Understanding facebook commerce (f-commerce) actual purchase from an artificial neural network perspective

• Main Authors: Leong, Lai Ying, Jaafar, Noor Ismawati, Ainin, Sulaiman
• Format: Article
• Published: California State University, Long Beach 2018
• Subjects: HF Commerce; QA75 Electronic computers. Computer science
• Online Access: http://eprints.um.edu.my/21483/; http://www.jecr.org/sites/default/files/19_1Paper5.pdf

Despite the abundance of studies in electronic commerce, few studies have validated the antecedents of actual purchase from the perspective of Facebook commerce or f-commerce. Most of the existing e-commerce studies have focused on purchase intention and little attention has been paid on consumers' actual purchase especially from the f-commerce context. This study intends to examine the effects of demographic variables, Web Usage Theory, Trust Transference Theory and F-commerce usage behaviors in predicting f-commerce actual purchase. The instrument was rigorously developed and validated using expert panel, Q-sort procedure, pretest and pilot test. Several issues of validity in previous studies were addressed. Unlike existing studies which engaged compensatory linear models such as SEM, PLS, MLR and etc., in this study 808 f-commerce users were selected and the data is analyzed using the non-compensatory and non-linear artificial neural network (ANN) model. ANN can overcome challenges encountered by conventional statistical analysis that relies on p-value caused by false correlations. The findings reveal that consumers' experience is the strongest predictor followed by Facebook usage, hedonic motivation, browsing, age, trust motivation, participation, utilitarian motivation, number of children, monthly income and educational level. Theoretical and managerial contributions were provided for scholars and practitioners of f-commerce.