This study examines the relationship between work–life balance (WLB) and job performance (JP) among university personnel, and whether perceived organizational support (OS) moderates this relationship. Using secondary survey data from 242 staff, we combined path analysis with a complementary machine-learning pipeline. Data were preprocessed via iterative imputation, winsorization, removal of near-constant items, and robust scaling; six principal components summarized WLB/OS features for clustering. Results show that WLB positively predicts OS, and OS positively predicts JP, whereas the direct path from WLB to JP is not significant, indicating complete mediation (indirect effect = 0.28). Employee segmentation yielded a four-cluster solution selected for parsimony and interpretability, with high stability (median ARI = 0.956). Clusters differed in overall job performance (Kruskal–Wallis H (3) = 37.49, p = 3.63 × 10⁻⁸), and were interpretable as Thriving, Supported but Stretched, Constrained, and At-risk. The findings position OS as both the mechanism translating WLB into performance and a practical axis for actionable segmentation. Implications include tiered support strategies that maintain conducive conditions for Thriving staff, targeted supervisory and informational supports for Supported-but-Stretched groups, the removal of systemic frictions for Constrained personnel, and integrated packages for At-risk employees, offering a scalable roadmap for HR in higher education.
