Determinants of user continuance intention towards mobile money services : the case of M-pesa in Kenya

Includes bibliographical referencesThe turn of the millennium witnessed the uptake and proliferation of mobile technology in developing regions. This occurrence has provided a medium for mobile telecommunication vendors within the region to create and offer services that are now accessible across socio-economic classes. A notable case of a widely adopted mobile technology-enabled service in the developing world is a mobile money service in Kenya called M-pesa. Since its inception, M-pesa has witnessed a mass adoption which has generally been attributed to prior lack of access by majority of individuals' in the country to affordable regulated financial services. M-pesa's presence has now been anticipated to afford a larger population the initial opportunity to harness economic benefits such as: increase money circulation, increase employment opportunities, facilitate social capital accumulation, facilitate savings, and promote financial autonomy, amongst others. Also, M-pesa based transactions in Kenya are reported to exceed those of western union globally. Whilst M-pesa presently vaunts large user adoption numbers, it is the first of its kind in the region to amass such achievement. Further, historically: products and services of similar nature to M-pesa have been unsustainable. A case of M-pesa's demise would have dire implication for the Kenyan economy and 30% of the households in the country that rely on it for remittances. To understand this phenomenon, extant studies have examined the drivers of adoption of this service but have slacked in subsequent investigations to understand user continuance with the service. As such, the information systems literature cautions that initial adoption of technology, although crucial, does not guarantee sustained use. Therefore it is imperative to investigate drivers of continuance. In general, extant research has not focused on investigations of user continuance intention in Africa. In response, this thesis presents an African based study on the determinants of user continuance intention towards M-pesa. Specifically, the purpose of this study was to i) identify and discuss factors from the literature that are most likely to influence user continuance intention towards M-pesa, (ii) develop a research model that is grounded in theory, (iii) test the model within the sample context to identify the antecedents and determinants of user continuance intention towards M-pesa in Kenya. A broad, critical review of the relevant literature provided basis for hypothesized relationships between the identified factors. A formal survey of users of M-pesa in Kenya comprised the phase of data collection and resulted in a usable data set of (n=434). The data collected from the respondents within Kenya was relied upon to test the hypotheses. The survey instrument used to measure the study's constructs was developed via a process of literature review, expert pre-testing, pilot testing, and statistical validation. Partial Least Square and Artificial Neural Network analyses were used to examine the study's measurement and structural model comprising variables of : behavioural beliefs (post-usage usefulness, confirmation, satisfaction), control-beliefs (utilization and flow), object-based beliefs (perceived task-technology fit, system quality, information quality, and service quality), and attitudinal belief (trust). Collectively, the afore-listed ten independent variables and one dependent variable (continuance intention) comprised the study's model. Four of the independent variables (utilization, satisfaction, flow, and trust) were hypothesized to directly determine continuance intention. Of these four, all emerged as determinants of continuance intention. However, trust emerged as the strongest determinant, subsequently, utilization, flow, and satisfaction respectively. The result was unexpected, as satisfaction (a behavioural belief) has been presented in the extant literature as the dominant determinant of continuance intention but does not hold a consistent predictive strength in a developing world. Its predictive power was diluted by trust, utilization, and flow amongst the Kenyan sample. The study's model revealed an R² of 0.334. The analyses demonstrated that user continuance intention is determined by factors across object, control, attitudinal, and behavioural beliefs. The unexpected finding of the rankings of predictive strength of the factors turns a new leaf and introduces areas of further inquiry in future studies. The study concludes with realized contributions to theory and important guidelines for current and future technology-enabled service vendors in developing regions