More than words can say: a multimodal approach to understanding meaning and sentiment in social media

Abstract

ABSTRACT The authors develop a multimodal social listening analysis (MSLA) approach as a framework for managers to understand how meaning is constructed in social media posts using both text and other media. The research adds to AI and text analysis approaches by considering the whole meaning of a post rather than an analysis of subsets of information in text and other media. The use of MSLA is validated across the social media platforms of Facebook, Twitter and Instagram. The findings show that MSLA helps (i) reveal structures in what appear to be unstructured multimodal posts; (ii) identify all the sentiment items in a post; (iii) identify implicit meanings, such as irony, humour and sarcasm; and (iv) further identify emotions and judgements in multimodal communication. Importantly, this paper explains how decisions and opinions are made online and how marketing strategies can be tailored towards meanings derived from multimodal communication in social media.