ModelingThe Effects of Verbal and Visual Marketing Content in Social Media Settings:
TheMediating Role of Consumer Engagement
刘蕾，中央财经大学广告学系副教授。2015年获得清华大学管理学博士学位。主要研究兴趣是社会化媒体营销、社会互动、整合营销沟通、广告策略。主要研究成果发表于Journal of Advertising Research，Journal ofMarketing Development and Competitiveness，《心理学报》《清华大学学报（哲学与社会科学版）》《管理科学》《中国出版》等期刊。
讲座摘要：Dueto the relentless pace of developments in social media and informationtechnology, firms increasingly rely on a combination of verbal and visualelements to communicate with consumers. The present research investigates theimpacts of text-image information on customer engagement and corporate value.Based on a large scale data on Sina Weibo, we employ a Natural LanguageProcessing algorithm to characterize text contents as function-oriented(information that is helpful for increasing consumer knowledge about theproduct, brand, or company) and social-bond oriented (i.e., informationconveyed on a social level, to create connections with customers) types.Further, using Deep Learning techniques, we respectively calculate eachmessage’s relevancy and expectancy levels between text and images, which arethe two dimensions of incongruency. The results indicate that compared tofunction-oriented text, social bond-oriented text that aims to connect withconsumers emotionally exerts a larger effect on firm value measured by abnormalreturns (AR) through the mediating role of the number of likes. Moreinterestingly, we find that forward, comment and “Like” are virtually differentin antecedent factors and influence. Specifically, there exist inverted-Urelationships between expectancy and the number of forwards, the number ofcomments as well as the number of likes. But expectancy doesn’t exert a furthereffect on firm value. In contrast, there is an U-relationship between relevancyand the number of likes, which further drives AR.