How EWOM Convergence Influence The Perceived Helpfulness In The Perspective Of Social Influence

With more and more customers relying on electronic word-of-mouth(eWOM)information to assist their purchasing decisions,the powerful online comment force makes more and more brands shift from the traditional marketing method of "goods" to emphasizing the value of managing "people" in the social retail era,pursuing social fission and encouraging consumers to speak for brands.At the same time,however,with the prosperity and development of e-commerce,overwhelming comments are pouring out,which makes it difficult for consumers to filter information.On the one hand,consumers would like to find as much reference information as possible.On the other hand,they want to avoid information overload and simplify their own purchasing process.To solve this problem,some research has proposed the concept of "perceived helpfulness/diagnosis" to measure what kind of eWOM information is more useful or what eWOM information can better assist customers understand and evaluate the performance of products from the perspective of readers.After reviewing previous literatures,some studies have already discussed the influencing factors of perceived usefulness in a comprehensive way,but most of the studies assume that each online comment exists independently.However,both publishers and readers of comments will inevitably be influenced by word-of-mouth information released by others before writing comments or reading comments.Publishers of word-of-mouth may follow other people’s product opinions in the aspect of content and emotion,because of informational social influence and normative social influence.Therefore,readers may perceive a lower sense of helpfulness for these similar online comments because of information redundancy and speculation of herd motivation.From this point of view,based on the dual-process theory of social influence and information adoption model,this study introduces the concept of eWOM convergence,which captures the similarity between one comment and other people’s comments in the aspect of information and emotion,reflecting the degree of social influence on one’s eWOM.Here,information convergence is calculated by cosine similarity of text,while emotion convergence is measured by cosine similarity of emotion score calculated based on LSTM model.Furthermore,the negative binomial regression method is used to measure the influence of the similarity on the perceived usefulness of comments.Factors of information quality(the length and extreme valence)and information source credibility(the professionalism and connectivity of publishers)are added as moderators.The model further tests the interaction effect between social influence and moderators(information quality and source credibility)on perceived helpfulness.The theoretical significance of this study lies in revisiting the perceived helpfulness of eWOM from the perspective of social influence,making up for the neglect of the influence from other people’s eWOM in previous literatures,innovating and expanding the original related theories.The practical significance of this research lies in helping enterprises understand the communication mode of electronic word-ofmouth from the perspective of social influence,digging out more useful word-ofmouth information,and providing theoretical support for enterprises to carry out word-of-mouth publicity,design and simplify the purchasing decision-making process for customers and manage its virtual communities.