Predicting Determinants of IoT Adoption In Botswana's Smes Using Hybrid Sem-Ann Approach

Abstract

Previous research on Internet of Things (IoT) adoption in Small and Medium Sized Enterprises (SMEs) has predominantly focused on contexts outside Botswana, consequently restricting the findings' direct applicability to the local setting. This study examines factors influencing IoT adoption in Botswana’s SMEs. A cross-sectional survey of 440 SMEs in Gaborone city tested a multi-theoretical model using a hybrid Structural Equation Modelling - Artificial Neural Network (SEM-ANN) approach. Both the hidden and output layers of the Multilayer Perceptron Artificial Neural Network were activated using the sigmoid function. Model training utilized a Batch-based approach, and the optimization of the model parameters was achieved through the scaled conjugate gradient algorithm. Findings revealed significant factors including; relative advantage, complexity, compatibility, trust, security, technology readiness, competitive pressure, top management support, financial resources, government support, perceived usefulness, vendor support, and perceived ease of use. Additionally, sensitivity analysis underscored compatibility as the most significant factor, followed by security and privacy, technology readiness, top management support, and financial resources. Educational level and computer experience significantly moderated IoT adoption among Botswana's SMEs. Findings contribute to better understanding IoT adoption by Botswana’s SMEs, providing valuable insights for policymakers and decision-makers. The proposed framework can guide IoT adoption policies in Botswana and other emerging economies. Future research could investigate additional factors and approaches to enhance understanding of IoT adoption in SMEs.