Classifying Business Marketing Messages on Facebook

Bei Yu, Syracuse University
Linchi Kwok, Syracuse University

Article comments

Yu, B., & Kwok, L. (2011, July). Classifying business marketing messages on Facebook. Empirical full paper presented in the Internet Advertising (IA 2011) Workshop at the 34th Annual International ACM SIGIR (Association for Computing Machinery; Special Interest Group on Information Retrieval) Conference, Beijing, China.

Abstract

Companies are increasingly using social media for marketing purposes. In this study we first demonstrate that although the majority of company posts on Facebook are aimed for direct sales and promotions, it is their communication messages that received the most attention from customers. We then trained an SVM classifier to automatically separate these two kinds of messages, hoping to use this tool to analyze messages from many companies and consequently monitor the evolution of their social media use over time. We found that the classifier trained with tf-idf weighted part-of-speech features performed best. It is better than classifiers trained with word features. Combining feature sets did not improve the performance. Feature ranking results show that this best-performed classifier captured the genre characteristics of direct marketing and communication messages.