Sentiment Assessment of Brand Advertising on Gender Issues on Social Network: A Case Study of Femvertising on Sina Weibo in China

With the popularity of social media platform and feminism on the internet, the marketing activities on female issues also known as femvertising has become a new and promising field of brand marketing. Considering the growing female awareness in social networks and the lack of quantitative studies, online brand marketing on female and gender issues remains a tough challenge for enterprises particularly. In this paper, we investigate three real cases of brand marketing on female issues and quantitate assessment of different-gender customers' attitudes towards femvertising through social media networks is conducted. Representative machine learning methods like SVM and state of art deep learning methods: Bert with LSTM and Bert with Attention mechanism plus LSTM model are compared to train the sentiment classification model for Weibo posts. Among them, Bert with Attention mechanism plus LSTM model outperforms others with an f1 macro score of 98.0998%. Then we use the best model to predict the polarity of posts. It turns out that female customers are more sensitive to femvertising than male ones. And femvertising focusing on empowering girls and women could be a great boost to positive influence on brand building. While simply employing controversial celebrities on gender issues won't please either side.