Presenta: Francisco Villarroel, University of Bologna.
Abstract: This presentation shares insights on how to use unstructured data (UD) for social media marketing research. I will discuss my experiences with developing measurements from multimodal content (e.g., using Vision Transformers and GenAI), exploring mediation effects in field data, and addressing modeling considerations such as endogeneity and robustness checks. The session will also highlight how research findings can be translated into interactive, research-driven apps that foster both academic replication and practical implementation. To illustrate these ideas, I will present four recent projects in which I have used text, audio, and images to understand behavioral outcomes. These projects address timely marketing topics, including optimizing images and text overlays in brand posts, the role of virtual influencers, conversational podcasts, and customer experience derived from online reviews. Collectively, these examples demonstrate diverse approaches to constructing quantitative studies that generate actionable insights for social media marketing.
