Abstract
Adopting a combination of qualitative textual analysis, human-based content analysis, and machine learning–based data mining, we propose a procedure to analyze user-generated content (UGC) on social media using Dove’s “Campaign for Real Beauty” as a case for demonstration. We provide a guideline to explicate all six steps of the analysis procedure: topic identification through qualitative textual analysis, generation of labeled data through human coding, data preprocessing, evaluation of machine learning–based classifiers, topic classification of unlabeled data, and conducting research. The study has important methodological implications for advertising scholars and practitioners.