What resonates most? An investigation on the factors of customer satisfaction in senior living communities

Abstract

Senior living communities (SLCs), specialized residential facilities designed to support older adults, have attracted increasing scholarly and practical attention amid rapid global population aging. As demand for these SLCs grows, prospective residents and their families increasingly rely on online customer reviews as a critical source of information to inform their decision-making. Drawing upon expectation-confirmation theory, this paper proposes an explainable machine learning model that integrates eXtreme Gradient Boosting and Shapley Additive exPlanations algorithms to investigate the impact of key variables on customer satisfaction in SLCs. In addition, we employ text clustering and multiple correspondence analysis to uncover dominant experiential themes and stakeholder-specific concerns embedded in online review narratives. Our findings reveal that care services make the largest marginal contribution to customer satisfaction and identify 37 clustered topics reflecting diverse expectations across facility types and customer segments. By integrating interpretable artificial intelligence with ECT, this study offers both methodological advances and actionable insights into satisfaction formation in institutional care settings.