From words to insights: Text analysis in business research

Business success relies on effective stakeholder communication, much of which occurs via or can be transcribed into text. Yet, business researchers often lack coherent frameworks to conceptualize business-relevant communication and its underlying logics. We thus consider business research from a message design logic lens to offer a conceptual foundation for research seeking to understand the content, style, and structure of business communication. Business researchers also underutilize modern tools for analyzing text data. Hence, our comparison of current methodologies for analyzing text (i.e., topic models, dictionaries, supervised machine learning, and large language models) points out their respective advantages, limitations, and applications. An overview of recent studies in the Journal of Business Research identifies how these methods are used to extract insights from business communication. We offer practical guidelines for authors and reviewers on method selection, implementation, and evaluation, and conclude by proposing future directions for business research using text data.