meagher 2018). The results of our study highlight
a third overlooked aspect in the BoP literature. Companies like
DWL and Jibu are reaching more the middle-income category
of consumers, not necessarily the very poorest despite the
stated mission of reaching poor consumers. Moreover, as
exhibited by the differences between consumer categories, con-
sumers at the BoP are not as homogenous as assumed in some
of the earlier BoP literature (Guarı´n and Knorringa 2014).
DWL and Jibu are providing a much-needed good to an under-
served market, even though they are likely not reaching their
original target group.
Future research should continue the quantitative angle on
BoP consumers and further explore the rising “new middle
class” and the role for western companies in reaching this new
demographic. Additionally, to gain a grounded understanding
of the heterogeneity amongst the “new middle class” it is crit-
ical to understand the behavioral aspects that might trigger
consumption and thereby foster the diffusion of clean water
at the BoP. An in depth understanding of consumers’ hetero-
geneous needs is vital for companies marketing to not only the
“new middle class” but also the original target – poor consu-
mers at the BoP.
14 Journal of Macromarketing XX(X)
Table A1. Selection of Variables.
Variable Description Related literature
Measured by time of purchase
Household head male (Byrnes, Miller, and Schafer
1999; Chikweche, Stanton,
and Fletcher 2012; Cohen
et al. 2017; de Queiroz
et al. 2013)
Received a recommendation
from a friend or relative or
contact with a company
(Chikweche and Fletcher,
2012; Lu¨thje, Herstatt, and
Von Hippel 2005; Ramani,
SadreGhazi, and Duysters
2012; Subrahmanyan and
Tomas Gomez-Arias 2008)
Purposeful search for bottled
(Smith and Beatty 1987)
PPI score (Francisco 2014; Quansah,
Okoe, and Angenu 2015;
´squez 2017)
Categories of education
(primary, secondary,
(Adkins and Ozanne 2005;
Quansah, Okoe, and
Angenu 2015; Ramani,
SadreGhazi, and Duysters
2012; Subrahmanyan and
Tomas Gomez-Arias 2008;
UNDP 2008)
Sales location DWL sales area (11 locations)
where respondent is
located or Jibu franchise
where water was
purchased, controls for
local context
(Ernst et al. 2015)
Table A2. Frequencies of Consumers Per Category by Sales Area/
Variables Consumer Categories per area/zone
Panel A: Kenya Sales Areas Early Middle Late Total
Bamburi 19.83 7.22 12.04 13.50
Kilifi 1.65 4.12 9.26 4.91
Magongo 5.79 11.34 12.04 9.51
Mikindani 2.48 5.15 0.93 2.76
Mtwapa 28.93 15.46 23.15 23.01
Tudor 11.57 13.40 6.48 10.43
Ganjoni 2.48 14.43 13.89 9.82
Ukunda 3.31 1.03 3.70 2.76
VOK 11.57 9.28 8.33 9.82
Ratna 6.61 9.28 4.63 6.75
Mishomoroni 5.79 9.28 5.56 6.75
Total 100 100 100 100
Panel B: Uganda Zones Early Middle Late Total
Bugolobi 0.00 8.77 4.46 5.41
Downtown 6.25 1.75 1.79 2.16
Entebbe 0.00 0.00 5.36 3.24
Ggaba 0.00 1.75 7.14 4.86
Kabale 0.00 0.00 0.89 0.54
Kabuusu 6.25 0.00 2.68 2.16
Kamwokya 6.25 7.02 3.57 4.86
Table A2. (continued)
Panel B: Uganda Zones Early Middle Late Total
Kawempe 43.75 10.53 5.36 10.27
Kireka 0.00 7.02 4.46 4.86
Kisaasi 0.00 1.75 2.68 2.16
Kitooro 0.00 1.75 8.04 5.41
Lugala 0.00 1.75 9.82 6.49
Lweza 0.00 3.51 3.57 3.24
Makindye 0.00 5.26 3.57 3.78
Mbarara 0.00 3.51 3.57 3.24
Najjanankumbi 12.50 0.00 1.79 2.16
Namugongo 0.00 1.75 1.79 1.62
Namuwongo 6.25 36.84 19.64 23.78
Nansana 18.75 1.75 3.57 4.32
Ntinda 0.00 5.26 6.25 5.41
Total 100 100 100 100
Panel C: Rwanda Zones Early Middle Late Total
Gatsata 0.00 1.94 0.00 0.69
Gikondo 1.47 5.83 4.27 4.17
Kabeza 7.35 2.91 5.98 5.21
Kabuga 0.00 1.94 2.56 1.74
Kagugu 4.41 1.94 0.85 2.08
Kanombe 7.35 4.85 1.71 4.17
Kibagabaga 0.00 1.94 0.85 1.04
Kicukiro 7.35 9.71 6.84 7.99
Kimironko 22.06 15.53 5.13 12.85
Kimisagara 0.00 5.83 6.84 4.86
Kinamba 26.47 15.53 3.42 13.19
Masaka 0.00 1.94 3.42 2.08
Niboye 1.47 0.97 2.56 1.74
Nyamata 0.00 2.91 5.98 3.47
Nyamirambo 11.76 16.50 16.24 15.28
Rubavu 0.00 0.00 14.53 5.90
Ruyenzi 1.47 0.97 8.55 4.17
Rwamagana 0.00 0.00 2.56 1.04
Sonatube 2.94 2.91 2.56 2.78
Wherever 5.88 5.83 5.13 5.56
Total 100 100 100 100
Note: Shares of consumers are expressed in percentage terms.
Table A3. Logit Model of DWL and Jibu Consumers and Non-
consumers for Kenya and Rwanda.
Kenya Rwanda
(1) (2) (3) (4)
Male household head 0.058 0.059 0.120** 0.147***
(0.062) (0.060) (0.049) (0.048)
PPI score 0.000 0.014***
(0.001) (0.002)
Low income group -0.002 -0.200***
(0.035) (0.055)
Education (Excluded category: No formal education and primary
Secondary 0.139*** 0.139*** 0.029 0.0347781
(0.033) (0.032) (0.047) (0.048)
University 0.172*** 0.171*** 0.098*** 0.127***
(0.033) (0.035) (0.037) (0.038)
Sales point/Location FE Yes Yes Yes Yes
Observations 630 630 572 572
Howell et al. 15
Table A4. Results from Multinomial Logit of Consumer Categories Using Asset Poor Variable.
Kenya Rwanda Uganda
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Early Consumers
Male household head 0.012 0.020 0.027 0.029 0.154*** 0.151*** 0.148*** 0.144*** 0.024 0.021 0.025 0.021
(0.060) (0.057) (0.054) (0.056) (0.036) (0.043) (0.045) (0.048) (0.038) (0.039) (0.033) (0.032)
Asset poor -0.059 -0.054 -0.050 -0.051 0.054 0.060 0.060 0.060 -0.004 -0.006 -0.002 -0.008
(0.067) (0.071) (0.069) (0.071) (0.060) (0.059) (0.059) (0.055) (0.043) (0.046) (0.045) (0.043)
Education (Excluded category: No formal education and primary education)
Secondary -0.018 -0.014 -0.014 -0.025 -0.028 -0.036 0.003 0.018 0.020
(0.091) (0.097) (0.100) (0.072) (0.072) (0.072) (0.051) (0.050) (0.050)
University 0.059 0.078 0.082 0.010 0.011 0.008 -0.023 -0.014 -0.008
(0.093) (0.099) (0.103) (0.045) (0.045) (0.043) (0.025) (0.024) (0.023)
Received personal information about water -0.104 -0.128* -0.020 0.007 0.050 0.084
(0.068) (0.070) (0.052) (0.065) (0.063) (0.066)
Purposeful search -0.071 0.1190.111***
(0.052) (0.075) (0.036)
Middle Consumers
Male household head 0.065 0.076* 0.074* 0.070* -0.098* -0.097* -0.088 -0.095 -0.109* -0.106** -0.111** -0.110**
(0.043) (0.039) (0.039) (0.040) (0.053) (0.056) (0.061) (0.059) (0.056) (0.054) (0.052) (0.051)
Asset poor -0.055 -0.038 -0.039 -0.040 0.066 0.079* 0.075* 0.080* -0.082 -0.074 -0.081 -0.073
(0.087) (0.087) (0.084) (0.090) (0.049) (0.047) (0.045) (0.049) (0.073) (0.080) (0.081) (0.078)
Education (Excluded category: No formal education and primary education)
Secondary 0.077 0.076 0.072 -0.033 -0.039 -0.041 -0.044 -0.046 -0.042
(0.069) (0.068) (0.069) (0.078) (0.076) (0.075) (0.120) (0.120) (0.126)
University 0.137* 0.133* 0.124* 0.058 0.050 0.054 0.079 0.072 0.076
(0.070) (0.069) (0.069) (0.104) (0.105) (0.106) (0.084) (0.082) (0.089)
Received personal information about water 0.021* 0.078 0.108 0.116 -0.049 -0.110*
(0.046) (0.056) (0.075) (0.086) (0.058) (0.062)
Purposeful search 0.184* 0.092 -0.188*
(0.098) (0.057) (0.112)
Table A4. (continued)
Kenya Rwanda Uganda
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Late Consumers
Male household head -0.077 -0.095* -0.101* -0.099* -0.057 -0.054 -0.060 -0.049 0.086* 0.085* 0.086* 0.090*
(0.053) (0.057) (0.056) (0.056) (0.042) (0.044) (0.043) (0.047) (0.051) (0.050) (0.048) (0.048)
Asset poor 0.1140.092 0.089 0.091 -0.121** -0.140** -0.135** -0.140** 0.085 0.080 0.083 0.081
(0.073) (0.061) (0.062) (0.067) (0.049) (0.056) (0.055) (0.055) (0.082) (0.087) (0.091) (0.091)
Education (Excluded category: No formal education and primary education)
Secondary -0.059 -0.062 -0.058 0.058 0.066 0.077 0.042 0.028 0.021
(0.094) (0.100) (0.094) (0.068) (0.068) (0.065) (0.111) (0.119) (0.127)
University -0.196* -0.211* -0.206* -0.068 -0.061 -0.063 -0.056 -0.058 -0.068
(0.109) (0.116) (0.113) (0.124) (0.124) (0.122) (0.084) (0.087) (0.091)
Received personal information about water 0.083 0.050 -0.088 -0.123** -0.002 0.026
(0.062) (0.054) (0.064) (0.059) (0.063) (0.056)
Purposeful search -0.113 -0.211*** 0.077
(0.094) (0.064) (0.121)
Sales point/Location FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 274 274 274 274 278 278 278 278 161 161 161 161
Chi2 35.74 41.78 44.91 50.97 130.20 134.39 137.29 146.22 73.43 75.99 77.01 79.75
p-value 0.058 0.045 0.039 0.018 0.000 0.000 0.000 0.000 0.001 0.002 0.003 0.003
Pseudo R2 0.060 0.070 0.075 0.085 0.218 0.225 0.230 0.245 0.259 0.268 0.272 0.282
Note: Results from a multinomial logit regression. Standard errors are in parentheses. ***/**/* indicates statistical significance at the 1/5/10% level, respectively; p-value¼12%,  p-value¼11%
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
The author(s) disclosed receipt of the following financial support for
the research, authorship, and/or publication of this article: The
research of this study was funded by NWO grant 313-99-314 by the
Netherlands Organization for Scientific Research.
Rachel Howell
1. In economic consumer theory symbolic would be referred to as
“preference” and material as “revealed preference” (Lancaster,
2. Water delivery is regulated through two agencies under the Min-
istry of Water: Water Services Regulatory Board (WASREB) and
Water Resources Management Authority (WMRA).
3. accessed 4 May 2017.
4. Interview with the research director of the National Water and
Sewage Corporation (NWSC) on 5 May 2017.
5. Accessed 11 May 2017.
6. Interview with the director of water regulation from Rwanda Util-
ities Regulatory Agency (RURA) on 11 May 2017.
8. For more information see: Last
accessed: 20 June 2018.
9. The disadvantage of employing the multinomial logit is that we
throw away information about the ordering in time. An ordinal
logit model would preserve that information but imposes stronger
assumptions, i.e. the proportional odds assumption. But, for the
case at hand we consider it as an advantage that the multinomial
logit allows us to estimate different coefficient estimates for the
predictors in every category whereas the ordered logit only identi-
fies individual intercepts for every class but the same predictor
Adkins, N. R. and J. L. Ozanne (2005), “Critical Consumer Education:
Empowering the Low-Literate Consumer,” Journal of Macromar-
keting, 25 (2), 153-62. doi:10.1177/0276146705280626.
AfDB, A. (2016), African Development Report 2015-Growth, Poverty
and Inequality Nexus: Overcoming Barriers to Sustainable Devel-
opment. Abidjan: African Development Bank.
Agnihotri, A. (2012), “Revisiting the Debate over the Bottom of the
Pyramid Market,” Journal of Macromarketing,32(4),417-23.
Alur, S. and J. P. L. Schoormans (2013), ‘Retailers and New Product
Acceptance in India’s Base of Pyramid (BoP) Markets: Proposi-
tions for Research,” International Journal of Retail & Distribution
Management, 41 (3), 189-200. doi:10.1108/09590551311306246.
Banbury, C., L. Herkenhoff, and S. Subrahmanyan (2015),
“Understanding Different Types of Subsistence Economies: The
Case of the Batwa of Buhoma, Uganda,” Journal of Macromarket-
ing, 35 (2), 243-56. doi:10.1177/0276146714528954.
Bonsu, S. K. and P. Polsa (2011), “Governmentality at the Base-
of-the-Pyramid,” Journal of Macromarketing, 31 (3), 236-44.
Brei, V. and S. Bo¨hm (2011), “Corporate Social Responsibility as
Cultural Meaning Management: A Critique of the Marketing of
‘Ethical’ Bottled Water,” Business Ethics: A European Review,
20 (3), 233-52. doi:10.1111/j.1467-8608.2011.01626.x.
Byrnes, J. P., D. C. Miller, and W. D. Schafer (1999), “Gender Dif-
ferences in Risk Taking: A Meta-analysis,” Psychological Bulletin,
125 (3), 367.
Cameron, A. C. and P. K. Trivedi (2009), Microeconometrics Using
Stata. College Station, TX: Stata Press.
Chikweche, T. and R. Fletcher (2011a), “Branding at the Base of Pyr-
amid: A Zimbabwean Perspective,” in Marketing Intelligence &
Planning, Vol. 29, M. Omar, ed. 247-63. West Yorkshire, England:
Emerald Group Publishing. doi:10.1108/02634501111129239.
Chikweche, T. and R. Fletcher (2011b), “Franchising at the Bottom of
the Pyramid (BOP): An Alternative Distribution Approach,” The
International Review of Retail, Distribution and Consumer
Research, 21 (4), 343-60. doi:10.1080/09593969.2011.588717.
Chikweche, T. and R. Fletcher (2012), “Revisiting the Marketing Mix
at the Bottom of Pyramid (BOP): From Theoretical Considerations
to Practical Realities,” Journal of Consumer Marketing,29(7),
507-20. doi:10.1108/07363761211275018.
Chikweche, T. and R. Fletcher (2014), “Marketing to the “Middle of
the Pyramid” in Emerging Markets Using a Social Network Per-
spective: Evidence from Africa,” in International Journal of Emer-
ging Markets, Vol. 9, Keith Perks and Dr D. Phani Tej Adidam,
eds. 400-23. West Yorkshire, England: Emerald Group Publishing.
Chikweche, T., J. Stanton, and R. Fletcher (2012), “Family Pur-
chase Decision Making at the Bottom of the Pyramid,” Journal
of Consumer Marketing, 29 (3), 202-13. doi:10.1108/
Cohen, A., Q. Zhang, Q. Luo, Y. Tao, Jr J. M. Colford, and I. Ray
(2017), “Predictors of Drinking Water Boiling and Bottled Water
Consumption in Rural China: A Hierarchical Modeling
Approach,” Environmental Science & Technology, 51 (12),
6945-56. doi:10.1021/acs.est.7b01006.
Cook, J., P. Kimuyu, and D. Whittington (2016), “The Costs of Cop-
ing with Poor Water Supply in Rural Kenya: COPING COSTS OF
POOR WATER,” Water Resources Research, 52 (2), 841-59. doi:
Dahlberg (2017), “The Untapped Potential of Decentralized Solutions
to Provide Safe, Sustainable Drinking Water at Large Scale,”
(accessed July 1, 2017), [available at www.safewaterentre].
Davies, I. A. and A. Torrents (2017), “Overcoming Institutional Voids
in Subsistence Marketplaces: A Zimbabwean Entrepreneurial
Case,” Journal of Macromarketing, 37 (3), 255-67. doi:10.1177/
de Queiroz, J. T. M., F. Doria M. de, M. W. Rosenberg, L. Heller, and
A. Zhouri (2013), “Perceptions of Bottled Water Consumers in
18 Journal of Macromarketing XX(X)
Three Brazilian Municipalities,” Journal of Water and Health,11
(3), 520-31. doi:10.2166/wh.2013.222.
Ernst, H., H. N., Kahle, A. Dubiel, J. Prabhu, and M. Subramaniam
(2015), “The Antecedents and Consequences of Affordable Value
Innovations for Emerging Markets: Affordable Value Innova-
tions for Emerging Markets,” Journal of Product Innovation
Management, 32 (1), 65-79. doi:10.1111/jpim.12171.
Francisco, J. P. S. (2014), “Why Households Buy Bottled Water: A
Survey of Household Perceptions in the Philippines: Why House-
holds Buy Bottled Water,” International Journal of Consumer
Studies, 38 (1), 98-103. doi:10.1111/ijcs.12069.
Geroski, P. A. (2000), “Models of Technology Diffusion,” Research
Policy, 29 (4–5), 603-25. doi:10.1016/S0048-7333(99)00092-X.
Grady, C. A., E. C. Kipkorir, K. Nguyen, and E. R. Blachleyt (2015),
“Microbial Quality of Improved Drinking Water Sources: Evi-
dence from Western Kenya and Southern Vietnam,” Journal of
Water and Health, 13 (2), 607. doi:10.2166/wh.2014.206.
Guarı´n, A. and P. Knorringa (2014), “New Middle-Class Consumers
in Rising Powers: Responsible Consumption and Private
Standards,” Oxford Development Studies, 42 (2), 151-71. doi:10.
Hahn, R. (2009), “The Ethical Rational of Business for the Poor –
Integrating the Concepts Bottom of the Pyramid, Sustainable
Development, and Corporate Citizenship,” JournalofBusiness
Ethics, 84 (3), 313-24. doi:10.1007/s10551-008-9711-6.
Hall, J., S. V. Matos, and M. J. C. Martin (2014), “Innovation Path-
ways at the Base of the Pyramid: Establishing Technological
Legitimacy through Social Attributes,” Technovation, 34 (5–6),
284-94. doi:10.1016/j.technovation.2013.12.003.
Jagadale, S. R., D. Kadirov, and D. Chakraborty (2018), “Tackling the
Subaltern Quandary: Marketing Systems of Dignity,” Journal of
Macromarketing, 38 (1), 91-111. doi:10.1177/0276146717740680.
Jaiswal, A. K. and S. Gupta (2015), “The Influence of Marketing on
Consumption Behavior at the Bottom of the Pyramid,” Journal of
Consumer Marketing, 32 (2), 113-24. doi:10.1108/JCM-05-2014-
Karnani, A. (2007), “The Mirage of Marketing to the Bottom of the
Pyramid: How the Private Sector Can Help Alleviate Poverty,”
California Management Review, 49 (4), 90-111. doi:10.2307/
Karnani, A. (2009), “Romanticising the Poor Harms the Poor,”
Journal of International Development, 21 (1), 76-86. doi:10.1002/
Kenya National Bureau of Statistics, Ministry of Health/Kenya,
National AIDS Control Council/Kenya. (2015) Kenya Demo-
graphic and Health Survey 2014. Rockville, MD, (accessed
November 15, 2016), [available at
Kotler, P., N. Roberto, and T. Leisner (2006), “Alleviating Poverty: A
Macro/Micro Marketing Perspective,” Journal of Macromarket-
ing, 26 (2), 233-39. doi:10.1177/0276146706291039.
Lancaster, K. J. (1966), “A New Approach to Consumer Theory,”
Journal of Political Economy, 74 (2), 132-57. doi:10.1086/259131.
¨thje, C., C. Herstatt, and E. Von Hippel (2005), “User-innovators
and “Local” Information: The Case of Mountain Biking,”
Research Policy, 34 (6), 951-65.
Mair, J. and I. Marti (2009), “Entrepreneurship in and Around
Institutional Voids: A Case Study from Bangladesh,” Journal
of Business Venturing, 24 (5), 419-35. doi:10.1016/j.jbusvent.
Marinakis, Y. D., S. T. Walsh Prof dr, and R. Harms dr (2016),
“Catalyzing New Product Adoption at the Base of the Pyramid,”
in 25th International Association for Management of Technology
Conference, IAMOT 2016, pp. 775-94, (accessed February 1,
2017), [available at].
Meagher, K. (2018), “Cannibalizing the Informal Economy: Frugal
Innovation and Economic Inclusion in Africa,” The European
Journal of Development Research, 30 (1), 17-33. doi:10.1057/
Ministry of Water and Irrigation (2017), “Water Boards,” (accessed
June 15, 2017), [available at
Montgomery, N., A. Peredo, and E. Carlson (2012), “The BOP Dis-
course as Capitalist Hegemony,” Academy of Management Pro-
ceedings, 2012 (1), 14505. doi:10.5465/AMBPP.2012.274.
Nakata, C. and K. Weidner (2012), “Enhancing New Product Adop-
tion at the Base of the Pyramid: A Contextualized Model: New
Product Adoption at the BOP,Journal of Product Innovation
Management, 29 (1), 21-32. doi:10.1111/j.1540-5885.2011.
Patsiaouras, G., M. Saren, and J. A. Fitchett (2015), “The Marketplace
of Life? An Exploratory Study of the Commercialization of Water
Resources through the Lens of Macromarketing,” Journal of
Macromarketing, 35 (1), 23-35. doi:10.1177/0276146714538454.
Piacentini, M. and K. Hamilton (2013), “Consumption Lives at
the Bottom of the Pyramid,” Marketing Theory, 13 (3), 397-400.
Prahalad, C. K. (2005), The Fortune at the Bottom of the Pyramid.
Upper Saddle River, NJ: Wharton School Publishing.
Prahalad, C. K. (2012), “Bottom of the Pyramid as a Source of Break-
through Innovations: BOP as Source of Innovations,” Journal of
Product Innovation Management, 29 (1), 6-12. doi:10.1111/j.
Prahalad, C. K. and A. Hammond (2002), “Serving the World’s Poor,
Profitably,” Harvard Business Review, 80 (9), 48-59.
Quansah, F., A. Okoe, and B. Angenu (2015), “Factors Affecting
Ghanaian Consumers’ Purchasing Decision of Bottled
Water,” International Journal of Marketing Studies, 7 (5), 76-87.
Rahman, M., M. R. Hasan, and D. Floyd (2013), “Brand Orientation as
a Strategy That Influences the Adoption of Innovation in the Bot-
tom of the Pyramid Market,” Strategic Change, 22 (3–4), 225-39.
Ramani, S. V. and V. Mukherjee (2014), “Can Breakthrough Innova-
tions Serve the Poor (Bop) and Create Reputational (CSR) Value?
Indian Case Studies,” Technovation, 34 (5–6), 295-305. doi:10.
Ramani, S. V., S. SadreGhazi, and G. Duysters (2012), “On the
Diffusion of Toilets as Bottom of the Pyramid Innovation:
Lessons from Sanitation Entrepreneurs,” Technological Fore-
casting and Social Change, 79 (4), 676-87. doi:10.1016/
Howell et al. 19
Ritter, M., C. Velcine, E. Camille, D. Lantagne, and R. Guillaume
(2017), “Optimizing Household Chlorination Marketing Strategies:
A Randomized Controlled Trial on the Effect of Price and Promo-
tion on Adoption in Haiti,” The American Journal of Tropical Med-
icine and Hygiene, 97 (1), 271-80. doi:10.4269/ajtmh.16-0820.
Sen, A. (2001), Development as Freedom. Oxford, UK: Oxford Uni-
versity Press, (accessed February 1, 2017), [available at http://].
Sheth, J. N. (2011), “Impact of Emerging Markets on Marketing:
Rethinking Existing Perspectives and Practices,” Journal of Mar-
keting, 75 (4), 166-82. doi:10.1509/jmkg.75.4.166.
Silvestre, B. S. and S. Neto R e (2014), “Capability Accumulation,
Innovation, and Technology Diffusion: Lessons from a Base of the
Pyramid Cluster,” Technovation, 34 (5–6), 270-83. doi:10.1016/j.
Smith, S. M. and S. E. Beatty (1987), “External Search Effort: An
Investigation Across Several Product Categories,” Journal of Con-
sumer Research, 14 (1), 83-95. doi:10.1086/209095.
Subrahmanyan, S. and J. Tomas Gomez-Arias (2008), “Integrated
Approach to Understanding Consumer Behavior at Bottom of
Pyramid,” Journal of Consumer Marketing, 25 (7), 402-12.
United Nations Development Programme (2008) “Creating value for
all: strategies for doing business with the poor”. In: Growing Inclu-
sive Markets Initiative, [available at
Va´squez, W. F. (2017), “Understanding Bottled Water Consumption in a
High-poverty Context: Empirical Evidence from a Small Town in Gua-
temala: Understanding Bottled Water Consumption,” International
Journal of Consumer Studies, 41 (2), 199-206. doi:10.1111/ijcs.12327.
Viswanathan, M. and J. A. Rosa (2010), “Understanding Subsistence
Marketplaces: Toward Sustainable Consumption and Commerce
for a Better World,” Journal of Business Research, 63 (6), 535-37.
Viswanathan, M., K. Jung, S. Venugopal, I. Minefee, and I. W. Jung
(2014), “Subsistence and Sustainability: From Micro-Level Beha-
vioral Insights to Macro-Level Implications on Consumption, Con-
servation, and the Environment,” Journal of Macromarketing,34
(1), 8-27. doi:10.1177/0276146713499351.
Water Services Regulatory Board (2014), “Water Tariffs,” (accessed
March 15, 2017), [available at
WHO/UNICEF Joint Water Supply and Sanitation Monitoring Pro-
gramme, World Health Organization, WHO/UNICEF Joint Monitor-
ing Programme for Water Supply, Sanitation, & UNICEF. (2005).
Water for life: making it happen. World Health Organization.
Wooldridge, J. M. (2016), Introductory Econometrics: A Modern
Approach. 6th ed. Boston, MA: Cengage Learning.
Yurdakul, D., D. Atik, and N. Dholakia (2017), “Redefining the Bot-
tom of the Pyramid from a Marketing Perspective,” Marketing
Theory, 17 (3), 289-303. doi:10.1177/1470593117704265.
Zanello, G., X. Fu, P. Mohnen, and M. Ventresca (2016), “The Creation
and Diffusion of Innovation in Developing Countries: A Systematic
Literature Review,” Journal of Economic Surveys, 30 (5), 884-912.
Author Biographies
Rachel Howell is a PhD candidate in Economics at Delft University of
Technology in the Netherlands; in the Economics section. Her PhD is
part of the Leiden Delft Erasmus Centre for Frugal Innovations in
Africa. She has conducted quantitative research on off grid electrifica-
tion and frugal innovation primarily in sub-Saharan Africa. Rachel’s
research interests are in impact evaluation, innovation management,
development economics, behavioural economics and East Africa.
Kinsuk Mani Sinha is an innovation scholar. Her research interests
lie in inclusive innovation, bio-based innovation, value co-creation
and technological business incubator. She has led and participated
in fieldwork across Eastern Africa, India, USA and the Netherlands
to study the role played by farmers in bio-based innovation, innovation
at Bottom of the pyramid, role-played incubators in fostering local
economic development and the learning process of female farmers.
For purpose of data analysis, Kinsuk applies econometric models,
stakeholder analysis and python codes.
Natascha Wagner is Associate Professor of Development Economics
at the Institute of Social Studies, Erasmus University Rotterdam
(Netherlands). Her research interests lie in international economics,
development, health and education. She has participated in various
impact evaluation projects in Africa and Asia ranging from public
health, HIV testing and counselling to support for people living with
HIV as well good governance and rural infrastructure programs. She
has published articles in, among others, Health Economics,Economics
of Education Review,Journal of Development Studies and World
Neelke Doorn is distinguished Antoni van Leeuwenhoek Professor
“Ethics of Water Engineering” at the Department of Philosophy of
Delft University of Technology and former assistant director of the
3TU. Centre of Ethics and Technology. Together with Diane P.
Michelfelder, Neelke is Editor-in-Chief of the peer-reviewed journal
Techne: Research in Philosophy and Technology. Her research inter-
ests are ethics of water engineering, philosophy of technology, water
ethics, resilience, responsible innovation, philosophy of risk and ethics
and design.
Cees van Beers is Professor of Innovation Management and head of
the section Economics of Technology and Innovations. He is also one
of the co-leaders of the Leiden Delft Erasmus (LDE) Centre for Frugal
Innovations in Africa. He holds a doctorate in economics (Ph.D) from
the Free University Amsterdam. Some of his research interests are
determinants of responsible innovation and entrepreneurship, and
inclusive business models for frugal innovations.
20 Journal of Macromarketing XX(X)
... Knizkov and Arlinghaus (2019) found that this market generates economic value for organizations and social value for the communities located in this segment. Including the BOP consumers will help the companies improve their economic, social, and ecological performance (Howell et al., 2020;Rosca & Bendul, 2019). ...
... Even explained that the growth of brands could originate from the preference of poorer consumers at the base of the economic pyramid. When these consumers faced with a brand choice for the first time, their loyalty is low (Chikweche, 2013;Chikweche & Fletcher, 2011;Hillemann & Verbeke, 2014;Nakata & Weidner, 2012), but their brand experiences are perceived as quality products, which increases their loyalty (D'Andrea, 2006;Howell et al., 2020;Prahalad, 2005). According to Kashyap (2016), rural consumers in developing countries take more time to buy a brand, but once bought and convinced of its value; they generate high loyalty rates, a higher percentage than urban consumers ...
... Baishya and Samalia (2020) also highlighted that the BOP market generates a business opportunity for companies that produce smartphones, which should improve their pricing strategies and implement services based on Information and Communication Technologies (ICTs). The findings by Howell et al. (2020) also confirm the results of this research, finding that bottled water companies may only be reaching middle-class consumers and not low-income ones, demonstrating that the BOP market can buy more than one commodity rather than other factors sociodemographic, subject only to income. ...
... Knizkov and Arlinghaus (2019) found that this market generates economic value for organizations and social value for the communities located in this segment. Including the BOP consumers will help the companies improve their economic, social, and ecological performance (Howell et al., 2020;Rosca & Bendul, 2019). ...
... Even explained that the growth of brands could originate from the preference of poorer consumers at the base of the economic pyramid. When these consumers faced with a brand choice for the first time, their loyalty is low (Chikweche, 2013;Chikweche & Fletcher, 2011;Hillemann & Verbeke, 2014;Nakata & Weidner, 2012), but their brand experiences are perceived as quality products, which increases their loyalty (D'Andrea, 2006;Howell et al., 2020;Prahalad, 2005). According to Kashyap (2016), rural consumers in developing countries take more time to buy a brand, but once bought and convinced of its value; they generate high loyalty rates, a higher percentage than urban consumers ...
... Baishya and Samalia (2020) also highlighted that the BOP market generates a business opportunity for companies that produce smartphones, which should improve their pricing strategies and implement services based on Information and Communication Technologies (ICTs). The findings by Howell et al. (2020) also confirm the results of this research, finding that bottled water companies may only be reaching middle-class consumers and not low-income ones, demonstrating that the BOP market can buy more than one commodity rather than other factors sociodemographic, subject only to income. ...
Full-text available
The study is conducted in families of a rural population of the Republic of Ecuador, classified based on their social, cultural, and economic characteristics in the BOP market. Fifty groups were selected from these populations. Two hundred eighteen participants underwent detailed interviews and natural observation. The findings indicated a strong preference for national and traditional products, although also a preference for regional and global developments, showing differences in lifestyle and a combination of consumer needs. It is found that these consumers can be loyal to brands representing different social statuses. Therefore, companies should design more effective marketing strategies to target this market, whose growth is attractive for the development and increase of their sales volume.
... Globally, the BOP is estimated to be more than four billion in number. Although often excluded from the mainstream global capital narrative, the group represents a significant aggregate buying power and the segment requires more research attention (Baishya & Samalia, 2020;Charman & Petersen, 2017;Howell et al., 2020). ...
... The BOP is not one homogenous segment but is heterogeneous and constantly evolving (Dembek et al., 2020). Part of this evolution involves potential mobility between the BOP tiers of extreme poverty, subsistence and low-income groups, which was framed by Rangaan et al. (2011) to reinforce the heterogeneity of the segment (Howell et al., 2020). The extreme poverty and subsistence groups within the BOP represent the majority of the BOP in Africa whose income thresholds of less than US$1, or between US$1 and US$1.90, makes them vulnerable and marginalised. ...
... The BOP is also not always self-defined as vulnerable because of the different levels of wealth even within the segment (Neethling, 2017;Simpson & Lappeman, 2017;Charman & Petersen, 2017). Recent studies on the overall BOP segment have focused on mapping the evolution of the BOP (Dembek, et al., 2020;Follman, 2012;Howell et al., 2020;Kolk et al., 2014). This is further expanded in the section which contextualises the MBOP where we outline our rationale and justification for separating the MBOP from the broad BOP segment. ...
Africa is home to 27 of the world’s 28 poorest, marginalised and most vulnerable countries. There are limited studies on the challenges and opportunities of researching marginalised consumers in Africa, even though more than 440 million marginalised consumers consume products and services that require market research. This paper aims to critically identify and discuss the challenges and opportunities for researching marginalised consumers in Africa using insights from empirical studies conducted in various countries between 2009 and 2022. The paper proposes a conceptual framework that outlines the methodological and operational challenges and enablers for market research and concludes with practical guidelines of considerations that researchers should take for researching the marginalised in Africa.
... In Africa, sachet packaged water is a significant primary drinking water source (Howell et al. 2020, Olukoju 2007, Stoler 2017, Stoler et al. 2013. In Ghana, sachet water constitutes up to 43% of urban and up to 12% of rural household drinking water (Ghana Statistical Service 2009 and 2013, Morinville 2017, Wright et al. 2016. ...
... In Nigeria, sachet water is a primary source for 12% of urban households (National Population Commission 2013). The growing use of bottled water by poor communities in Rwanda, Uganda, Kenya and Ghana is driven by a desire for clean, safe drinking water (Howell et al. 2020, Quansah et al. 2015. ...
Technical Report
Full-text available
The report examines facts and perceptions about bottled water in the global context. It analyses the geography, structure, trends, and drivers of the global bottled water market. It examines the existing knowledge on the quality of bottled water, its impacts on water resources, and its role in plastic pollution. It raises the question of the bottled water industry’s contribution to the sustainable development goal on universal access to safe drinking water. The analysis considered only those types of bottled water that have little or no difference in taste from the tap water provided by regular municipal water supply. It is shown that bottled water is widely consumed in the both Global North and South although prices can be orders of magnitude higher than tap water. The current global bottled water sales are estimated at almost 270 billion US$ and 350 billion liters. The report maps and ranked the top 50 countries in the world by total and per capita bottled water sales both in dollars and liters. The Asia-Pacific region constitutes about half of the global bottled water market, and the Global South countries together about 60%. The USA, China and Indonesia combined comprise half of the global market. Germany is the biggest market in Europe, Mexico in the LAC region and South Africa in Africa. Singapore and Australia stand out as the leaders in both annual revenue and volume of bottled water sold per capita, with the USA and China per capita indicators being much smaller. The report indicates that bottled water market drivers differ significantly between the Global North and the Global South. In the former, bottled water is often perceived as a healthier and tastier product than tap water and is more a luxury good than a necessity. In the Global South, bottled water sales are stimulated primarily by the lack or absence of a reliable public water supply. Based on around 60 case studies from more than 40 countries from every region of the world, the report illustrates that there have been numerous cases of inorganic, organic, and microbiological contamination of hundreds of bottled water brands of all bottled water types and that such contamination often exceeded local or global standards. This represents strong evidence against the misleading perception that bottled water is an unquestionably safe drinking water source and argues that the provision of a safe and reliable drinking water supply in any country may not be achieved at the expense of one water source over another. Withdrawals for bottled water can contribute to groundwater resource depletion in areas of bottled water procurement, although case studies that illustrate this are rare. However, even if such withdrawals are small in absolute terms globally or compared to larger water consumers like irrigated agriculture, local impacts on water resources may be significant. The lack of data available on water volumes extracted by the bottled water industry is largely due to the lack of transparency and a legal foundation that would have forced bottling companies to publicly disclose extracted water volumes and assess the environmental consequences of their activities. The Global South, where safe drinking tap water is not always available, represents potential future markets for bottled water. Lack of national policies for water management may promote uncontrolled groundwater withdrawals for bottled water procurement with little or no contribution to a sustainable long-term drinking water supply. The report collates scattered information on plastic pollution associated with bottled water, pointing out that the world currently generates around 600 billion plastic bottles amounting to approximately 25 million tonnes of plastic waste, which is not recycled but is disposed of in landfills or as unregulated waste. While there are signs of growing social awareness of the adverse impacts of plastics on the environment, a breakthrough solution that could radically reduce the environmental impacts of plastics does not yet appear to exist. Hence plastic pollution will likely continue in the years to come. The report argues that while progress toward universal access to safe drinking water for all is significantly off-track, the expansion of bottled water markets slows this progress down, distracting attention and resources from accelerated public water supply systems development. Estimates suggest that less than half of what the world pays for bottled water annually would be sufficient to ensure clean tap water access for hundreds of millions of people without it – for years. There are recent high-level initiatives that aim to scale up financing for the Sustainable Development Goals, including water-related ones. Such initiatives are an opportunity for the bottled water sector to become an active player in this process and help accelerate the progress toward sustainable water supply, particularly in the Global South.
... The limited focus on actual poverty alleviation and the consideration of BoP mainly as a consumer is noticed, for example, in the case of cell phone trade in Thailand (Pipitwanichakarn & Wongtada, 2019), the case of logistics services for the delivery of orders in Brazilian slums (Duarte, Macau, Silva, & Sanches, 2019), and in the case of bottled water trade in African countries (Howell, Sinha, Wagner, Doorn, & van Beers, 2020), reported in the literature on BoP. ...
Full-text available
There is a need for further research on the value creation from both business and bottom of the pyramid perspectives. In our study, attention was given to this gap and it can be argued that partnerships between sustainable entrepreneurship ventures (SEVs) and bottom of the pyramid (BoP) communities are a relevant approach to creating sustainable value in BoP ecosystems. Accordingly, we propose to find answers to our main research question: How do partnerships between sustainable entrepreneurship ventures (SEVs) and the bottom of the pyramid (BoP) communities create sustainable values and generate benefits and opportunities for BoP? The methodology comprised three case studies of SEV-BoP partnerships in Brazil. Results indicated that SEV-BoP partnerships have the potential to deliver economic, social, and environmental values that enable progress in both businesses and low-income communities. Benefits and opportunities for BoP communities included income generation, professionalization, and business experience, which pushed the BoP communities to develop their enterprises and start new businesses independently. These results show that SEV-BoP partnerships represent an alternative approach to creating sustainable value in the BoP ecosystem, including multidirectional benefits. Keywords: sustainable entrepreneurial ventures; bottom of the pyramid; sustainable values; entrepreneurial opportunities
... Abimbola (2006) presents a strong case for expansion of research on branding to include the African region by arguing that understanding branding is important for marketers who intend to penetrate this market, which is the world's second largest and second most populated continent (Freire 2014). The limited prior research on branding and product design in emerging regions such as Africa has primarily focused on the bottom of pyramid segment, and not necessarily on the middle of pyramid (Akula 2008;Blankson and Coffie 2020;Chikweche and Fletcher 2011;Howell et al. 2020;Kumar, Guruvayurappan, and Banerjee 2007). ...
This research reports on a cross-country based investigation of the floating middle class in Africa demonstrating how their behavior and structure influences the design of marketing mix strategies used by marketers. Questionnaires were used to collect data in ten African cities. Key findings include the importance of branding and core drivers for purchase such as durability, origin, popularity, affordability, and a unique circle of quality variables. Innovative adaptations for pricing, distribution and marketing communications are outlined to assist marketers develop appropriate context relevant strategies to target the floating middle class. The discussion also outlines theoretical and practical implications.
Bottom of the pyramid (BOP) consumers differs in marketplace behavior owing to their socio-economic and long exposure to poverty. The purpose of this study is to summarize the latest literature on consumer behavior studies of the BOP consumers and understand the reasons for the peculiar behavior of BOP consumers. Insights gained from this study are very useful for the organizations and policymakers focusing on the BOP segments. Systematic literature review of previous literature was undertaken using TCCM (Theory, Contexts, Characteristics, Methodology) framework. Findings suggest that BOP literature is still emerging, as it is evident from the recent conceptual papers and qualitative studies. Most of the studies focused on the new product adoption behavior at BOP followed by the aspirational consumption and role of corporate social responsibility. It is one of the literature reviews which focuses on the consumer behavior of BOP consumers holistically rather than focusing on specific aspects of BOP consumer behavior. Insightful directions for organizations and policymakers serving the BOP segments were given.
Full-text available
This is the English version of a chapter entitled "Proposition de valeur mixte (BVP), triple bilan (TBL), création de valeur partagée (CSV) et base de la pyramide (BoP), quelles différences ? Une analyse comparative avec la méthode de Morse" that will be published in french in "Management international et valeurs by Vuibert. It aims to delineate and compare four concepts often used synonymously in management literature: blended value, triple bottom line, creating shared value, and the base of the pyramid using the methodology of Morse.
The Indian bottom of pyramid (BoP) segment contributes around 85% of the total national household market. This study attempts to ascertain the purchase behaviour of customers at the Indian urban BoP. It endeavours to appreciate the viewpoint of the urban BoP consumers in the purchase process with reference to their purchase basket comprising of products mainly across categories such as grocery, perishables and basic consumer durables. The study starts with qualitative grounded theory followed by quantitative survey-based approach. It presents and validates emergent themes to give insights about purchase behaviour of consumers at urban BoP. The empirical findings of the study discovered five consumer motivations through factor analysis. The subsequent result of the cluster analysis showed that the urban BOP market is heterogeneous. Since the size of cluster is substantial, companies must make marketing efforts to target them on a priority basis. The study proposes a conceptual model of consumer motivation supported by the self-determination Theory for the urban BoP market in India.
Full-text available
In recent years, the “Bottom of the Pyramid” (BOP) market has become an attraction for researchers, business managers, and policymakers. Numerous researchers have contributed theoretically and empirically to the BOP literature, however, little focus is accorded to BOP customers’ buying behavior. It is interesting to study the shopping behavior of BOP customers characterized by low disposable income, poor standard of living, and geographical isolation. Therefore, the present study aims to investigate the shopping behavior of BOP customers to identify the drivers and hindrances in the buying decisions. The study follows a qualitative approach by conducting in-depth personal interviews of 100 BOP customers visiting Vishal Mega Marts (VMMs) in a northern Indian city. The results indicate that BOP customers aspire for a better shopping experience like affluent customers, however, income constraint restricts them to buy superior brands. Indeed, BOP customers are persuaded by excessive discounts, gifts, coupons, peer influence, and seasonal offers by VMMs. The study provides implications for researchers, retailers, and policymakers.
Full-text available
The subaltern quandary refers to the failure of a fast-growing economy to improve the abysmal living conditions of marginalized groups. To gain a better insight into this issue, we investigate the subaltern group’s experiences of marketing systems in the context of neo-liberal reforms in rural India. The qualitative analysis of subaltern narratives shows that subaltern experiences are shaped by marketization processes that imbue market relations with new stylized meanings of dignity. Despite these meanings perpetuating limited and distorted constructions, subalterns use them, exemplified in their attempts to minimize their perceived dissimilarity to other marketing system actors, in order to gain access to predominant, albeit flawed, marketing systems. Thus, the status quo is rarely challenged. This research suggests that the subaltern quandary can only be resolved when market development initiatives take human worth as a main goal, while subalterns are empowered with market system creation, design and governance capabilities.
Full-text available
Approximately two billion people drink unsafe water. Boiling is the most commonly used household water treatment (HWT) method globally and in China. HWT can make water safer, but sustained adoption is rare and bottled water consumption is growing. To successfully promote HWT, an understanding of associated socioeconomic factors is critical. We collected survey data and water samples from 450 rural households in Guangxi Province, China. Covariates were grouped into blocks to hierarchically construct modified Poisson models and estimate risk ratios (RR) associated with boiling methods, bottled water, and untreated water. Female-headed households were most likely to boil (RR=1.36, p<0.01), and among boilers those using electric kettles rather than pots had higher income proxies (e.g., per capita TV ownership RR=1.42, p<0.01). Higher-income households with younger, literate, and male heads were more likely to purchase (frequently contaminated) bottled water, or use electric kettles if they boiled. Our findings show that boiling is not an undifferentiated practice, but one with different methods of varying effectiveness, environmental impact, and adoption across socioeconomic strata. Our results can inform programs to promote safer and more efficient boiling using electric kettles, and suggest that if rural China's economy continues to grow then bottled water use will increase.
Full-text available
This case study explores an entrepreneurial venture as it deals with the challenges of operating in Zimbabwe, a country with high levels of poverty, limited institutional frameworks, and no national currency. It investigates how the venture has overcome institutional voids to challenge the existing monopoly, and succeeded in establishing itself, and thrived, in a contracting Sub-Saharan economy. In so doing the case has reshaped the existing marketing system by increasing competition, lowering prices, providing income for more than 2,000 people, and better quality products for the urban subsistence community. Specifically this paper looks at what resources are critical to reshaping subsistence marketing systems, and provides insights into the marketing channels and networks that have led to this change.
Full-text available
The objective of the study was to examine the factors influencing consumers’ choice of bottled drinking water. The survey research design was employed. Questionnaire was used as the data collections instrument. The items measuring the constructs were adapted from the extant literature. A sample size of two hundred and forty (240) bottled water consumers answered the questionnaire. Data was analysed using ANOVA and correlation test. The Statistical Package for Social Sciences Software was used to analyse the data. The results show that there is a relationship between age categories, income groups, educational levels and bottled water buying behaviour in the Ghanaian market. Again, the study found a relationship between perception and beliefs of bottled water usage. Furthermore, quality, brand price, availability and package were found to influence consumers’ choice of bottled water. Recommendations have been provided at the end of the study. Studies of this nature are very rare in Ghana making this study novel.
Full-text available
In this study, we review the literature on the creation and diffusion of innovation in the private sectors (industry and services) in developing countries. In particular, we collect evidence on what are the barriers to innovation creation and diffusion and the channels of innovation diffusion to and within developing countries. We find that innovation in developing countries is about creation or adoption of new ideas and technologies; but the capacity for innovation is embedded in and constituted by dynamics between geographical, socio-economic, political and legal subsystems. We contextualize the findings from the review in the current theoretical framework of diffusion of innovations, and we emphasize how the institutional context typical of developing countries impacts the diffusion itself.
This paper argues that, far from collaborating with informal economic systems and actors, frugal innovation tends to treat informal economies as a pool of workers and organizational resources to be tapped for the benefit of corporate actors. I will examine how frugal innovation models selectively transform informal economic and institutional systems around formal economic interests, reconfiguring informal opportunities and the distribution of gains in ways that promote adverse incorporation of informal actors rather than mutual benefit. I will examine four mechanisms of adverse incorporation operating within frugal innovation models: copying, free-riding, bypassing nodes of accumulation and shifting risk. Drawing on case studies of M-Pesa and micro-insurance, I will illustrate the often selective and disempowering effects of frugal innovation, which operate to reconfigure informal economic systems in ways that divert profits and control away from informal operators. © 2017 European Association of Development Research and Training Institutes (EADI)
Household water treatment can reduce diarrheal morbidity and mortality in developing countries, but adoption remains low and supply is often unreliable. To test effects of marketing strategies on consumers and suppliers, we randomized 1,798 households in rural Haiti and collected data on purchases of a household chlorination product for 4 months. Households received randomly selected prices ($0.11–$0.56 per chlorine bottle), and half received monthly visits from sales agents. Each $0.22 drop in price increased purchases by 0.10 bottles per household per month (P < 0.001). At the mean price, each 1% drop in price increased purchases by 0.45% (elasticity = 0.45). There is suggestive evidence that household visits by some sales agents increased purchases at mid-range prices; however, the additional revenue did not offset visit cost. Choosing the lowest price and conducting visits maximizes chlorine use, whereas slightly raising the retail price and not conducting visits maximizes cost recovery. For the equivalent cost, price discounts increase purchases 4.2 times as much as adding visits at the current retail price. In this context, price subsidies may be a more cost-effective use of resources than household visits, though all marketing strategies tested offer cost-effective ways to achieve incremental health impact. Decisions about pricing and promotion for health products in developing countries affect health impact, cost recovery, and cost-effectiveness, and tradeoffs between these goals should be made explicit in program design.
Marketing literature has remained mostly silent on the issue of conceptualization of poverty, relying instead on the available definitions of the bottom-of-the-pyramid (BOP) “poor” from economics and sociology. Consequently, in marketing theory, the analytic bases and the practical implications of poverty-centered discourses sometimes remain ambiguous. This study provides a broader, culture-linked conceptualization of poverty and BOP from a consumer research perspective, initiating a dialogue on bottom-up approaches to understanding what poverty means through the lenses of the poor. Via qualitative methods such as semistructured individual in-depth interviews, observations, and fieldnotes, deeper insights were sought on how poverty can be defined from the perspectives of the poor. Transcending the economic-only approach, this study contributes to the literature by extending the contours of “felt poverty” and of the “poverty line” beyond the biogenic and stark utilitarian needs and incorporates the sociocultural dimensions of consumption. Our contribution comes from including the effects of the global consumer culture as a major source of social deprivation. Furthermore—while supporting the positions that the definitions of BOP are relative to contexts, cultures, and times—we also situate the discussion of BOP within the broader discourse on globalization of markets and consumption practices.
This research investigated the determinants of bottled water consumption using household survey data from a small, poor town in Guatemala. Hurdle (two-part) models were estimated to account for 73.3% of sampled households that did not consume bottled water. Findings indicated that the vast majority of respondents perceived minimal health risks from drinking bottled water. In contrast, few respondents (3.2%) believed that tap water is totally safe to drink. Estimation results indicated that bottled water consumption was positively associated to health risk perceptions, household income, education and market access. Household size had a negative effect on the likelihood of consuming bottled water. However, once the household had decided to consume bottled water, its consumption increased with each additional household member.
As the disease burden of poor access to water and sanitation declines around the world, the non-health benefits - mainly the time burden of water collection - will likely grow in importance in sector funding decisions and investment analyses. We measure the coping costs incurred by households in one area of rural Kenya. Sixty percent of the 387 households interviewed were collecting water outside the home, and household members were spending an average of two to three hours doing so per day. We value these time costs using an individual-level value of travel time estimate based on a stated preference experiment. We compare these results to estimates obtained assuming that the value of time saved is a fraction of unskilled wage rates. Coping cost estimates also include capital costs for storage and rainwater collection, money paid either to water vendors or at sources that charge volumetrically, costs of treating diarrhea cases, and expenditures on drinking water treatment (primarily boiling in our site). Median total coping costs per month are approximately US$20 per month, higher than average household water bills in many utilities in the United States, or 12% of reported monthly cash income. We estimate that coping costs are greater than 10% of income for over half of households in our sample. They are higher among larger and wealthier households, and households whose primary source is not at home. Even households with unprotected private wells or connections to an intermittent piped network spend money on water storage containers and on treating water they recognize as unsafe. This article is protected by copyright. All rights reserved.