Purpose -The current research aims to explore variables that explain the differences in online reviewers' behavior. Design/methodology/approach -The authors' panel dataset is compiled from Amazon.com's book section and uses publicly available information about reviewers in combination with the reviews they wrote. The authors utilize the Pareto/NBD model with time-invariant covariates. The model's parameters are estimated using maximum likelihood estimates (MLE) with MATLAB software. Findings -This study contributes to the literature by exploring how the characteristics of reviews and the reviewers might shape consumer review frequency and continuity. Specifically, the authors' results show that review ratings, comments on a review, and helpful votes have a positive association with review frequency and continuity. Furthermore, the length of the textual review has a positive relationship with review frequency, but a negative relationship with review continuity. Relative to anonymous reviewers, people who write reviews and use their real names post reviews less often, but their review continuity is longer. Originality/value -This paper is the first to identify empirically variables that explain review frequency and continuity, thus enabling companies to gain a better understanding of their reviewer base and hence to manage online word-of-mouth more efficiently. It is the first study to apply a well-known behavioral model to address online review behavior.

Heterogeneity in Internet diffusion across countries is the core issue of the global digital divide-a phenomenon that has attracted wide academic discussions and various international initiatives. However, prior studies on the global digital divide have addressed the issue mainly from a crosssectional perspective. This article attempts to examine the topic from a longitudinal perspective by using the Bass new product diffusion model, which has been successfully applied in diffusion studies over the past four decades. Summarizing information from an empirical study on 48 countries, the modeling framework finds three types of divide dynamics; assesses the severity of the global digital divide at various Internet diffusion stages; and identifies cultural, economic, and educational factors as the major causes of the global digital divide at different stages. The article concludes with a discussion on marketing and policy-making implications.

No abstract