Rethinking adoption theories for developing countries investigating the goodness of fit for M-Pesa and blockchain

(1)

Rethinking adoption theories for developing countries

Investigating the goodness of fit for M-Pesa and Blockchain

DOI and UTAUT2 Revisited

Laura Derks 11935561

Date of submission and version draft 06-06-2018 / final 21-06-2018 MSc. in Business Administration – Digital Business Track

Amsterdam Business School, University of Amsterdam Supervisor Prof. Dr. Hans P. Borgman

(2)

2

Statement of Originality

This document is written by Laura Derks who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

List of Figures

Figure 1 - UTAUT2 model (Venkatesh, 2012) ...25

List of Tables

Table 1 - M-Pesa's success variables ...20

Table 2 - Success factors explained by DOI ...23

Table 3 - Success factors explained by UTAUT2 ...27

Table 4 - DOI and UTAUT2 applied ...30

Table 5 - Missing elements and case success factors ...30

(5)

5

Abstract

Anno 2018, sending money abroad, especially to Sub-Saharan Africa, is still rather expensive and takes up several days. M-Pesa in Kenya is the only serious alternative service offered for money transfers and remittances. It found its way to majority adoption in just a few years. This article explores the goodness of fit of two technology adoption theories in explaining the success of M-Pesa. Namely, the diffusion of innovation (DOI) theory and the unified technology acceptance theory 2 (UTAUT2). Based on the gap found in the explanations of these two theories and its explaining factors, a third perspective is developed to fit the local context. This research shows that market share, regulatory environment, consumer trust and the country readiness for technology are crucial elements to take into consideration for predicting the adoption of technologies in developing countries. The importance of these new elements is shown in the initial application of this new framework on the current developments around blockchain and its possible applications for remittances in developing countries.

Keywords technology adoption, DOI, UTAUT2, developing countries, blockchain, blockchain adoption, remittances, M-Pesa

(6)

6

Introduction

The motive for this study is rise of blockchain-based solutions and their possible applications to remittances (Swan, 2015; Biggs & Diana, 2016; The Economist, 2018; Kshetri, 2017). How can we understand and predict the diffusion of new technologies in developing countries? Thinking about the potential of blockchain for remittances, money transfers made typically by migrant workers to relatives in their home country, we asked ourselves this question. In the absence of an established banking structure or the availability of trusted third parties, this remittance market is now generally dominated by few commercial companies that charge around 7 per cent commission per send transfer. The Sub-Saharan African continent is the most expensive region to receive money, with an average cost of 9.81 per cent (World Bank, 2017). Research cites remittances play a similar role in economic development as Foreign Direct Investment (FDI) and other capital flows (Chami, Fullenkamp & Jahjah, 2005; OECD, 2015). It is argued that mobile money tools have tremendous potential in developing countries (Gosavi, 2015). Hence, blockchain could open up this sector and greatly benefit customers. The question is if, when and how this will happen. Currently, start-ups are setting up such initiatives. However, the way to mass adoption by customers still seems far. An example of this is BitPesa, a Kenyan based blockchain based payment start-up founded in 2013. Despite its advantages, BitPesa shifted their initial focus from remittances to corporates. This research investigates the reasons for and challenges to technology adoption by consumers in developing countries.

A clear example is of a technology successfully adopted in a developing country is M-Pesa, a highly publicised success story of a mobile payment technology service in Kenya. In explaining its growth and success, researchers have primarily taken the diffusion of innovation and technology acceptance perspective, identifying the focus to the bottom of the pyramid by intensive training for understanding the technology by the large, frequently ignored and unbanked population (Foster & Heeks, 2013). Safaricom, M-Pesa’s parent company, is

(7)

7 mentioned as another important country specific reason behind its success (Mas & Morzaxcysky, 2009). These explanations have been challenged by various scholars. Also, what holds in 2007 for mobile payments in Kenya, may very well not hold for blockchain in 2018. Besides, these reasons might not hold for the moderate successes and failures in other (similar) countries. Therefore, it is important for us to understand this adoption process. To gain insights into the chances and challenges to consumer adoption of blockchain for remittances in developing countries, an understanding of prevailing technology adoption theories is essential. Many scholars stress the necessity of theoretical adjustments for different local contexts (Foster & Heeks, 2013; Kumar, van Dissel & Bielli, 1998; Venkatesh, 2012; Baptista & Oliveira, 2015). Therefore, M-Pesa’s successes and challenges are first analysed. The DOI is the most used theory to explore adoption of technologies by individuals (Venkatesh, 2003; Robertson, 1976; Al-Jabri & Sadiq Sohail, 2012). The DOI especially fits this research since previous studies have consistently found evidence of Roger’s (2010) elements to be prominent in explaining mobile services adoption among consumers (Koenig-Lewis et al., 2012; Lopez-Nicolas, Molina-Castillo & Bouwman; 2008; Montazemi & Qahri-Saremi, 2015). So, to understand the adoption process, the diffusion of innovation (DOI) theory is selected as the first theory. However, it is not only important to understand the adoption process, but also the motivation behind the adoption and the acceptation of a technology (Davis, 1980; Venkatesh, 2012). Therefore, also the unified theory of acceptance and use of technology (UTAUT2) is selected. To learn to understand the adoption process and the motivation behind the acceptance of M-Pesa these two theories are separately critically examined. The UTAUT2 also complements the DOI by incorporating individual differences. Therefore, the choice of these two theories lies in their complementary characters and high level of application within the field of global information management.

(8)

8 The purpose of this research is to develop a comprehensive framework and to understand and predict the acceptance and adoption of blockchain-based solutions for remittances in developing countries. Therefore, we provide a new comprehensive theoretical framework predicting and explaining the adoption of blockchain for remittances in developing countries.

Consequently, this research contributes to the extent Information Systems (IS) literature on the diffusion of new technologies in developing countries by adding insights and discussions on the new and nascent blockchain technology to it. Furthermore, it stresses the necessity of local context adjustments, hereby supporting notions from previous literature (Foster & Heeks, 2013; Venkatesh, 2012; Kumar, van Dissel & Bielli, 1998). This is done by presenting a case study on M-Pesa, being a successfully adopted technology in a developing country.

The research is structured as follows, the diffusion of innovation and the technology acceptance theory are explored individually to examine their usability for explaining these successes and challenges. Thereafter, these two prevailing theories in the field of technology adoption and acceptance are separately applied to M-Pesa to explore their (lack of) fitness to explain M-Pesa’s successes and challenges. Based on this, a new theoretical framework is developed, namely a comprehensive theoretical framework that transcends the separate theories and offers a more including perspective based on the case study.

Lastly, we apply the developed framework to the current expectations concerning blockchain solutions. This results in concluding remarks for the prediction of blockchain for remittances in developing countries and leads to a theoretical framework usable for predicting technology adoption in developing countries.

(9)

9

Research Method

As the research question implies, the aim of this study is to develop a comprehensive theoretical framework that survives the out-of-context implementation of traditional technology adoption and acceptance theories. It is conducted along a non-western case study and predictions of new innovative technologies within a non-western context. A qualitative approach is chosen since more qualitative research is needed in the area of technology adoption in developing countries (Foster & Heeks, 2013). To strengthen the reliability and validity of this research, it combines theoretical development with case study applications (Eisenhardt & Graebner, 2007). Furthermore, in depth interviews with business experts and professional research analysts are conducted which were selected with convenience sampling and snowball sampling (Yin, 2013). Due to distance and limited access to the local market this is inevitable. Research Description

As mentioned earlier, the motivation for this study is investigating the possibilities of blockchain-based solutions for the remittance sector. Therefore, this study answers the following research question: “What theoretical framework can best be used to predict and explain the adoption of a blockchain-based solution for remittances in developing countries?” In order to answer this question a few sub-questions are designed.

Firstly, the research revisits M-Pesa and explores its challenges and successes. This part of the research focuses on exploring how current theories explain the success of M-Pesa. The first sub-question therefore: “To what extent are DOI and UTAUT2 fit to explain the success factors of M-Pesa’s adoption in Kenya?” Two highly applied theories in current research are used to explore how they explain the case of M-Pesa, namely the diffusion of innovation (DOI) and the unified technology acceptance and use of technology model 2 (UTAUT2). By conducting this interview validation for new variables is found and missing variables are identified.

(10)

10 The second part of this research develops a comprehensive framework on technology adoption in developing countries. This leads to the second sub-question: “What framework can be designed to predict and explain the adoption of technology innovations in developing countries?” The answer to this question is based upon the following, the explanations of the consumer adoption of M-Pesa emerging from the case study, the separate application of the DOI and UTAUT2 to explain this adaption and foremost the gap in the explanations of these theories to explain the M-Pesa’s success factors. This results in a complementary theory explaining the technology adoption probability in developing countries surviving the out-of-context implementation of the earlier mentioned traditional theories. With this, the research question is answered.

The third part of the research applies the new theoretical framework to blockchain-based solutions for remittances. Answering the third sub-question “How can the new complementary theoretical framework predict and explain the adoption of blockchain for remittances in developing countries?” Therefore, the third part is an initial application of the newly developed framework to a technological innovation developing countries. This results in a complementary theory explaining the probability of success of blockchain-based solutions for remittances in developing countries. With that, the research question is answered.

The new framework is enriched with literature research and is explored along the case study. To strengthen validity and reliability, four interviews and two email conversations are conducted with experts. The interviews are conducted with experts in technology adoption, blockchain, blockchain in developing countries, blockchain entrepreneurs and with local market experts. The combination of inductive and deductive approaches in the case study investigation and expert interviews enrich the newly constructed theoretical framework, developed to explain and understand the prediction of blockchain in developing countries (Yin, 2013). Each of the interviewees has been interviewed for approximately 45 minutes in a

(11)

semi-11 structured interview. The interview questions focus on the expertise of the individual interviewees and the aim was to gain knowledge on the potential of blockchain in developing countries, the factors moderating this and to make sense of the events (Saunders et al. 2016; Saunders and Lewis, 2012, Yin, 2013). Whenever a theory is applied to a situation other than the culture in which it is developed, different conditions apply, which can test the limits of the theory and its applicability (Kumar, van Dissel & Bielli 1998). Furthermore, the interview questions focus on critical factors for adoption of technologies, market characteristics of developing countries and blockchain specific characteristics. All interviews were recorded and transcribed. Furthermore, all data is only used with interviewee’s consent.

Theory triangulation is conducted combining multiple theories and developing a more comprehensive framework. Methodological triangulation complements the research in using multiple data gathering methods such as interviews, questionnaires and articles.

With this research, a significant contribution is made to existing research on technology adoption in developing countries. Consequently, this research contributes to current literature by developing a more universal theoretical framework on diffusion of technology and innovation in emerging markets, which is applicable to blockchain in these markets.

(12)

12

Revisiting the M-Pesa Case: Goodness of Fit of DOI and UTAUT2

To answer the first sub-question “To what extent are DOI and UTAUT2 fit to explain the success factors of M-Pesa’s adoption in Kenya?”, this chapter investigates to what extent two prevailing and highly used theories are separately fit to explain successes and challenges of M-Pesa. Literature research and expert interviews on the case are conducted to fill in gaps and add variables to explain the adoption of technology in developing countries. This is done along the case of M-Pesa, a money transfer system developed by Safaricom in Kenya in 2007. This is the first technological innovation to achieve mass adoption in a developing country within a few years. Given the size of its success and its specific local context, M-Pesa is the selected case study (Foster & Heeks 2013; Eisenhardt & Graebner, 2007).

As mentioned earlier, the choice has been made to apply both the DOI and the UTAUT2 separately to investigate their fitness in explaining M-Pesa’s factors for success and its challenges. The DOI is considered to be a useful framework for investigating the adoption of innovations and the UTAUT2 is increasingly gaining popularity among IS scholars for examining consumer focussed issues (Huff & Munro, 1985; Venkatesh, 2003; Tamilmani, Rana & Dwivedi, 2017). The DOI is the most used theory to explore adoption of technologies by individuals (Venkatesh, 2003; Robertson, 1976; Al-Jabri & Sadiq Sohail, 2012). Multiple previous studies have consistently found evidence of Roger’s (2010) elements to be prominent in explaining mobile services adoption among consumers (Koenig-Lewis et al., 2012; Lopez-Nicolas, Molina-Castillo & Bouwman; 2008; Montazemi & Qahri-Saremi, 2015). Furthermore, the DOI is used in multiple researchers investigating the adoption of mobile money systems in developing countries (Al-Jabri & Sadiq Sohail, 2012; van der Boor, Oliveira & Veloso, 2014; Tobbin & Kuwornu, 2011; Lin, 2011). One research applies the DOI to M-Pesa to identify the role of the early adopters (Ngugi, Pelowski & Ogembo, 2010).

(13)

13 Previous research combined the DOI and the UTAUT to identify the factors responsible for the rapid adoption of M-Pesa (Kinuthia, 2016). However, the UTAUT2 is increasingly gaining popularity among IS researchers (Tamilmani, Rana & Dwivedi, 2017). This theory builds upon eight technology adoption models, including worldwide recognised technology acceptance model (TAM) by Davis (1989), DOI and UTAUT, but focuses specifically on consumer adoption (Venkatesh, 2012). This research follows the suggestion of Venkatesh (2012) to explore the UTAUT2 in different countries and technologies and identify relevant factors to extend it and integrate it into a new framework (Venkatesh, 2012; Baptista & Oliveira, 2015). Also, the UTAUT2 has been used to understand mobile banking, e-government services and online services in various researches by integrating it with other theories like Hofstede’s cultural moderators and also revisiting UTAUT2 in order to provide new insights (Baptista & Oliveira, 2015; Lian, 2015; Slade, Williams & Dwivedi, 2013; Morosan & DeFranco, 2013).

These theories have been widely accepted in and applied to many different research fields. Kiwanuka (2015) combines the theories in order to come up with a complete model explaining the adoption process (Kiwanuka, 2015). Also, Oliveira et al. (2016) combine the strengths of UTAUT2, DOI and cultural moderators successfully in order to explore mobile payments (Oliveira et al., 2016). For these reasons these two theories are selected.

As the DOI investigates the innovation’s factors affecting the individual’s technology acceptance over time, the UTAUT2 investigates the acceptance and use of technology in the consumer context. It integrates the characteristics of the innovation itself in perceived advantage in performance expectancy, trialability in performance expectancy, observability in performance expectancy, complexity in effort expectancy, and compatibility in social influence from the DOI (Kiwanuka, 2015). Some scholars argue that the UTAUT2 model is more complete in explaining the behavioural intention and technology adoption (Martins et al., 2014; Venkatesh, 2012). This research also explores the correctness of this statement.

(14)

14 The aforementioned theories are first applied to M-Pesa separately, since an immediate combined application of the theories together could lead to a more subjective and biased new model, weakening the theoretical development of this new theoretical framework. Shortcomings resulting from a combined application could be the direct result of the way of combining, instead of an objective analysis of the theories themselves.

The first paragraph of this chapter explores the success factors of the M-Pesa’s case. Within the second and third paragraph, the DOI and UTAUT2 are separately explained and applied to M-Pesa. Thereafter a reflection on the findings follows and conclusions are drawn upon these. Lastly, the foundation for the new theoretical framework is constructed.

M-Pesa in 2018: Successes and Challenges

In the current scientific literature there is a discussion going on whether M-Pesa resulted in higher financial inclusion of the unbanked people in Kenya (Barasa & Lugo, 2015; Omanga & Dreyer, 2017). In IS journals it is stated that M-Pesa is a clear example of a technology used in a simple way by the ‘bottom of the market’ to gain scale and disrupt the existing market (Omanga & Dreyer, 2017; Foster & Heeks, 2013). This results in a significant impact on the financial behaviour of consumers (Bwemelo, 2018). The debate discusses whether blockchain could do the same. Swan (2015), Underwood (2016), Scott (2016), Kewell et al. (2017) and Kshetri (2017) suggest the potential of blockchain-based solutions for developing countries in various sectors, from social, to political to humanitarian ones. Variating from land registration ownership administration to health care to electronic money solutions (Swan, 2015; Underwood, 2016; Kewell et al., 2017; Kshetri, 2017). This study explores the opportunities of blockchain-based solutions as a new innovation for mobile money technologies in developing countries.

Currently, most research on global information management focuses on the economic and social impact of mobile money systems. These studies involve needs assessment, the design

(15)

15 of mobile money systems and the impact of assessments (Harry et al., 2014). However, these studies do not show an extensive, inclusive explanation for the adoption of technology innovations regarding mobile money systems in developing countries. This is supported by Duncombe and Boateng (2009), Dahlberg et al. (2008) and Dahlberg et al. (2015), stating that a better theoretical foundation is required for analysing mobile payments research.

In order to develop a framework for predicting the adoption of blockchain by consumers for remittances in developing countries, previously adopted innovations need to be analysed (Foster & Heeks 2013). M-Pesa is currently used by about 80 per cent of the adult population in Kenya (Finclusion, 2016). The service allows registered users to purchase electronic money (e-money) and send this to other (non-)registered users or withdraw it at a M-Pesa agent. The e-money can be used for various complementary services such as paying bills, school fees, electricity, the gardener and the use of a taxi (M-Pesa, 2018; Ratha, 2011; Reinartz, 2018). Furthermore, the platform extended its services, to be used for example for microloans, called M-Shwari (Vodafone, 2018; Reinartz, 2018). M-Pesa currently has 30 million users in 10 different countries. In 2016 there were 6 billion transactions made through this money system (Safaricom, 2017). Jack and Suri (2011) estimate that M-Pesa as a mobile money system increased the per capita consumption level and lifted 194,000 households - about 2 percent of the Kenyan population - out of poverty. At the moment, M-Pesa is Safaricom’s largest revenue growth driver, and it contributes to over half of the total service revenue growth (Safaricom, 2017). M-Pesa is expanding to other countries, but not with the same rate of success as in Kenya. In 2015, they withdrew their activities in South Africa after limited success. Additionally, it expended to Tanzania, but it only managed to attract a maximum of 250.000 customers in the first few years (Mas & Morawczynski, 2009). This indicates the presence of special local conditions in Kenya. Knowledge on M-Pesa’s factors of success can be used for

(16)

16 learning more about (future) possibilities for similar blockchain solutions in developing countries.

The first important factor explaining M-Pesa’s success is the large unsaturated financial market in Kenya in 2007 (Mas & Radcliffe, 2011; Mas & Radcliffe, 2010; Mas & Morawczynski, 2009; Ngugi, 2010). Mas and Morawczynski (2009) state that the large demand for a domestic money transfer service encouraged the quick growth of M-Pesa. This is supported by Mas and Radcliffe (2011). In 2007, 17 per cent of the households depend on remittances as their primary income.

Another factor leading to the success of M-Pesa was the failure of existing financial institutions to meet the needs of the unbanked people (Mas & Radcliffe, 2011; Mas & Radcliffe, 2010; Mas & Morawczynski, 2009; Ngugi, 2010). Comparing the population being excluded in 2006 to 2016, nine years after foundation of M-Pesa, the financially excluded population of Kenya decreased with from 41.3 per cent to 17.4 per cent. Comparing this to other Sub-Saharan African countries Kenya is, after South Africa, the most financially included country. In an interview with the previous deputy CEO of Safaricom’s main competitor, Telkom, he states that “... the reason why M-Pesa a success is, is that the service focussed on the unbanked audience.” (P. Reinartz, personal communication, May 17, 2018).

A third condition contributing to the success of M-Pesa in Kenya is the large market share of about 79 per cent Safaricom enjoyed in the country in 2007. This size allows the flexibility to try out new things without fear (Ngugi et al., 2010; Mas & Morawczynski, 2009; Mas & Radcliffe 2010). This is supported by Reinartz and Giodini (Red Cross blockchain team lead). Furthermore, is one of the reasons why M-Pesa and similar services have moderate success elsewhere. In order countries in Africa the market is more balanced (P. Reinartz, personal communication, May 17, 2018; S. Giodini, personal communication, May 02, 2018).

(17)

17 The fourth reason that allowed Safaricom to set up M-Pesa successfully is its alliance with the Central Bank of Kenya (CBK) (Hayes & Westrup, 2010; Mas & Radcliffe, 2010; Mas and Morawczynski, 2009; Jack & Suri 2011; Ngugi et al., 2010; Alexander, Mas & Radcliffe, 2011). The CBK published an article explaining their supportive relationship with M-Pesa as a “step towards making financial services accessible to all Kenyans who have access to a mobile phone” (Daily Nation, 2008). Safaricom and CBK agreed upon depositing all customer funds in a trust, a regulated financial institution of the CKB. According to the bank this was done for legal protection and satisfaction for the banking sector (Mas & Ng’weno, 2010). Reinartz (2018) even goes further and states that the bank never gave approval but was simply too late to disapprove and was forced into an alliance with Safaricom (P. Reinartz, personal communication, May 17, 2018). This is also a reason why M-Pesa or similar services have moderate success in other countries. The banks keep close track on similar activities to prevent making the same mistake (P. Reinartz, personal communication, May 17, 2018).

The fifth success factor is Safaricom’s focus on its customers. M-Pesa puts a lot of effort in lowering the adoption barriers for customers. Much time is spent on explaining the service and providing trainings. It is free to register, there is no minimum balance, it is possible to send money to non-registered users and prices are low and transparent (Mas and Radcliffe, 2011; Mas & Ng’weno, 2010). Giodini emphasises the importance of customer engagement and education regarding the technological services (S. Giodini, personal communication, May 02, 2018).

The sixth variable identified is the consumer trust in M-Pesa and Safaricom’s brand. This relates to the comfortability of consumers to deposit and send money via M-Pesa, which is supported in case studies on M-Pesa and in research in technology adoption in developing countries. (Mas and Ng’weno, 2010; Lepoutre & Oguntoye, 2017; Morawczynski & Miscione, 2008; Sadoulet & Furdelle; 2015; Al Sukkar and Hasan, 2005). According to Esser (2018)

(18)

18 “trust is a big advantage”, which is validated by Giodini (2018) (A. Esser, personal communication, May 16, 2018; S. Giodini, personal communication, May 02, 2018).

The seventh factor contributing to M-Pesa’s success is minimised consumer adoption barriers in Kenya, including a simple interface and the option to try out the service before adopting it (Mas & Radcliffe, 2010; Hughes and Lonie, 2007; Lepoutre and Oguntoye, 2017). This results in high trialability and a low complexity element in the DOI. Which is opposed by Al-Jabri and Sadiq Sohail (2012) stating that these two elements do not weigh as much in developing countries. However, Esser (2018) emphasises the importance of network effect and worth-of-mouth, which is in line with the trialability.

The eighth factor contributing to M-Pesa’s success is network externalities or network effect. The value of the service increases when a new member is added (Lepoutre and Oguntoye, 2017). Esser and Giodini validate this element emphasising the importance of word-of-mouth (A. Esser, personal communication, May 16, 2018; S. Giodini, personal communication, May 02, 2018).

The ninth factor leading to M-Pesa’s success, confirmed in a conversation with Reinartz (2018), is the inactivity of official regulators. Three official regulators regarding competition, finance and telecommunication in Kenya are “sleeping-on-the-job” (Reinartz, personal communication, May 17, 2018). Respectively, the competition authority of Kenya, the financial regulations institute and lastly the telecommunication authority of Kenya. Table 1 below provides an overview of the identified factors relevant for the success of M-Pesa integrated with interviews expert respondents.

Success factors

M-Pesa Scholars Code Interviewee Area

1.Large

unsaturated market Mas and Radcliffe, 2011 Mas and Morawczynski, 2009

Ngugi, 2010

(19)

19

2. Failure to meet needs consumers by banks

Mas and Radcliffe, 2011 Mas and Morawczynski, 2009

Ngugi, 2010

Banks failure Reinartz (2018):

“M-Pesa’s focus op de unbanked audience”

Environmental

3. Large market

share of the brand Mas and Radcliffe, 2011 Mas and Morawczynski, 2009;

Ngugi, 2010 Alexandre, Mas and Radcliffe, 2010 Sadoulet and Furdelle, 2015

Market share Giodini (2018):

“It has to be connected to something that is very wide spread solution” Reinartz (2018) Goldie-Scot (2018) Environmental 4. Alliance with Central Bank of Kenya

Alexandre, Mas and Radcliffe, 2011 Mas and Morawczynski, 2009

Jack and Suri 2011 Ngugi et al., 2010 Lepoutre and Oguntoye, 2017

Mas and Ng’weno, 2010

Alliance with bank Reinartz (2018):

Not an alliance. There was no choice by the Bank Esser (2018)

Giodini (2018)

Organisational

5. Focus on

customer support Jack and Suri, 2011 Mas and Radcliffe, 2011 Customer support Giodini (2018): Educational engagement in regarding the service “Of the things you always

have to take into account is one, community

engagement…”

Organisational

6. Trust in brand Mas and Ng’weno, 2010 Lepoutre and Oguntoye, 2017

Morawczynski and Miscione, 2008 Sadoulet and Furdelle, 2015

Al Sukkar and Hasan, 2005 Trust Giodini (2018) Reinartz (2018) Esser (2018) Individual 7. Low adoption barriers (simple interface,

trialability and low prices)

Hughes and Lonie, 2007 Lepoutre and Oguntoye, 2017

Mas and Ng’weno, 2010

(20)

20

8. Network externalities (effect)

Lepoutre and Oguntoye, 2017

Network effect Giodini (2018):

Network effect:

“Peer-to-peer validation of the product”

“social networks between people”

Esser (2018):

“most successful models come from word of mouth”

Individual

9. Inactivity official regulators

Regulators inactivity Reinartz (2018): Unawareness of the regulators

“Verschillende regulators die in dit domein actief zijn of zouden moeten zijn, zijn dat nooit geweest.”

Environmental

Table 1 - M-Pesa's success variables

As seen in the previous paragraph, there are many factors explaining M-Pesa’s success. To answer the first sub-question stated at the beginning of this chapter, this paragraph investigates to what extend the DOI explains the success of M-Pesa. First the DOI is explained, after which it is applied to M-Pesa.

Diffusion of Innovation

This paragraph explores to what extent the diffusion of innovation (DOI) theory, developed by Rogers (1976), is fit to explain M-Pesa’s success in Kenya.

In order to explain how new ideas diffuse among societies, Rogers’ (1976) DOI argues that individuals within a society show different levels of willingness to adopt a new innovation (Rogers, 2003). Within this society, Rogers (1976) distinguishes five categories of customers, based on their relative probability to try out new things; innovators, early adopters, early majority, late majority and laggards. Rogers (2010) incorporates several characteristics affecting the rate of adoption of an innovation in the different consumer categories: observability, trialability, complexity, compatibility and relative advantage. First of all, observability is the degree to which the results of a new innovation are visible to others. With trialability Rogers (2010) refers to the degree to which a new innovation can be tested or

(21)

21 experimented before full commitment needs to be made. Furthermore, the element of complexity is to be understood as the degree to which a new innovation is perceived as difficult to use and understand. With the fourth element of compatibility is meant the degree to which an innovation is consistent with the values and experiences of the consumer. Finally, relative advantage is to be understood as the degree to which an innovation is perceived as better than the idea it replaces (Rogers, 2003; Rogers, 2010). The market share increases when the early majority starts adopting the new idea or innovation. New innovations are more likely to be adopted by consumers who show a high degree of observability, trialability, relative advantage and compatibility, while displaying a lower degree of complexity.

DOI explanation of M-Pesa

In order to analyse to what extent the application of DOI is fit enough to explain M-Pesa’s success, its nine success factors (large unsaturated market, failure to meet needs consumers by banks, large market share, alliance with the KCB (Kenya Central Bank), focus on customer support, trust in the brand, low adoption barriers, network effect and activity official regulators) are explained through the DOI’s five adoption characteristics (observability, trialability, complexity, compatibility and relative advantage).

From the customer perspective, a large unsaturated market comes from the lack of alternative solutions. Therefore, the first success factor of M-Pesa can be explained by the use of the concept relative advantage. The market generates a growing demand for services. (Ngugi et al., 2010)

Also, from the customer perspective, the failure to meet needs consumers by banks, being the second success factor of M-Pesa, can be explained by use of the concept relative advantage. The focus of the banks and other services on the higher layers in the pyramid (Reinartz, 2018) leads to no serious alternative, resulting in a high relative advantage (Al-Jabri & Sadiq Sohail, 2012).

(22)

22 The large market share, the third success factor of M-Pesa, cannot be explained through the use of a concept of DOI.

From a customer perspective, the alliance with the KCB, being the fourth success factor of M-Pesa, can be partially explained by the use of the concept compatibility as an individual characteristic. Even though, compatibility is low and not in line with previous values and experiences of consumers. Compatibility would lower the adoption probability. Therefore, this would imply a lower probability of success. Also, the alliance with the KCB emphasises the influence of external factors and therefore the success is only explained partially by the DOI.

The focus on customer support, being the fifth success factor of M-Pesa, can be partially explained through the use of the concept trialability and complexity.

The trust in the brand, the sixth success factor of M-Pesa, can be partially explained by the use of the concept compatibility. The trust in the brand includes consistency with the values and experiences of the consumer. However, this is opposed by Esser (2018), emphasising the preference in cash. Therefore, compatibility only explains partially the trust in brand factor.

The low adoption barriers (simple interface and trialability), the seventh success factor of M-Pesa, can be partially explained by use of the concept observability and trialability. However, observability does not automatically imply a simple interface and the possibility to try out the service despite being able to see the service. Therefore, observability only explains partially this success element.

The eighth success factor of M-Pesa, network externalities or effect, can be explained by the use of the concept observability. Since Esser (2018) and Giodini (2018) emphasise the importance of network externalities regarding word-of-mouth (A. Esser, personal communication, May 16, 2018; S. Giodini, personal communication, May 02, 2018).

The inactivity of official regulators, being the ninth success factor of M-Pesa, cannot be explained by an element of DOI. However, Zhu et al. (2006) suggest the importance of the

(23)

23 regulator environment for the diffusion of innovation in developing countries. Reinartz (2018) validates this by emphasising the unawareness of multiple regulators (financial regulator, competition authority and telecommunication regulator Kenya) as an important factor for Safaricom to be able to set up M-Pesa. (Reinartz, personal communication, May 17, 2018; Zhu et al., 2006).

In the text above, it is shown that the DOI elements lack in explaining all nine success factors identified for M-Pesa. It lacks to explain the factors large market share and inactivity official regulators. The DOI elements only partially explain low adoption barriers, alliance with KCB, large unsaturated market and trust in brand. However, the DOI successfully explains network externalities and large unsaturated market.

Fully explained Partially explained Not explained

Network externalities

Large unsaturated market Low adoption barriers Alliance KCB Large unsaturated market Trust in brand

Large market share

Inactivity official regulators

Table 2 - Success factors explained by DOI

The next paragraph explores the extent to which another prevailing and highly applicated theory, the UTAUT2, is fit to explain M-Pesa’s success by its separate application to the case.

As demonstrated above, the DOI is on its own not fit enough to fully explain M-Pesa’s success factors. Mainly because the DOI focuses on the innovation itself, which is not the only reason behind its success. Since the UTAUT2 model is considered to be the most complete model to explain behavioural intention and technology adoption (Martins et al., 2014; Venkatesh, 2012) it is explored if it explains all elements of M-Pesa better than DOI. Therefore, this paragraph investigates to what extend the UTAUT2 is well fit to explain these success factors on its own. First, the UTAUT2 is explained, after which it is applied to M-Pesa.

(24)

24 UTAUT2

The UTAUT2 is suitable for explaining the technology adoption process. It is built upon eight earlier recognised adoption models and studies the acceptance and use of technologies in a consumer context (Venkatesh, 2012). UTAUT2 is an extension of the UTAUT model with a focus on the consumer context. The UTAUT2 complements the DOI by incorporating individual differences. The elements influencing the behaviour intention according to the UTAUT2 are performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, habit and experience.

Performance expectancy is the degree to which using a technology will provide benefits to consumers in performing certain activities (Venkatesh, 2012). Effort expectancy is the degree of ease associated with consumers’ use of the technology (Venkatesh, 2012). Social influence is the extent to which consumers perceive that important others believe they should use a particular technology (Venkatesh, 2012). Facilitating conditions refer to consumers’ perceptions of the resources and support available to perform a behaviour (Venkatesh, 2012). Hedonic motivation is the enjoyment experienced when using the technology (Venkatesh, 2012). Price value is the trade-off consumers make between the perceived benefit and the monetary cost of using it (Venkatesh, 2012; Dodds et al., 1999). Habit is the prior behaviour and measured as the extent to which individuals believe the behaviour to be automatic (Venkatesh, 2012; Kim & Malhotra, 2005; Limayem et al., 2007). Age, gender and experience moderate the relationships. Since experience is the only moderating element having also a direct influence on the use behaviour, this is the only moderator taken into consideration for the case study (Venkatesh, 2012).

(25)

25 Figure 1 - UTAUT2 model (Venkatesh, 2012)

UTAUT2 Explanation of M-Pesa

In order to analyse to what extent the application of UTAUT2 is fit enough to explain M-Pesa’s success, its nine success factors (large unsaturated market, failure to meet needs consumers by banks, large market share, alliance with the KCB, focus on customer support, trust in the brand, low adoption barriers, network effect, activity official regulators) are explained through the UTAUT2’s adoption elements (performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, habit and experience).

The large unsaturated market in Kenya, being the first success factor of M-Pesa, cannot be explained by the use of the UTAUT2 elements, because UTAUT2 only incorporates the individual adoption behaviours (Venkatesh, 2012).

The failure to meet needs consumers by banks (consumer side) is the second success factor of M-Pesa and can be explained by the use of the concept of performance expectancy. The focus of the banks and other services on the higher layers in the pyramid leads to no serious alternative, resulting in a high-performance expectancy (P. Reinartz, personal communication, May 18, 2018).

(26)

26 The alliance with the KCB, being the fourth success factor of M-Pesa, cannot be explained by use of a concept of UTAUT2, because UTAUT2 only incorporates the individual adoption behaviours (Venkatesh, 2012).

The focus on customer support, being the fifth success factor of M-Pesa, can be explained through the use of the concept facilitating conditions. This is because focus on customer support is in line with the definition of facilitating conditions, consumers’ perceptions of the resources and support available to perform a behaviour (Venkatesh, 2012).

The trust in the brand, being the sixth success factor of M-Pesa, cannot be explained by the use of a concept of UTAUT2 since no element of UTAUT2 implies a trust element.

The low adoption barriers (simple interface, trialability), being the seventh success factor of M-Pesa, can be explained by the use of the concept of effort expectancy and facilitating conditions.

The network externalities or effect, the eighth success factor of M-Pesa, can be explained through the use of the concept of social influence. According to Esser (2018) and Giodini (2018), word-of-mouth is significantly important and can also be explained through the concept of social influence. Social influence is defined as when an important consumer adopts is, others will follow (A. Esser, personal communication, May 16, 2018; S. Giodini, personal communication, May 02, 2018; Venkatesh, 2012).

The inactivity of official regulators, being the ninth success factor of M-Pesa, cannot be explained by the use of a concept of UTAUT2.

In the text above, it is shown that the UTAUT2 elements lack in explaining all nine success factors identified for M-Pesa. It lacks to explain the factors large market share, trust in brand, inactivity official regulators, large unsaturated market. The DOI elements only partially explain alliance with KCB. However, the DOI successfully explains failure to meet needs consumers, network externalities, focus on customer support and low adoption barriers.

(27)

27

Fully explained Partially explained Not explained

Failure to meet needs consumers Network externalities

Focus on customer support Low adoption barriers

Alliance with the KCB Large market share Trust in brand

Inactivity official regulators Large unsaturated market

Table 3 - Success factors explained by UTAUT2 Reflection

The DOI fits to explain consumer trust. DOI only partially explains unsaturated market, failure to meet needs consumers, network effect, low adoption barriers, alliance with KCB and focus on consumer support. It fails to explain the large market share and inactivity of official regulators.

The UTAUT2 fits to explain consumer need, network effect, low adoption barriers and focus on consumer support. UTAUT2 only partially explains alliance with KCB and unsaturated market. It fails to explain consumer trust, large market share and inactivity official regulators.

The combination of these two theories shows an explanation of the success factors that could not be explained with the application of both theories separately. Concluding, the DOI and UTAUT2 complement each other in explaining consumer trust, failure to meet needs consumers, network effect, focus on consumer support and adoption barriers. Alliance with the KCB is only partially explained. However, this still leaves the factors of regulatory environment and market share unexplained by a combination of both theories. Thus, even the two theories combined fail to explain all success elements of M-Pesa. Therefore, the next section states which elements need to be included within the new theoretical framework.

The first element that needs to be included within the new model is the regulatory environment. Identified in the case analysis and during interviews, this is an important element leading to a successful adoption of a technological product. Zhu et al. (2006) support this view, stressing this factor is more important in developing countries than in developed countries (Zhu

(28)

28 et al., 2006; P. Reinartz, personal communication, 2018; A. Esser, personal communication, 2018).

The second element to be included within the new theoretical framework is market share. This success factor could not be not explained through the use of either DOI and UTAUT2, but it is argued to be important in the case analysis, literature review and interviews.

The third element to be included in the model is market saturation. The availability of alternative solutions within a specific market determines the probability of adoption by consumers.

A fourth element influencing the success of technology adoption to be included is a country’s technology readiness. At time of the development of the DOI theory, Rogers (1962) stated the diffusion process to be cross-cultural (Rogers (1972; Rahim, 1961; Deutschmann & Borda, 1962). However, there are some recent applications of the theory of DOI in developing countries emphasising the need for modifications to fit local contexts (Al-Jabri & Sadiq Sohail, 2012; Zhu, Kraemer & Xu, 2006; Meade & Islam, 2002; Talukdar, 2002). Technology readiness relates to the infrastructure, other relevant systems, skills and previous technology experience in a country (Rogers, 2010; Meade & Islam, 2002; Talukdar, 2002). Researchers state this is a factor strongly influencing technology adoption in developing countries (Zhu et al., 2006; Mustonen-Ollilia & Lyytinen, 2003; Mustonen-Ollila & Lyytinen, 2003; Musa, 2006). Reinartz (2018), Esser (2018) and Giodini (2018) validate this by stating that infrastructure is especially important in developing countries. Particularly when the service that is being offered depends on the availability of a technological service (Reinartz, personal communication, May 17, 2018; Giodini, personal communication, May 02, 2018; A. Esser, personal communication, May 16, 2018).

A fifth element contributing to the success of technology adoption not being explained through the use of DOI or UTAUT2, is consumer trust. Following from the case analysis, trust

(29)

29 in Safaricom and the lack of trust in alternative solutions determined the success of M-Pesa. Supported by researches and interviewees (Mas and Ng’weno, 2010; Lepoutre & Oguntoye, 2017; Morawczynski & Miscione, 2008; Sadoulet & Furdelle; 2015; Al Sukkar and Hasan, 2005; A. Esser, personal communication, May 16, 2018; S. Giodini, personal communication, May 02, 2018).

These above-mentioned five factors influencing technology adoption need to be included in the newly constructed theoretical framework, for they could not have been explained by using both DOI and UTAUT2.

Throughout the case study research, some elements were found not to be important in determining M-Pesa’s success. In an interview with BitPesa co-founder Goldie-Scot he argues that complexity is not a barrier for consumer adoption in Kenya (Goldie-Scot, personal communication, May 30, 2018). This view is supported by Al-Jabri & Sadiq Sohail (2012) who state that complexity is not determining for the adoption probability of a technology in developing countries.

Second, Al-Jabri and Sadiq Sohal (2012) suggest trialability is not important for the adoption process of a technological service. However, following the case study, complexity and trialability appear to be important. Therefore, we suggest future research should investigate these factors to find proof of relevance.

Third, the experience of the consumers would only negatively influence the intention to use M-Pesa, since it was a technology the people were not used to.

Furthermore, hedonic motivation, habit and price value did not come up as important elements determining the success of M-Pesa. Therefore, we suggest future research should investigate all the above-mentioned factors to find proof of the relevance of these factors.

This chapter integrates the DOI and UTAUT2 into an overarching and more comprehensive theory, complemented by variables that appear to be important within the

(30)

30 context of developing countries. This paragraph presents this combined and revised theory and by doing so, answers the first sub-question: “To what extent are DOI and UTAUT2 fit to explain the success factors of M-Pesa’s adoption in Kenya?”.

DOI explained success factors

UTAUT2 explained success factors

Gaps

Fully

Network externalities Large unsaturated market

Fully

Failure to meet needs consumers Network externalities

Focus on customer support Low adoption barriers

DOI

Large market share Inactivity official regulators

UTAUT2

Large market share Trust in brand Inactivity official regulators Partially

Low adoption barriers Alliance KCB

Large unsaturated market Trust in brand

Partially

Alliance with the KCB

Table 4 - DOI and UTAUT2 applied

Following table 4, there are several success elements of M-Pesa that cannot explained by both theories. These are market share, consumer trust and regulatory environment. Hereby the above mentioned sub-question is answered and its summarised in table 5.

Theories’ elements Success factors not explained by theories

DOI Complexity Trialability Relative advantage Compatibility Observability UTAUT2 Performance expectancy Effort expectancy Social influence Facilitating conditions Hedonic motivation Price value Habit Experience Not explained Market share Regulatory environment Partially explained Consumer trust Alliance with KCB

(31)

31

Towards a New Theory for Technology Adoption in Developing

Countries

This chapter presents a new theoretical framework for explaining technology adoption, as a result of the outcomes of the previous chapter. This model integrates the validated concepts of the case study, DOI and UTAUT2 into a new theoretical framework for technology adoptions in developing countries. The new, more comprehensive perspective presented below can be applied to different technology adoption processes within different contexts. With this, the second sub-question is answered “What framework can be designed, based on the case study, to predict and explain the adoption of technology innovations in developing countries?”

Four elements of the DOI and UTAUT2 are integrated due to overlap and to fit the global developing countries context. First, the element alliance with the KCB integrates into the element regulatory environment element and second, effort expectancy integrates with the complexity.

Next, the added elements in the newly constructed theoretical framework are briefly explained. The first element resulting from the case study is consumer trust. Alalwan, Dwivedi and Rana (2017), Lin (2011), Hanafizadehetal (2014), Gefen et al. (2003), Zhou (2012) A. Esser (personal communication, May 16, 2018) and S. Giodini (personal communication, May 02, 2018) support the view of trust as an important factor influencing the adoption process. We use the definition of Alalwan et al. (2017) who define consumer trust as the “belief of integrity, benevolence and ability that could enhance customer willingness to depend on a particular service” (Alalwam et al., 2017).

The second element resulting from the case study is market share and is defined by the current share the firm enjoys when launching the new technology service. Safaricom was able to experiment with M-Pesa to their large market share (P. Reinartz, personal communication, May 17, 2018).

(32)

32 The third element resulting from the case study is regulatory environment and is supported by Zhu et al. (2006) and Reinartz (2018) and Goldie-Scot (2018).

New model Complexity Trialability Relative advantage Compatibility Observability Performance expectancy Social influence Facilitating conditions Hedonic motivation Price Value Habit Experience Market share Regulatory environment Consumer trust

Country’s technology readiness

Table 6 - New framework technology adoption developing countries

Application to Blockchain for Remittances in Kenya

This chapter applies the most essential elements of the new theoretical framework, presented in the previous chapter, to blockchain expectations in order to predict the adoption for remittances in developing countries. The following research question is answered “How can the new complementary theoretical framework predict and explain the adoption of blockchain for remittances in developing countries?”.

This study applies the framework presented in the previous chapter to expectations on blockchain-based technologies. First, the definition of blockchain and possibilities for applying the technology in the remittance sector are given. Following, the newly developed theoretical framework is applied to predictions on blockchain for remittances in developing countries.

Blockchain-based technologies can change different sectors and industries worldwide. One way it can be used, is as a mobile money system for remittances (Kshetri, 2017; Scott, 2016; Swan, 2015; Peter & Panayi, 2016). It is transparent, safe and there is no need of a trusted third party because of its following four characteristics. Firstly, blockchain is a distributed

(33)

33 network, making it possible to use at any location, at any time. BitPesa co-founder Goldie-Scot identifies this as most important advantage of the blockchain technology (D. Goldie-Scot, personal communication May 30, 2018). Secondly, agreed transactions are stored in a database as smart contracts, making sure everyone plays by the agreed upon rules (IBM, 2017). Thirdly, agreement of validation between all parties involved is necessary, which minimises potential fraudulent transactions. Fourthly, its data is immutable, meaning it can never be changed once the transaction is completed. This helps to know the origins and history of every asset or transaction (IBM, 2017).

Questioning why the remittance market is dominated by a few commercial companies charging high prices, the blockchain technology can open up this sector and can provide advantages flowing from its characteristics. Therefore, it is particularly interesting to apply the new theory to predictions around blockchain for developing countries. Since blockchain is a new and disruptive technology which offers many solutions, it is interesting to properly investigate its possibilities for remittances in developing countries. Since remittances contribute to economic development in developing countries, blockchain could open up this sector and greatly benefit the customers (Chami, Fullenkamp & Jahjah, 2005; OECD, 2015). Consequently, this research contributes to the current literature in developing a more universal theoretical framework on diffusion of innovation and technology in developing countries, applicable to blockchain in these markets.

In the next paragraph, elements of the new model are explored in the application of it to blockchain technology developments for remittances. Blockchain is seen as the successor to M-Pesa for mobile payments and remittances (Swan, 2015).

The first element of the new model explaining the potential adoption of a new technology is complexity. Goldie-Scot argues this is not of significant relevance for developing countries and blockchain-based solutions: “People are ready in Kenya, an old lady said to me

(34)

34 full enthusiasm: when can we start using it?” (D. Goldie-Scot, personal communication May 30, 2018).

Another element of the framework, relative advantage appears to be important. The advantages of blockchain mentioned earlier, including the decentralised ledger and low (potential) transaction costs, indicate a good relative advantage compared to current solutions in the remittance sector (IBM, 2017, Biggs, 2016; Kewell et al., 2017; Kshetri, 2017). When the relative advantage is large, consumers are more willing to accept a higher complexity. This supports the view of Goldie-Scot (2018) stated above.

The next element of the new theoretical framework that appears to be important is the one of performance expectations regarding blockchain-based solutions for remittances (Swan, 2015; Biggs, 2016; Kewell et al., 2017; Kshetri, 2017). The recent set up of the Blockchain Task Force in Kenya could positively increase the seventh element of the new theoretical framework, social influence (Kachwanya, 2018).

Regarding the next element of the framework, the mobile usability is high in Kenya providing good facilitating conditions. In 2017, 72 per cent of the Kenyan population have registered mobile money accounts (Finclusion, 2018).

Another element of the framework is price value. Blockchain-based solutions for remittances are relatively cheap compared to the current alternative solutions (Swan, 2015). This could positively influence the adoption of blockchain-based solutions.

The thirteenth element of the framework is consumer trust. An advantage of blockchain is that its data is immutable, meaning it can never be changed once the transaction is completed. This helps to know the origins and history of every asset or transaction (IBM, 2017). This increases the trust of consumers in the service. Furthermore, Kenya is considered to be a more collectivistic society and therefore less likely to trust a person not part of their group (Hofstede insights, 2018; Jarvenpaa, Tractinsky & Saarinen, 1999). This is in line with the reasoning of

(35)

35 trust in the brand, as a success factor for M-Pesa. M-Pesa is considered as a homegrown service, leading to increased trust in the brand (Mas, 2010).

The fourteenth element of the framework is market share. Regarding this element it can be stated that each start-up offering a solution experiences a disadvantage. There is a lack of awareness of the service amongst potential customers because it is a start-up’s service. Goldie-Scot states that the main reason for BitPesa to switch from remittances to corporate payments, were the customer acquisition costs (D. Goldie-Scot, personal communication, 2018). Another element for BitPesa to move from a remittance focus to a corporate one, is the next element of the framework, the regulatory environment. Reinartz emphasises the importance of regulating bodies in developing countries (P. Reinartz, personal communication, May 17, 2018). BitPesa experiences difficulties launching in Kenya due to regulating bodies (P. Reinartz, personal communication, May 17, 2018; D. Goldie-Scot, personal communication, May 30, 2018). The power of regulating bodies in developing countries is more important than in developed countries. Therefore, this element is important to incorporate in a framework for technology adoption in developing countries.

The last element of the framework is a country’s technology readiness. Due to the experience with mobile money, Kenya’s infrastructure is suited to support blockchain-based solutions for remittances. The necessary infrastructure is thus available within this specific context. This might not be the case for other developing countries. However, besides the existence of the necessary infrastructure, accessibility to this infrastructure is relevant in explaining a country’s technology readiness. Availability of the crucial infrastructure does not automatically imply that the provider of this infrastructure allows access to it. Dependability on specific infrastructure (e.g. network providers) results in an essential relationship with providers of this infrastructure, and thus the potential competitors. Since Safaricom owns over 70 per cent

(36)

36 of the mobile network, allowance of using the network by a (potential) competitor is questionable.

Consequently, the goal of this chapter is to demonstrate the applicability of the newly constructed framework and answers the research question “What theoretical framework can best be used to predict and explain the adoption of a blockchain-based solution for remittances in developing countries?”. This new framework seems appropriate in its initial application. Taking only into account the traditional elements of the DOI and the UTAUT2, this study would suggest that blockchain initiatives could find its way to consumer adoption for remittances in developing countries. These two models show a good basis for explaining the adoption process of this blockchain-based solutions. These models suggest that, if the service or product shows potential according to the elements of these models, its adoption and acceptance is likely. However, taking into account the new elements, a successful adoption process is less likely to occur. It is questionable if entrepreneurs take these elements (enough) into account. Drawing on the earlier mentioned example of BitPesa, the start-up started with the intention to become a platform for remittances in Kenya. However, when experiencing too many difficulties with consumer acquisition and regulating bodies, such as the bank (consumer trust and regulatory environment), they shifted their focus to domestic and international corporate payments in the Kenyan market. When looking at the Nigerian market, BitPesa business is more successful due to close ties with the government and banks. (D. Goldie-Scott, personal communication, May 30, 2018). This shows the possibility of being successful in other countries, emphasising the importance of incorporating these elements.

The power of the regulatory environment and a country’s technology readiness is more important than start-ups might expect. It is clear that entrepreneurs and businesses consider the power of regulating bodies more in developing countries than in developed countries. However, as seen above, it can be questioned whether all initiatives consider the dependability on

(37)

37 infrastructures enough. If the technology uses or depends on a certain network owned by a (potential) concurrent, it is possible this owner never allows the service or business to use it. For example, Safaricom owns 70 per cent of the mobile network in Kenya. They are in power to prevent a business from using their network. The government and Safaricom are closely linked (P. Reinartz, personal communication, May 17, 2018).

Concluding, the new theoretical framework includes important elements that were not incorporated in the traditional DOI and UTAUT2, namely market share, regulatory environment, consumer trust and country’s technology readiness. Next to the traditional elements of DOI and UTAUT2, complexity, relative advantage, performance expectations and facilitating conditions, also these new elements from the comprehensive framework, market share, regulatory environment, consumer trust and country’s technology readiness are seen to be essential for the prediction of technology adoption in developing countries. As stated above, these elements are more important than in developed countries and it is therefore important to keep these in mind.

The results of this study indicate that start-ups focusing on blockchain based solutions might, for the time being, better focus on business to business (b2b) and corporate sectors instead of on end-consumers. This is stressed by importance of both the traditional and the new elements found in this study.

This gives hope to non-governmental organisations (NGOs) which are not involved with rules regarding regulatory environment. Blockchain for humanitarian causes is often mentioned as a potential application (Swan, 2015; Kewell et al., 2017; B. Ndemo, personal communication, June, 15, 2018). For example, the red cross blockchain initiative mentioned earlier. Due to the absent power of the regulatory environment otherwise, NGOs might achieve quicker consumer adoption.

(38)

38

Discussion, Limitations and Future Research

Blockchain and its possible applications is a rising topic in IS journals. Furthermore, along the global development, the realisation of the necessity of local adjustments regarding theories has become crucial (Foster & Heeks, 2013; Venkatesh, 2012; Kumar, van Dissel & Bielli, 1998). Therefore, this paper contributes to the existing IS literature by providing a comprehensive theoretical framework for blockchain-based solutions in developing countries. This framework combines learnings from M-Pesa, DOI and UTAUT2. DOI and UTAUT2 are proven to be powerful frameworks and when complemented with contextual elements, they can be used to understand the phenomena and consumer technology adoptions in developing countries (Venkatesh, 2012).

Despite the increasing attention to developing countries, there is still a lack in top IS journals of research on this part of the world. Also, there is still a lack in goodness of fit of western-based theories in their global application. This suggests the need of modifications and the need for new frameworks within the respective context.

The first limitation of this research concerns the generalisability of the theoretical framework, because it is bounded to technology adoption analysis in developing countries. This research focused on Kenya and therefore, for future research, it is suggested to explore this framework in other countries. A second limitation concerns the convenience- and snowball-sampling used in order to find interviewees, due to limited access to the market. Thirdly, this study only investigates one case study and one type of technology. Therefore again, for future research, it is suggested to explore the applicability of this framework to other technologies, like other mobile services.

Rethinking adoption theories for developing countries investigating the goodness of fit for M-Pesa and blockchain