Understanding dexterity is a critical factor when designing physical and interactive products, shaped by the unique proprioceptive and musculoskeletal traits of users. The extent to which these individual differences manifest during physical product interactions and the methods to effectively quantify them remain largely unexplored. Measurement of subtle characteristics in hand–object interactions could assist researchers and practitioners to better understand how users interact with products, leading the way for more refined, accessible, bespoke or adaptive products tailored to individuals’ dexterity and usage. This paper investigates (1) individual differences within object interactions for single-handed highly dexterous tasks and (2) the feasibility of data-driven measurement of dexterous interaction. A study explores the ability of data-driven techniques to identify individual differences and characterise dexterous interaction, for (i) an unconstrained hand–object interaction scenario and (ii) a constrained hand–tool–object manipulation scenario. Despite a reduction in performance variance during the constrained task, the classification of user actions remained heavily dependent on participant-specific features. Models trained on group data failed to generalise to new users, highlighting the significant inter-participant variability in dexterous strategies, even under constrained conditions. Our results demonstrate that user-specific data capture could aid personalised product development and provide recommendations for implementation in future work.