A collaborative filtering model incorporating media promotions and users' variety-seeking tendencies in the digital music market

Abstract

Understanding customer preferences and providing the right products at the right time to customers via personalized recommendations have been among the major interests of online retailers and service providers. This paper proposes an improved collaborative filtering model that incorporates a firm's marketing effort variables (i.e., media promotional variables) to improve the prediction of customers' digital music choices. In addition, we assert that the predictive model's effectiveness is different for consumers depending on their varietyseeking tendencies in music. We compared our predictive model to benchmark models and demonstrated that our proposed model is superior in predicting users' download behavior. We also found that the overall predictive performance is higher for active variety seekers who consume diverse types of music via streaming. We provide some evidence that this may be due to differences in the degree to which the two groups are influenced by different types of media promotions. The results suggest that considering psychological characteristics such as variety-seeking tendencies provides more advantages in prediction and recommendation systems, which opens new avenues for improvement.