A 2020 perspective on “Predicting the influence of viral messages for VM campaigns on Weibo”

Abstract Viral marketing (VM) in social networks (SNs) has the advantages of rapid distribution with comparatively little effort and is easy to reach the “hard-to-get” audience. Consequently, it has been being actively used by marketers to entice users and advertise products. The massive and diversified data in SNs presents opportunities to uncover the key to ensure the persistent and broad spread of a viral message, thus leading to successful VM campaigns. First, considering that a viral message is an integration of text, images, and videos, we concentrate on extracting significant features embedded in images and videos. Second, given that users often interact with each other and companies, the viral message may be enriched through its propagation. To accurately predict its influence, it is appropriate to account for the influence of the message itself and its derivatives. Third, identifying frequent interactions among users who have similar preferences or interests will contribute to well-targeted and effective VM campaigns.