ADOPTION INTENTION OF LOW-TECH B-2-B SERVICE INNOVATIONS IN DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

(1)

DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

Fasta Cycle Messengers’ low-tech b-2-b parcel delivery service in Dar es Salaam, Tanzania

(2)

ADOPTION INTENTION OF LOW-TECH B-2-B SERVICE INNOVATIONS IN

DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

Master thesis, MscBA, specialization Business Development University of Groningen, Faculty of Economics and Business

April, 25, 2013 ELMAR DE VREEDE Student number: 1634992 Westersingel 21-A 9718 CB Groningen tel.:+31(0)649858276 e-mail:elmar_dv@hotmail.com First supervisor F.D. Streefland Second supervisor C. Reezigt

Acknowledgement: Thanks to Frank Streefland for his overhead view, focus on structure and personal approach to supervising. Thanks to Bartjan Pennink for starting of my journey by getting me excited about doing research in Tanzania and for facilitating a lifes lesson that when you really want something you can find a way. Thanks to Reachel Kayeye and her ‘’foot in the door’’ approach in helping me to get appointments with the right people for my interviews. Thanks to Elaine Baker, Tendekay Guni and little Tara for their hospitality in Dar es Salaam. Thanks to Allen Andrew and Onesmo Williams for providing me with rich cultural experiences that have helped me to get feeling with the research context. Thanks to my personal bajaji (tuctuc) driver Mr. Begga for taking care of the logistical part of the research. Thanks to everyone else in Tanzania who took part in making my stay an unforgettable (learning) experience. Thanks for the helpful comments of Clemens Lutz, Cees Reezigt and Maira de Vreede.

(3)

1. INTRODUCTION

In a rapidly changing world, the importance and urgency for organizations to innovate increases (Rujirawanich et. al, 2011; Dewar and Dutton, 1986; Frambach and Schilleweart, 2002). Innovation can give supplying organizations the opportunity to attract new customers, retain existing customers, stay ahead of competition and strengthen the ties with their distribution network (Rujirawanich et.al, 2011; Chandra and Neelankavil, 2008). At an aggregated level innovation can improve the international competitiveness and economic success of countries (Porter et. al 2002; Goedhuys, 2007; Rujirawanich et.al, 2011). Statistics however point out that innovation is a risky business, because most innovations fail to be adopted, leaving considerable costs after introduction (Dewar and Dutton, 1986; Frambach and Schilleweart). Not merely innovation, but successful innovation seems to be the key to future growth and development of economies (Porter et. al 2002; Iakovlea et. al, 2011). Successful innovations are those that match with the needs of buyers and users and have an adoption rate high enough to ensure that the payback on investments occurs within a relatively short period of time (Frambach and Schilleweart, 2002; Rogers, 1995; Robertson and Patel 2007). This thesis and its accompanying research will zoom in on successful innovations in the context of developing countries. Adoption and diffusion literature in organizational science examines the factors and processes that underlie successful innovation, by focusing on the acceptance of innovations. Adoption can be defined as ‘’the decision of an individual or organization to make use of an innovation’’, whereas diffusion can be defined as ‘’the accumulated level of users of an innovation in a market’’ (Frambach and Schilleweart, 2002). Innovation in this sense refers to ‘’any idea, product, service, program or technology that is perceived to be new by the relevant unit of adoption’’ (Hameed et. al 2012). For innovating organizations in developing countries a bigger understanding of innovation adoption can help to identify the needs of potential adopters and innovations can be developed and positioned accordingly (Frambach and Schilleweart, 2002). This benefit could potentially facilitate and increase the success of innovations, which in turn is vital in order to make progress in the process of economic development towards catching up with the rest of the world (Musa et. al, 2005; Musa, 2006). From existing literature the argument can be derived that certain types of innovation are likely to be more successful in the context of a developing country and thus can contribute more to the development of developing economies compared to other types of innovations (Molla and Licker, 2005). In order for developing countries to be able to make the shift towards dynamic high-tech developing sectors first the right conditions have to be shaped to provide more viable entry points into high-tech industries (Robertson and Patel, 2007; Tunzelman and Acha, 2005). Evidence from developed countries tells us that the bulk of the customers of high-tech offerings are organizations in low-tech sectors that adopt and adapt those offerings as input for innovations (Robertson and Patel, 2007). For developing countries the adoption and adaptation still seems to fail, because of a lack of resources, knowledge, welfare and infrastructure (Musa, 2006). Increasingly healthy low-tech sectors could help to accumulate these factors in order to set the right conditions under which high-tech sectors become viable (Aubert, 2005; Musa, 2006). Once a certain level of wealth is reached it becomes optimal for an economy to import technological capital and begin to produce new technologies (Bruno et. al, 2008). The low-tech sectors in developing countries

(6)

- 2 - thus seem to play a significant initial role in the development process of developing economies (Robertson and Patel, 2007; Kirner et. al, 2008; Santemaria et. al, 2008).

For developing countries it is proposed that input and output relations between organizations are necessary for innovation and innovation adoption, which arguably makes the business to business (b-2-b) type of innovation adoption especially relevant to economic development (Robertson and Patel, 2007). Service innovations have also proven their relevance to the context of developing economies, because service sectors provide an increasing share of employment and an increase in GDP to developing economies (Seyoum, 2007). Services can also help goods producers become more efficient. Sectors such as transportation or financial services, for example, set the conditions under which goods, labor and capital can flow (Seyoum, 2007).

Despite of the potential relevance of low-tech b-2-b service innovations in the context of developing countries the majority and the most cited and redeveloped mature models seem to focus on acceptance of high-tech products that are offered to individual adopters in a developed country (Venkatesh et. al, 2003; Venkatesh et. al 2012). Extensions to these main models to other types of innovations and other contexts are scarce and have only recently started emerging in innovation adoption literature (Venkatesh et. al, 2012). From these scarce studies it is evident that the existing models and theories cannot be generalized across innovation type and context (Baaren et. al 2011). Studying adoption of different types of innovation in different contexts has proved existing theorized relationships insignificant, altered the direction or magnitude of relationships, and has created new relations altogether (Venkatesh, 2012). It appears that low-tech innovations, b-2-b innovations, service innovations and the context of developing countries make different factors relevant in the adoption decision. Combining these aspects in one situation is expected to be deviate even more from the existing literature on innovation adoption. The combination of these aspects thus seem to represent a gap in existing literature that deserves further exploration. The following research question attempts to provoke the additional knowledge in order to close the gap:

What factors determine the intention to adopt a low-tech b-2-b service innovation in a developing country?

The next paragraph will further elaborate, explain and justify this research question. In the second paragraph existing theory will be assessed in order to see what is already known about the research and what is still to be discovered. The research methods used will be described in the third paragraph. The results of this research will be laid out in the fourth paragraph, followed by the final paragraph that will conclude and discuss the findings.

(7)

- 3 -

2. THEORY

2.1 Defining a developing country

Countries are often roughly placed in three categories according to their level of economic development (Hoskisson et al, 2000). Most countries in Asia, Latin America, and the Middle East are developing countries, China and Russia transition countries and western economies of the US and Europe are developed or industrialized countries (Hoskisson et al, 2000). The aspects that many of the developing countries have in common is that until the 1980’s international policy favored state planning and state ownership to lever economic development (Parker and Kirkpatrick, 2005). Because of ever declining economic trends in many of these countries reforms were introduced towards more privatization and a reduction of state participation in industry. These reforms brought partial success to developing countries, but did not achieve the expected transformation (Parker and Kirkpatrick, 2005). In recent years both policymakers and academics have started to look at innovation in relation to economic success and there is a growing understanding of both policymakers and academics that innovation processes underlie the international competitiveness of organizations and states (Parker and Kirkpatrick, 2005). As countries develop this usually goes hand in hand with an expansion of the service sector, R&D activities and knowledge (Iakovleva et. al, 2011). Economic development can be described as a sequential process that involves successive upgrading of organizations and their supporting environments to produce and compete in increasingly sophisticated ways (Porter et. al, 2002). From literature it appears that that certain types of innovation seem especially relevant in their contribution to economic development of developing countries. These are low-tech, b-2-b and service innovation.

2.2 Low-tech innovation adoption in developing countries

The first type of innovation that seems relevant to developing economies is low-tech innovation. The distinction between high-tech and low-tech in literature is often done on the level of industry (Santamaria et. al, 2009). High-tech industries are those that undertake relatively many R&D activities. Consequently there often is a high degree of new scientific and technical knowledge embedded in innovations developed in high-tech industries (Kirner et.al, 2009; Santamaria et.al, 2009). Innovations developed in low-tech industries do not rely so much on R&D or the latest scientific and technological knowledge. Low-tech innovations usually result from an incremental innovation strategy, which uses internal experiments with technologies that are acquired externally and then adapted to function under the new conditions (Santamaria et. al, 2009; Hirsch and Kreinsen, 2008). High-tech innovations are usually developed by following a more radical innovation strategy and are characterized by transcending existing technological concepts (Hirsch Kreinsen, 2008). The level of high-tech or low-tech innovations in a country is tightly linked to the level of economic development. It appears that before a developing economy can progress to the next stage in development towards more high-tech innovation and catching up with the rest of the world, a certain level of development of the low-tech sectors needs to be reached (Perez and Soete, 1988). In order for a sector to develop, previous capital is needed to produce new capital, previous knowledge is needed to absorb new knowledge and skills must available to acquire new skills (Perez and Soete 1988). In the first stage of economic development organizations mostly market commodities and relative simple offerings of long-standardized technologies. When enough capital knowledge and skills are accumulated a progression to the next stage of economic development can be established. In this second step existing technology is not only assimilated, but also improved up on. In the third

(8)

- 4 - step of economic development an economy includes significant investments in R&D, higher education and improved capital markets and regulatory systems, that all support organizations in producing offerings with new technology. The third stage of economic development is usually where developed countries are in (Perez and Soete, 1988). Lack of improvement in any economic area can lead to a plateau in productivity and stalled economic growth (Porter et. al, 2002). The sufficient accumulation of resources, knowledge, welfare and infrastructure in the early stages of economic development in low-tech sectors can facilitate the transition towards an economic focus on the development of new technologies (Bruno et. al, 2008; Aubert, 2005 Hirsch-Kreinsen, 2008). This implies that developing countries need to follow the sequential process of economic development (Perez and Soete, 1988; Porter et.al, 2002) and start development in all low-tech sectors of the economy (Heidenreich, 2008).

2.3 B-2-B innovation adoption in developing countries

The second type of innovation that seems to be important for developing countries is innovation offered to organizational adopters, so-called b-2-b innovations. Two main groups of innovation adopters can be identified from literature, which are b-2-b adopters or individual consumers (Frambach and Schilleweart, 2002). B-2-b adoption consists of two sequential parts; the first is the organization that adopts an innovation and the second part is the consequent adoption of the innovation by employees in the organization (Frambach and Schilleweart, 2002; Vowles et. al 2011). In case the unit of adoption is another organization, than innovations are to be viewed as new production inputs, machines, processes, and techniques adopted by organizations for their own use (Frambach, 1993). In this case the adoption is from business to business, which means that the adopting unit is a team, department or a complete organization. As explained in section 2.2; in developing countries the way to economic development is through a focus on low-tech industries (Porter et.al, 2002). In such industries the innovation expenditures are usually relatively low, which puts limits on the ways in which innovations can be developed in organizations. B-2-b relationships can help to overcome the limitation of a firms own resources and knowhow in developing new production and innovation potential (Heidenreich, 2008). Evidence from developed countries tells us that organizations in low-tech industries are not only generators of low-tech innovations, but are also key users of high-tech innovations (Santamaria, 2009; Robertson and Patel, 2007). Organizations in high-tech sectors play a big role for organizations in low-tech sectors as source of information and as providers of technology for innovations (Heidenreich, 2008). Low-tech organizations are often described as technology adopters and adapters whereas high-tech organizations are often described as technology developers and producers (Robertson and Patel, 2007). For development and innovation to take place in low-tech sectors it is often the case that long standardized offerings from high-tech sectors are adopted and adapted into a low-tech innovation. So for the organizations in low-tech sectors of economies to stay up to date they are often dependent on the input of technologies from organizations in high-tech sectors (Robertson and Patel, 2007). B-2-b innovation adoption thus seems a key for economic performance. The importance of the reciprocal input and output relations between different sectors in an economy, makes b-2-b innovation arguably an important road for developing countries towards catching up with developed countries (Hauknes and Knell, 2009; Robertson and Patel, 2007). It appears that first the low-tech sectors need to be developed in developing countries by adopting and adapting existing technologies. When sufficient accumulation of resources has taken place in low-tech sectors the high-tech sectors become more feasible, because there are more low-tech clients with enough resources and knowledge to adopt

(9)

- 5 - and adapt high-tech innovations or start developing high-tech innovations themselves (Porter et.al, 2002).

2.4 Service innovation adoption in developing countries

The third type of innovation that seems relevant to developing countries is service innovation. Services encompass a heterogeneous group of economic activities often having little in common other than that their principal outputs are largely intangible (Seyoum, 2007). It includes activities such as transport of goods and people, financial intermediation, communications, distribution, hotels and restaurants, education, health care, construction and accounting (Hoekman, 1999). In developing countries services appear to account for an increasing share of employment and GDP (Seyoum, 2007). Services are essential for overall economic development, because they are often inputs for the production of other goods and services. Producers thus depend on services to deliver their output to end users. Services can also help goods producers become more efficient. Sectors such as transportation or financial services, for example, set the conditions under which goods, labor and capital can flow (Seyoum, 2007). Although developed countries dominate global trade in services from an absolute perspective, developing countries are often highly dependent on service exports as a source of foreign exchange. The trend in developing countries in the past decades has been significant deregulation and privatization of state owned organizations. This trend has provided a bigger market access for service offerings (Seyoum, 2007) and has potential to grow in importance since service industries such as telecommunications, airlines, railway, water and electricity are often still under government monopolies in developing countries (Musa, 2005). The quality of these state owned services is often below standards which keeps the economies from developing further, because the price and quality of the services that are available in an economy have major impacts on all sectors and overall economic performance (Musa, 2005; Hoekman, 1999).

2.5 Innovation adoption models from theory

Although the low-tech, b-2-b and service innovations seem relevant in developing countries the main models from innovation adoption theory appears to have a different focus. The adoption and diffusion of innovations is a theme that originated from social sciences, which since then has been used in various disciplines (Smith et, al 2010). Gradually the acceptance literature spilled into the marketing domain of organizational sciences (Rogers, 1995; Venkatesh et. al, 2003). The majority of the innovation adoption models that were constructed over the past decades are based on the Theory of Planned Behavior (TPB), which came from psychology research on the Theory of Reasoned Action (TRA). TPB shaped three perceptions: attitude, subjective norm, and perceived behavioral control. TPB was first used to explain adoption of innovations in general. This model was later extended in several ways to explain specific behaviors that relate to high-tech innovations. The Technology Acceptance Model (TAM) is perhaps the most widely used extension and has been the basis for most research into technology adoption. TAM consists of the constructs perceived usefulness and perceived ease of use. In TAM2 this was extended with the subjective norm adapted from TRA/TPB. In general, these theories (TPB, TRA, TAM) imply that behavior is determined by the intention to perform the behavior. Intention to perform a specific behavior or adopt a technology and the actual performance or adoption were found to be highly linked (Yi et.al, 2006;Sun et. al, 2007;Sukkar and Hasan, 2005;Ozdemir and Trott 2009). Another dominant theory of innovation adoption, much used to explain adoption of innovations is the Diffusion of Innovation Theory (DOI). This model posits that the rate of adoption of an innovation is determined by several attributes of

(10)

- 6 - the innovation, which are relative advantage, compatibility, complexity, trailability and observability (Rogers, 1995; Rogers, 1983). This model was developed and used primarily for high-tech innovations (Strutton et. al, 1994). A more recent innovation adoption model is the Unified Theory of Acceptance and Use of Technology (UTAUT) model that has integrated eight different high-tech innovation adoption models (including TPB, TRA, TAM and DOI) and its components into a single model. This model until now has proven to be the most predictive for high-tech products with individual adopters in developed countries (Venkatesch et. al, 2003; Venkatesch et. al 2012).

2.6 Generalizability of models across different types of innovations

One of the first steps in establishing the efficacy of a theoretical model is examining its generalizability across time, populations, and contexts (Venkatesh et.al, 2007). Innovation adoption seems to be determined different across different types of innovations. This is carefully worded by Damanpour (1988): ‘’all types of innovations do not have identical attributes, their process of adoption is not the same, and they do not relate equally to the same predictor variables’’ p-547-548. Although extensions of existing innovation adoption models to other types of innovations and contexts have been undertaken, Venkatesh et. al (2007) states that there is still much promise for generation of new knowledge by focusing research on other types of innovation. The generalizability of existing knowledge to the low-tech, b-2-b and service types of innovation in specific seems questionable (Dwiwedi et. al, 2008).

2.6.1 Generalizability from high-tech to low-tech

A first limit on the generalizability of the main innovation adoption models depends on the degree of new technology embedded in the innovation. The main innovation adoption models seem to focus predominantly on high-tech innovations. High-tech innovation adoption appears to have different issues that are important to the adoption decision compared to the adoption of low-tech innovation (McDade et. al, 2002). High-tech innovations will in general be less familiar to potential adopters, which can heighten the perceived risk associated with adoption. For potential adopter of high-tech innovation a number of factors that can reduce novelty and increase familiarity with innovations like information from peers, media, opinions of experts or observing the outcomes of other actors, therefore seem more important (Weijnert, 2002). High-tech innovations compared to low-tech innovations also appear to have more issues regarding compatibility with other products that are currently owned or the availability of complimentary products, which help to make the product more usefull for the organization or individual (McDade et. al, 2002; Gao et. al, 2012). When the adopter is an organization the organization size also seems to be of influence. For small organizations high-tech innovations are perceived more risky because of the posession of little resources by the organization and the greater cost and complexity related to high-tech innovations. For large organizations the decision process for high-tech innovations is more complex because of the need for the involvement of more people in the decision, which can slow it down and resist change. With low-tech innovations, in general, less people need to be involved in the decision (McDade et. al, 2002). Adoption of high-tech innovations is dependent on the ability of an adopter to make a judgment, for which knowledge resources are needed (Heidenreich, 2008). And the depth of knowledge resources in an organization seems to be more important for high-tech innovation adoption than for low-tech innovation adoption (Dewar and Dutton, 1996). Low-tech innovations are often based on long standardized technologies and low-tech innovation developers are therefore less able to be different from competitors with their offering. Marketing, design and service innovations seem to be an alternative

(11)

- 7 - road for organizations that offer low-tech innovations to differentiate themselves. These factors therefore appear especially important in the judgment of low-tech innovations (McDade et. al, 2002).

2.6.2 Generalizability of individual adopter to b-2-b adopter

A second limit to the generalizability of the main innovation adoption models comes from the predominant focus on individual adopters. From theory it appears that individual adoption decisions differs from those made by decision making units in organizations. The adoption decision in this situation is often a multi-phase, multi-person, multi-department and multi-objective purchasing process, and will thus have different determinants and levels of determinants compared to looking at individual end-users (Vowles et. al, 2011). This multi-nature of b-2-b adoption decision makes these decisions in general more complex than individual decisions. B-2-b adoption normally takes place over a longer period of time, and is influenced by a greater number of forces inside and outside the buyer firm (Vowles et.al, 2011; Frambach and Schilleweart, 2002; Smith et. al, 2010). Vowles et.al (2011) explain that the ability to obtain information and form good judgment with that information presents a higher challenge for resource-constrained organizations in particular (Vowles et.al, 2011). Individual adoption decisions are affected by personal characteristics and individual needs, while organizational decisions are affected by organization characteristics and attitudes and expectations about the degree to which the innovation can support the adopting organization in carrying out value-adding activities (Frambach and Schilleweart, 2002; Frambach et. al, 1988). Because most of the time more people are involved in the initial organizational decision it usually follows a more formal procedure (Frambach and Schillewaert, 1998). Organizations are focused on achieving and sustaining competitive advantages and therefore innovation adoption usually appears to involve a longer term commitment with a higher degree of perceived risk involved than in the case of consumer products (Frambach et. al 1998; Rogers, 1995).

2.6.3 Generalizability of product to service

A third limit to the generalizability of the main innovation adoption models seems to be the predominant focus on products. The different nature of services compared to products appears to make the judgment and therefore innovation adoption different (Malhotra et. al 1994). Making a adoption decision can be seen as largely an evaluation process (Rogers, 1995). In general more difficulties seem to arise in the evaluation of services before adoption (Black et. al, 2001). Services are largely intangible, so there is no tangible item which can be viewed, held or tested to some degree before adoption. Production and consumption of services can often not be separated. This means that before consumption it is not produced so there is little to judge the service on (Den Hartog et. al, 2009; Malhotra et. al, 1994). Attention in evaluation can shift to other more tangible aspects of the service, or the prestige or reputation of the supplier (Frambach et. al, 1998).

In general for services the adopters will have higher levels of uncertainty and risk, because of the mentioned difficulties in evaluation (Black et. al, 2001).

2.7 Generalizability of developed countries to developing countries

Innovation adoption is not only determined differently for different types of innovations, also the context in which innovation adoption takes place seems to influence innovation adoption. Current innovation adoption models seem to have a predominant focus on developed countries, largely ignoring the context of developing countries (Da Silviera, 2001). Innovation adoption research conducted in developing countries points out that existing models have been found to be less

(12)

- 8 - predictive (Bandyopadhyay and Fraccastoro, 2007). Cultural and economic differences seem to be responsible for the differences in innovation adoption across countries (Chandrasekaran and and Tellis, 2008). The culture and economy of a country seem to affect the practicality and benefits of adoption, as well as an adopter’s willingness and ability to adopt an innovation (Wejnert, 2002).

2.7.1 Generalizability of economic conditions

Economic conditions appears to be the first of two broad drivers from the context that influence innovation adoption (Chandrasekaran and Tellis, 2008). Economics can be thought of as differences in opportunities and wealth that limit consumers’ ability to purchase innovations (Chandrasekaran and Tellis, 2008). Economic rationale suggests that consumers who adopt an innovation need to have the ability to pay, the willingness to pay and access to the product (Talukdar et. al, 2002). Economic conditions thus seem to shape the susceptibility of potential adopters to adopt innovations (Wejnert, 2002).

The conditions under which innovation and innovation adoption takes place in developing countries is embedded in several kinds of economic scarcities that are not as widely present in developed countries. Developing countries often face problems of infrastructure that is missing or not up to date. Telecommunications and internet infrastructure remains weak and limited, transport infrastructure as well as sanitation; water and other infrastructural systems also remain sub-standard (Musa et.al, 2005; Srinivas and Sutz, 2008). This lack of such complementary systems can decrease the willingness to pay for certain innovations that rely on these systems (Talukdar et. al, 2002). The masses in developing countries yearn for basic infrastructure for socioeconomic development from which access and exposure to modern technologies could follow (Musa, 2006). Lack of educated human resources often lowers the value certain offerings because potential adopters often have not passed through the requisite learning curve in developing countries (Musa, 2006). Some types of knowledge can only be build up by experience and cannot be transferred to a receptive individual who does not have the expertise. This means that organizations in developing countries should build upon resources and capabilities available at their level of development (Aubert, 2005). There is also a lack of availability and access to materials and equipment of the required quality or accuracy and a lack of people with appropriate education and skills to come up with ideas and run innovation projects (Srinivas and Suts, 2008), which limits the accessibility. A general lack of income and capital means that well-known solutions to problems often cannot be funded both on the supply as on the demand side (Musa, 2006; Shih et. al, 2008). In many developing countries governments have tried to attract investors and innovations from developed countries in order to catch up towards this technology frontier (Musa, 2006). It was expected that this would bring spill-overs of knowledge and technology to local firms, however this was not achieved (Musa, 2006). The imported innovations showed very low adoption rates. With even huge evidence of wasted resources in several of the least developed countries (Musa, 2006).

2.7.2 Generalizability between cultures

Cross-country variation in adoption is not only explained by economic conditions, but also by the prevailing culture (Erumban and de Jong, 2006; Waarts and van Everdingen, 2005; Vatanasakdakul et. al, 2004;Chandrasekaran and Tellis, 2008). A broad spectrum of factors of culture is studied in adoption and diffusion research. Things like belief systems (values, norms, language, religion, and ideologies), cultural traditions, cultural homogeneity and socialization have been found to influence

(13)

- 9 - innovation adoption (Wejnert, 2002). Organizations are found to be less likely to adopt innovations that conflict with societal culture, because such incongruence increases the risk of societal disapproval. On the other hand, when an innovation is consistent with local cultural traditions and belief system, these variables are strong determinants of the ceiling of adoption, which is the number of actual adoptions to the number of potential adopters (Wejnert, 2002).The significant influence of cultural differences on innovation adoption has often been proven with help of the framework of Hofstede (2001). Hofstede defines national culture as the collective programming of the mind that distinguishes the members of one group or category of people from another. His culture theory can be understood in terms of five dimensions: power distance, uncertainty avoidance, individualism/collectivism, masculinity/femininity and long term/short term orientation. See Table 1 for a more detailed description of these dimensions and how they relate to innovation adoption in organizations.

TABLE 1 Hofstede’s national culture dimensions and their effect on innovation adoption

Factor Description

1. Individualism/collectivism Individualism describes the relation between the group and the individual. Individualist cultures are characterized by a loosely knit social framework in which individuals take care of themselves and their immediate family, whereas collectivistic societies are integrated into strong cohesive groups (Gong et.al, 2006). In collectivistic countries people will act conform the norms of the group, which may lead to a delay in the adoption decision process. In contrast in individualistic countries people will make their own choices (Waarts and Everdingen, 2005) A positive relationship has been found between individualism and innovativeness (Gong et al, 2006). In individualistic cultures innovation adoption is found to be

higher than in collectivistic cultures.

2. Power distance The extent to which members of a society accept that power in institutions and organizations is distributed unequally (Gong et.al, 2006). In organizations this distribution of power is reflected in centralization of decision making; authority; the use of formal rules and the sharing of information is constrained by hierarchy (Waarts and Everdingen, 2005). High levels of power distance have been found to be associated with lower rates of innovation adoption. This is because there is less openness to new ideas and idea sharing (Gong et. al, 2007). Cultures with high power distance will

negatively influence innovation adoption (Gong et. al, 2007) .

3. Masculinity/Femininity Masculinity pertains to societies in which emphasis is on ambition, competition, material values. Organizations in masculine cultures focus on rewards and recognition of performance, and training and improvement of the individual, both characteristics that are common to innovative organizations. Femininity pertains to cultures in which values like equality, solidarity, social relationships and managers’ use of intuition and seeking consensus are important (Waarts and

(14)

- 10 - Everdingen, 2005). The findings about masculine/feminine cultures

and its influence on innovation adoption are mixed (Gong et. al,

2006).

4. Uncertainty avoidance The extent to which members of a society can tolerate uncertainty and ambiguity (Gong et, al, 2007). For organizations uncertainty avoidance usually means resistance to innovations, highly formalized management and the constraining of innovations by rules. In high uncertainty avoidance cultures, risk-averse attitudes imply that organizations will not take unnecessary risks and only adopts innovations if its value has already been proven in the market (Waarts and Everdingen, 2005). Cultures high on uncertainty

avoidance have been found to have a negative influence on innovation adoption.

5. Long term/ Short term orientation

Measure of people’s consideration of the future (Gong et. al, 2006). Values associated with long term orientation are persistence, adaptations of traditions to new circumstances, personal adaptability, and the idea that most important events in life will occur in the future (Gong et. al, 2006). Organizations in cultures with a long-term orientation focus on future results, and are more receptive to in-depth investments to effectuate a long-term change in the firm than organizations operating in a short-term orientation culture with a focus on the past and quick wins (Waarts and Everdingen, 2005). Cultures with a long-term orientation have a

significant positive influence on innovation adoption.

Using the dimensions of Hofstede, Medonca and Kanungo (1996) and Anandarajan et. al (2002) posit that the cultural environment of developing countries, when compared to developed countries, is more collectivistic, higher on power distance, more feminine and more uncertainty avoidance and have a shorter term orientation. An additional dimension ‘’abstractive versus associative thinking’’ has been suggested by Kedia and Bhagat (1988) to be useful in understanding the cultural differences between the developed and developing countries. According to Kedia and Bhagat (1988), in associative culture, people utilize associations among events that may not have much logical basis, whereas in abstractive cultures, cause-effect relations or rational types of thinking are dominant. Developing countries are more associative and determine their behavior on the context and developed countries are more abstractive in their thinking (Anandarajan et. al, 2002). A word of caution is needed when generalizing the cultures of developed and developing countries and the corresponding effects on innovation adoption. Culture is often state to be country specific and therefore between developing countries or differ between developed countries (Waarts and van Everdingen, 2005; Tan et.al, 2007).

2.8 Combining low-tech, b-2-b, service and developing country context

As was outlined in the previous sections, adoption of the low-tech b-2-b service innovation type seemed especially relevant for the context of developing countries. The available knowledge on innovation adoption however seems to have a different focus as it concentrates mainly on high-tech product innovations with individual adopters in developed countries (Venkatesch et. al, 2003). From

(15)

- 11 - theory it appears that taken separate the individual aspects high-tech, individual adoption, tangible product and developed country are likely to have limited generalizability to the situation of low-tech, b-2-b and service innovation adoption. A combination of these aspects in one situation is therefore expected to differ even more, given the possibility that the aspects influence each other and therefore altering the determinants of innovation adoption even more. The aim of the research is to explore this specific situation in which these factors are combined. For this exploration the following research question will be used:

What factors determine the intention to adopt a low-tech b-2-b service innovation in a developing country?

Because of a lack of knowledge about the combination of these aspects in one situation no propositions can be made at the outset of the research of this thesis. The lack of knowledge puts the research in an explorative stage, which calls for certain research methods that can help to close the gap in literature effectively.

(16)

- 12 -

3. METHODS

3.1 Theory development based on a case study

Most previous innovation adoption studies appear to have followed a quantitative research approach. This is a logical approach when testable propositions can effectively be derived from the extant base of literature. For the current research there appears to be a gap in literature that seems too wide to be able to develop such testable propositions. This puts the research in an earlier or exploratory stage (van Engelen and van der Zwaan, 1994). The goal of this research is therefore to explore innovation adoption of low-tech b-2-b service innovation in a developing country empirically, with the end product being theoretical propositions. At the outset of the research a broad range of factors can be expected to take part in determining innovation adoption. A broad and flexible method thus seems necessary in order to be able to capture the whole range of possible determinants and to prevent early exclusion of factors. A quantitative method seems inappropriate in this situation since such a method typically relies on the quantitative measurement of pre-specified factors. The need for exploration and flexibility calls for a qualitative method (Flick, 2006). This was also suggested in the innovation adoption study of Baaren et. al (2011) that states that: ‘’It is important to look at contemporary situations to define influential factors, rather than examine only those factors that have been tested before. Methods of more qualitative nature, such as face-to-face interviews can help define such factors and further explain their nature’’ p-87. Qualitative methods allow for the forming of definitions and can help explain why certain relations might exist. Quantitative methods can only establish whether relationships exist using existing definitions (Van Aaken et al, 2006; Rogers, 1995). Innovation adoption is a phenomenon that involves perceptions and therefore additional insights regarding irrational causes of why potential adopters adopt can only be explained with a qualitative method (Rogers, 1995). The case study method of Eisenhardt (1989) was selected as an appropriate approach to follow for the current study. Case study is a research strategy that focuses on understanding the dynamics present in single settings. The underlying logic of this method is that a process is followed that starts out in a broad way and finds more and more focus by using findings from earlier setting as increased understanding and input when entering later settings. The case study approach provides the needed flexibility to allow intermediate shifts in focus when progressive insights provide cues to do so (Eisenhardt, 1989).

3.2 Case selection

An organization called Fasta Cycle Messengers in Dar es Salaam, Tanzania was selected as it appeared to be an appropriate research setting for conducting the case study. Fasta Cycle Messengers is a local courier service supplier, which is a start-up enterprise. The service Fasta Cycle Messengers offers can be typified as a typical low-tech b-2-b service innovation. The offering consists of the delivery of outgoing parcels and letters from organizations within Dar es Salaam to other locations within the city. This service offering is the first and only courier service of its kind in Tanzania that uses bicycles or tricycles as mode of transport for the deliveries. Looking back at the definition of innovation proposed by Hameed et. al (2012), ‘’any idea, product, service, program or technology that is perceived to be new by the relevant unit of adoption’’, this offering can thus be described as an innovation, since it is new to the country and to the potential customers. Other organizations in the same industry all use motorized vehicles to make deliveries. The development of this service did not involve much research activity or knowledge intensity. It was simply adapting an existing technology, the bicycle, to make it function in a new environment. The innovativeness only

(17)

- 13 - comes from a new field in which the technology is applied, which is transport of outgoing parcels and letters as a service that is offered. The service can therefore be typified as low-tech. The fact that Fasta Cycle Messengers offers this service to other organizations as their paying customer means that this service can be labeled as a b-2-b service. Tanzania, the country in which this service is offered, can be typified as a typical developing country because of three reasons. First is that Tanzania’s economy is still heavily based on agriculture and is not much industrialized. Second, the country has been going through a series of typical reforms from a central and state planned economy towards a more liberal market-based economy. And a third typical characteristic is that there is a low level of resources and technological capabilities available among organizations in Tanzania (Goedhuys, 2007).

3.3 Process of data collection

Before entering the field the interview instrument has been developed, after which respondents were selected and the actual data collection could proceed.

3.3.1 Instrument and protocol

A special type of semi-standardized interview technique was selected, called the expert interview. Flick (2006) states that such interviews are appropriate for studying subjective viewpoints. This type of interview focuses more on the group that the interviewee represents instead of only on the personal views of the individual as would be the case in a normal semi-standardized interview (Flick, 2006). This was done because in this research the unit of analysis is the relevant decision making group of the organization, which is represented by an individual member of this group. The interviews were preceded by clarifying the reasons for the research and a promise that no organization sensitive information will be elicited and that the results will be processed anonymously, followed by asking for permission to use a tape recorder. After these issues of confidentiality were dealt with the interviewees were provided with information, which was needed to lead the interview in the right direction in order to avoid unproductive topics. Such directive actions are typical in expert interviews (Flick, 2006). Direction in this case was provided by means of a short mental simulation. This means that the interviewer gave information about the service, which allows a potential adopter to make judgments without having actually used it. This is a method proposed by Castano et. al (2008). Information about the service that Fasta Cycle Messengers offers was given in both a letter of recommendation that the respondents read before the interview (see Appendix 2) as well as by the researcher who provided an introduction to each interview (see Appendix 3). The provided information was kept as objective as possible and concisely explained what the service innovation entails. This was done in order to restrict the influence of this information on the respondents’ judgments of the service. After this information was provided the interviewee was asked to put him or herself in the imaginative situation in which the relevant decision making unit in his or her organization is considering to start to use or not to use the service innovation that was just described.

At the start of all interviews some qualifying questions about the respondent and the organization were asked to assure that this respondent indeed was in a position to provide a valuable contribution to the research. The main question that each interview focused on was an open ended question, which is typical in a semi-standardized interview to being with (Flick, 2006). The open question that was used (or a similar version) was: What things would you and your colleagues take into

(18)

- 14 -

consideration or evaluate before having actually used this service? After each topic that the potential

customer mentioned, the interviewer asked clarification and deepening questions like the following: What makes x important in this decision? Or: Can you explain x a little bit more? Or: You mentioned x, how would you define x? Or: Can you give an example of x? When the interviewee would ran out of other topics as an answer on the main question, the interviewer would proceed by asking about factors taken from adoption literature to see if the potential customer would also find these specific topics relevant in their adoption decision. A list of the constructs from literature was crossed of when mentioned during the interview and the topics that where not addressed yet were treated at this point in the interview. The type of question that was used here was: Could y also be something you would take into consideration in your decision? The topics that were perceived as very strange or totally not relevant to a respondent have been left out after hearing this in three interviews. An example of such a topic is if the importance of the degree of fun in using the service, in their decision, taken from Venkatesh et. al (2012).This proved quickly to be an irrelevant construct. The use of progressing insights allowed constructs mentioned in earlier interviews could be added to the list of constructs as input for later interviews. The type of question that was asked here was: Another organization also mentioned z, could z also be something you would take into consideration in your decision. The interviews were closed by asking whether the interviewee would like to add something to the things said. In general the later interviews allowed a more focused approach. When a factor was brought up that was mentioned a lot of the times in earlier interviews sometimes a confirming question was used to check if the respondent indeed was talking about that factor.

3.2.2 Selection of respondents

In order to achieve the needed scope, depth and consistency of data results multiple respondents have been selected. The respondents in this research were all potential customers of the courier service offered by Fasta Cycle Messengers. A potential customer of their service has been defined as: any organization in the city of Dar es Salaam that regularly has parcels or letters that need to be delivered to another location within the city. Because the aim of this research is to find out what constitutes intention to adopt instead of actual adoption, it did not involve actual customers. Methodologically and practically this has advantages and disadvantages that will be discussed in section 5.2. The selection of respondents for this research was conducted by means of convenience sampling, which refers to the selection of those cases that where the easiest to access under the given conditions (Flick, 2006). Appointments could only be made face-to-face, the relevant people in organizations proved not to be reachable through phone or E-mail. Respondents have been selected by means of office visits. Those people that were willing to cooperate have been included. The organizations that have been included in the research sample have been roughly selected on three criteria, which were possible to judge based on the appearance of the office. A mix of government organizations, businesses and non-governmental organizations and small and large organizations was included, because of the expectancy that the type of adopter might affect innovation adoption in b-2-b adoption situations (Frambach and Schillewaert, 2002). Only Tanzanian organizations were included in the sample, because of the expectancy that cultural and economic factors might play a role in innovation adoption (Wejnert, 2002; Musa, 2006). This rough selection was done to increase generalizability of the results. A formal recommendation letter from the supervisor from the University was constructed and used in order to get permission to speak to the relevant people and to convince people of the researcher’s good intentions (see Appendix 2). In this letter from the University the goal of the research, the content of the requested appointment, and the service of

(19)

- 15 - Fasta Cycle Messengers were explained. This proved a valuable tool to increase willingness to get an appointment with a relevant person. Also the use of a mediator that speaks Swahili proved a good way to establish contact more easily with the receptionist employees of organizations, which in turn helped to get in contact with a relevant person for this research. The high level mastery of the spoken English language of the relevant respondents allowed for effective communication and an elimination of the need for translation. Although B-2-B adoption usually involves multiple people or decision units a decision was made to only include one member of the relevant decision unit. Given the limited resources it was not expected feasible to get appointments with complete decision units. This method of including one person from a decision unit is a widely used method (Wilson and Lilien 1992; Frambach et. al 1998) and has been proven to be valid. “As long as an informant is reasonably knowledgeable about the buying process it matters little who is chosen as the informant’’ p-166 (Frambach et. al 1998). The individual that was asked for in this research was the person who would normally be responsible for making a decision about starting to use a courier service. The relatively high rank in organizational hierarchy of the respondents was another indicator of the interviewees’ capabilities to give sound insights in what would constitute the adoption intention of the relevant decision making unit of their organization.

3.3.3 Overall process of data collection

Two test interviews at two different organizations in Dar es Salaam were done in order to refine the interview instrument. The content of these interviews has not been used. The first interview contained mainly constructs from theory that were asked in a more closed manner. It was concluded that this approach straitjacketed the answers too much, because it resulted in short confirming or denying answers without the needed depth. It was expected that this approach to interviewing would exclude factors from being mentioned. The main interview question as mentioned in section 3.3.1 was tested in the second test interview. It was found that the open nature of this question was effective because the respondent was not steered towards an answer. Clarifying and deepening questions heightened the understanding of the answers. Theory driven questions were however not left out. Constructs from theory were introduced when they did not come up in the open questions.

A total of 23 interviews, each with duration of approximately one hour were conducted at 14 different organizations. The content of the interviews was amended after every interview to integrate the growing understanding from earlier interviews into later interviews. This means that data collection has been performed partly in parallel with data analysis. The progressing insights of the interviews raised new questions and gaps in understanding of earlier interviews. The strength of the data collection process used is that it allows for the needed flexibility and deep insights at a relatively small number of organizations. The choice was made to split up the interviews in 3 rounds. The first round consisted of 10 interviews at potential customers, in the second round the same 10 potential customers were interviewed a second time (from which 9 were reached again). There were 3 reasons to conduct second interviews with the same organizations. First of all it was expected that in the first interview the interviewees would hold back information because of trust issues and that in the second interview they would give deeper answers because of the established rapport and familiarity with the interviewer. This was a suggestion given to the researcher by Clemens Lutz, who is an associate professor who has experience with doing interviews in Tanzania. Second it was expected that the interviewer would have a learning curve and would thus get better in asking questions, which would allow for new learned skills to be used with earlier respondents. Third, it was

(20)

- 16 - expected that the communicative validity could be controlled this way. This means it gives the possibility to detect possible inconsistencies in answers between the first and second interviews (Flick, 2006). It was estimated that after 2 rounds of 10 interviews the incremental new insights would be minimized and theoretical saturation would be neared. After the first 2 rounds a third round of 4 interviews at 4 new organizations were done to see if the preliminary outcomes of the first 2 rounds would hold in order to help in generalizing and further confirming of the outcomes.

3.4 Process of data analysis

3.4.1 Grounded theory approach

All interviews were recorded using a sound recorder and afterwards transcribed as literally as possible (see Appendix 4 for an example of such a transcription). Transcripts were written down as literally as possible to prevent early interpretation. This is in contrast to taking notes, where the researcher is forced early on to make an interpretation of what has been said (Flick, 2006).

The ‘’grounded theory’’ approach described by Flick (2006) for analysis that is used to this end is especially suited for the exploration of unfamiliar territory. It offers a structure that can be followed in order to develop theoretical propositions from raw interview data. The grounded theory approach employs three broad procedures. The first is open coding, which involves labeling and categorizing. Codes are developed and continuously changed while coding to prevent from straitjacketing the data. The second step is axial coding, which involves finding relationships between the factors. There are several types of relationships that can be found in this step. This research is only interested in the cause-effect type relationships because it is aimed at identifying the determinants of adoption intention. The third step towards a grounded theory is called selective coding, which involves the further crystallization and refinement of the factors and relations that have been found. The analysis process finally stops when factors and relations are saturated and analysis of data no longer leads to new insights (Van Aken, 2007; Flick, 2006).

3.4.2 Use of Atlas.ti and detailed steps

Because of the relatively large amount of data a software tool called Atlas.ti has been used to help manage, extract, compare, explore and reassemble meaningful pieces of data in a systematic way, while leaving room for the needed flexibility (Muhr, 2004). In Table 2, 3 and 4 below the steps that lead to grounded theory are described in detail.

TABLE 2 Detailed steps in open coding Process 1: open coding Explanation

Step 1 After each interview codes have been attached to the empirical material as closely as possible to the transcriptions, which means as little interpretation as possible. Gradually, after more and more texts were analyzed similar concepts seem to appear. When this was the case the same label was used when possible, sometimes the label needed to be altered to fit both. During step 1 this was only done when very obvious similarities between concepts appeared. If a label would obviously match a concept from theory than the label was given this name. If however it was still ambiguous whether the label from theory was appropriate, a label closest to the transcription was

(21)

- 17 - used. This preliminary analysis was done directly after each interview so that new insights would be ready as input for later interviews. After having conducted all the interviews and have done the preliminary analysis a total of 137 codes resulted (see Appendix 1).Early impressions, associations, likely cause effect relations and other ideas of the researcher where linked to codes with a memo function of the Atlast.ti software program.

Step 2 With help of the network view function of the Atlas.ti software, the researcher could add concepts in a screen that were viewed as relevant. Relevance was determined in three ways. First, the amount of times a concept was mentioned in total. The more it was mentioned in total the higher the researcher perceived its relevance. Second, the amount of interviews in which the concept is mentioned. The more interviews in which a concept was mentioned the higher the research perceived the relevance. Third, the degree of ambiguity about the mentioned concept, which means the degree of clarity of the concept or the degree to which the researcher was certain that what the respondent said, was unambiguously tied to the concept. Some irrelevant concepts where thus left out of the network view screen judged on these three criteria. Because it involved a ‘’softer’’ issue of ambiguity it was no hard statistical choice, but rather a soft judgment by the researcher that took all three aspects of relevance together.

Step 3 In the network view screen the relevant concepts now appear in boxes that can be dragged around. The concepts that were similar to each other or where thought to at least have some similar aspects where dragged into groups. This led to the insight that many codes described more or less the same phenomenon. Multiple labels were put under single labels that described the underlying concept the most striking. This was not always a choice between which of the existing labels was superior, sometimes new codes where thought when it was necessary to combine labels to get a comprehensive code. Caution was taken at first to only combine the obvious combinations into a single label, still staying rather close to the text. By continuously going back to the original text new memos where attached that described the combinations from which the newer codes where made up. This resulted in 63 codes.

Step 4 In this step the researcher used the same strategy as in the previous step. Only now not only the obvious labels were combined but also the ones that have similar aspects. The coding at this point began to get more abstracted. A technique that Flick (2006) proposes was used at this point. Implying that the researcher moved continuously back and forth between inductive thinking (developing new codes) and deductive thinking (testing the codes against the transcriptions).This resulted in 41 codes on the network view screen. New memos had been attached and the code boxes that were thought to be somewhat similar or where in relation to each other where roughly dragged together.

(22)

- 18 -

TABLE 3 Detailed steps in Axial coding Process 2:

Axial coding Explanation

Step 5 In this step it became clear that there were two kinds of factors of the 41 concepts on the screen. The first where concepts that cause adoption intention and the second where concepts that describe the process of adoption intention. Since this research is only interested in factors that cause adoption intention, the process factors were left out at this point.

Step 6 The remaining factors were grouped around phenomena that the group had in common. After trying several ways to group the factors, the grouping method used by Frambach and Schilleweart (2002) was taken as an idea. Their model described adoption of high-tech b-2-b innovations in the context of developed countries. The factors that were identified could be put under one of four levels, the level of the innovation, the adopter, the supplier and the environment. The technique of Flick (2006) was used at this point again. By going back and forth from inductive thinking (developing concepts and relations from the transcriptions) and deductive thinking (testing concepts and relations against transcriptions) it was found that the groups of the supplier and the environment indirectly influenced the intention to adopt. It was only through first influencing the characteristics of the innovation that they seem to have influence. The factors on innovation and adopter level were found to influence the intention directly. At this point not all the labels of the factors and the groups where final yet.

TABLE 4 Detailed steps in selective coding Process 3:

Selective

coding Explanation

Step 7 This step begins with factors grouped among four levels. These factors are compared with factors used in existing literature. Oftentimes the factors from literature (and the definitions behind them) did not completely match the data and therefore the researcher used his own labels for the concepts. Some factors however seemed to match factors from existing theory perfectly. In these cases the label used in theory was used. Theory also helped with further condensing into more general codes.

Step 8 A definition was written down for all the codes, which was done by rereading the transcripts. This resulted in a sharpening of some factors to represent the concepts in a more striking and comprehensive way. The memos that where linked to the codes where used to help in the definition. These memos still partly held the history of what factors a concept was made up of. This was done until a point was reached where further coding or enrichment of categories no longer provided or promised new knowledge.

(23)

- 19 -

4. RESULTS

The data analysis process resulted in 11 determinants (see Figure 1), which seemed to either directly or indirectly influence the intention to adopt a low-tech b-2-b service innovation in a developing country. These identified determinants could be grouped into four levels of determinants, being perceived innovation characteristics, adopter characteristics, perceived supplier characteristics and environmental influences. Whilst the factors on the ‘’perceived innovation characteristics’’ level and the ‘’adopter characteristics’’ level seem to have a direct influence on the intention to adopt, the factors on the level of ‘’perceived supplier characteristics’’ and ‘’environmental characteristics’’ appear to affect ‘’intention to adopt’’ indirectly. The latter two factors appear to first influence the intervening factors on the ‘’perceived innovation characteristics’’ level. Figure 2 visualizes this basic relationship of the identified determinants of influence the identified factors have.

All levels and factors within these levels will first be explained and in turn the resulting model will be presented. The resulting model is called a multi-level model since the different determining factors could be grouped along four different sources that all represent a different level of influence.

4.1 Perceived innovation characteristics (direct determinants)

The first level of influence appeared to pertain to the perception the potential adopter had of the characteristics of the innovations. The identified factors in this group were found to directly influence the intention to adopt. Figure 3 shows the innovation characteristics that are described below.

- Risk (from innovation) - Relative advantage - Compatibility - Observability - Trailability - Switching barriers - Openness to innovation - Marketing efforts - Reliability - Included support

- Risk (from environment)

Indirect determining factors

Direct determining factors Dependent variable

Figure 1; List of 11 identified determinants of innovation adoption

ADOPTION INTENTION OF LOW-TECH B-2-B SERVICE INNOVATIONS IN DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

ADOPTION INTENTION OF LOW-TECH B-2-B SERVICE INNOVATIONS IN

DEVELOPING COUNTRIES: EVIDENCE FROM TANZANIA

TABLE OF CONTENTS

1. INTRODUCTION

2. THEORY

2.1 Defining a developing country

2.2 Low-tech innovation adoption in developing countries

2.3 B-2-B innovation adoption in developing countries

2.4 Service innovation adoption in developing countries

2.5 Innovation adoption models from theory

2.6 Generalizability of models across different types of innovations

2.7 Generalizability of developed countries to developing countries

2.8 Combining low-tech, b-2-b, service and developing country context

3. METHODS

3.1 Theory development based on a case study

3.2 Case selection

3.3 Process of data collection

3.4 Process of data analysis

4. RESULTS

4.1 Perceived innovation characteristics (direct determinants)