Purpose – This paper addresses a methodological gap in Human-Computer Interaction (HCI) research by demonstrating how Partial Least Squares Structural Equation Modelling (PLS-SEM) can be used to analyse complex user behaviours, perceptions, and experiences in interactive systems, particularly within exploratory and user-centred research contexts. Design/methodology/approach – The study situates PLS-SEM within HCI research and outlines practical steps for model specification, validation, and interpretation under conditions common in applied HCI studies, such as model complexity, non-normal data, and limited sample sizes. A demonstrative case example is included, drawing on survey data from 243 users of a mobile health (mHealth) application to illustrate the approach. Findings – The illustrative results indicate that perceived ease of use, system quality, and overall user experience influence behavioural intentions toward continued app use. User satisfaction and holistic experience emerge as stronger predictors than perceived usefulness alone, demonstrating how PLS-SEM can be used to examine multidimensional user experience constructs and relationships among latent variables. Originality – Moving beyond traditional regression and covariance-based SEM approaches, this paper contributes a methodologically focused perspective by illustrating the rigorous application of PLS-SEM in HCI research, including the use of reflective and formative measurement models within a single analytical framework. Practical implications – The paper provides a practical roadmap for HCI researchers and practitioners applying PLS-SEM in studies involving emerging digital applications, small to moderate samples, or non-normal data, supporting user-centred modelling in applied digital contexts.