Social network analysis (SNA) has emerged as a significant research field, that helps in understanding interpersonal relations and information dissemination within the social network. Community detection plays a pivotal role in SNA identifying cohesive groups of individuals which in turn helps in identifying the underlying structure, analyzing patterns, and facilitating targeted interventions in various domains. In this paper, a novel community detection technique based on similarity matches is presented. We propose a similarity-based score that leverages the inherent structures of the social network to detect communities and carry out impact analysis. We demonstrate the effectiveness of our algorithm by comparing it with machine learning (ML) techniques. Furthermore, we discuss the implications of our work for real-world applications highlighting its potential for enhancing community detection in diverse domains, in particular for paid promotion tweet analysis.