SEGMENTATION STRATEGIES FOR POLITICAL MARKETING ON SOCIAL MEDIA: A CLUSTER ANALYSIS

Abstract

The rise of social media and the "network society" (Castells, 1999) has transformed political communication, rendering obsolete electoral segmentation based solely on demographic data. Given the complexity of the digital environment, this study investigates how clustering techniques applied to behavioral data can extract more precise electoral profiles. Based on benefit-based segmentation (Haley, 1968) and micro-segmentation (Zuboff, 2019), the research adopted a quantitative-deductive approach. The AutoCluster framework was used to analyze Instagram data from Brazilian governors, evaluating metrics of engagement, frequency, and communication structure. The analysis identified three distinct profiles of digital activity: Institutional Manager with consistent activity and moderate engagement; Efficient Mobilizer with a high proportional rate of interaction; and Mass Communicator with a high volume of publications and wide reach. The study demonstrates that behavioral clustering surpasses traditional typologies, offering more sophisticated strategies for political marketing. For future research, the use of Natural Language Processing (NLP) is recommended for more in-depth qualitative analyses.