Steering Social Media Promotions with Effective Strategies
Kun Kuang,Meng Jiang,Peng Cui,Hengliang Luo,Shiqiang Yang
DOI: https://doi.org/10.1109/tbdata.2017.2734102
2018-01-01
IEEE Transactions on Big Data
Abstract:On social media platforms, companies, organizations and individuals are using the function of sharing or retweeting information to promote their products, policies, and ideas. While a growing body of research has focused on identifying the promoters from millions of users, the promoters themselves are seeking to know what strategies can improve promotional effectiveness, which is rarely studied in literature. In this work, we study a new problem of promotional strategy effect estimation which is challenging in identifying and quantifying promotional strategies, as well as estimating effectiveness of promotional strategies with selection bias in observational data. Here we study a series of strategies on both context and content levels. To alleviate the selection bias issue, we propose a method based on Propensity Score Matching (PSM) to evaluate the effect of each promotional strategy. Our data study provides three interpretable and insightful ideas on steering social media promotions, including (1) three significant and stable strategies, (2) a critical trade-off, and (3) different concerns for promoters of different popularity. These results provided comprehensive suggestions to the practitioners to steer social media promotions with effective strategies.