We describe a practical approach to tackling observed heterogeneity using partial least squares structural equation modelling (PLS-SEM) when the number of categorical variables is high and the context of the research is exploratory. The approach is based on combining classical multigroup PLS-SEM approach and pathmox analysis. We provide practical guidance on using our hybrid multigroup PLS-SEM and illustrate its application using real data for bank employees. In investigating work climate, specifically the relationship between satisfaction and loyalty considering specific drivers (empowerment, company reputation, leadership, pay, and work conditions) and different sources of heterogeneity (gender, age, marital status, education, job level, and antiquity), the hybrid multigroup PLS-SEM identified three partitions defined by juniors, seniors, and managers, and identified significant differences between those groups, specifically in indicating that leadership and pay were more important for juniors, empowerment for seniors, and company reputation and work conditions for managers.