Uncovering effects of virtual endorser's attractiveness on consumer behaviors from the perspective of national culture.

In the context of social media, more and more enterprises take virtual endorsers for brand marketing. Investigating people's preferences and concerns about virtual endorsers' characteristics and seeking to understand users' feedback on the effectiveness of endorsement with the help of reviews are important for companies to expand their digital marketing strategies. We use the machine learning approach, conduct text mining and regression analysis on the advertisements and reviews, and find the positive effect of attractiveness on purchase intention and the suppressing effect of advertising perception between virtual endorsers' attractiveness and purchase intention. Users' high perception of advertising information deepens their impression and improves the effectiveness of advertisements. However, that may lead to psychological revulsion and purchase inhibition, the regression coefficient of advertising perception toward purchase intention is negative and the effect is significant. Fortunately, we confirmed the moderating role of positive emotion to weaken the inhibitory effect. Additionally, the study validates the moderating effect of national culture on the relationship between virtual endorser's attractiveness and purchase intention. Our findings have positive implications for studying consumer psychology and behaviors by big data analysis, it gives practical suggestions for enterprises to adopt the virtual endorsers' marketing strategy more effectively and provides references for enterprises to facilitate the digital marketing transformation on social media. [ABSTRACT FROM AUTHOR]