Purpose – The purpose of this paper is to examine the characteristics of virtual social networks (VSNs) and to determine their salient attributes, including those that influence flow experience, trust and electronic word-of-mouth (eWOM) behaviors affecting users’ purchase intentions, and to provide important strategic implications contributing to the Internet marketing literature. Design/methodology/approach – Using a self-administered questionnaire, the data ( n = 167) are collected from two Iranian Internet social networking sites, namely, facenama.com and cloob.com . Using LISREL 8.5, hypothesized relationships are examined through structural equation modeling (SEM) analysis. Findings – According to the results of the study, despite all assumptions and studies to the contrary, eWOM behaviors in VSNs are derived from neither users’ flow experience nor their trust in VSNs, but they are mostly caused by VSNs’ attributes, from which four are investigated in this study, namely, communication and social relationships, entertainment, information disclosure and ease of use. Nevertheless, according to the results, VSN attributes also influence trust and flow experience, trust in a VSN environment influences users’ flow experience and eWOM in VSNs has significant impact on users’ purchase intentions. The findings also revealed that the level of education of a user affects how much he trusts the VSN environment. Practical implications – The author examined flow experience, trust, eWOM behavior, purchase intentions and the VSNs’ attributes to verify their relationships, providing a better understanding of an effective indirect marketing in VSNs. The results also have important implications for researchers. Originality/value – While flow experience, trust, word-of-mouth (WOM) behaviors and purchase intentions have been separately studied in Web sites, e-shopping malls and blogs, little research has sought to identify the existence of these elements within VSNs, their correlations with one another and how they are affected by VSNs’ attributes.
