Trends in global working life suggest that the occurrence of jobs characterized by long-lasting low-level loads or repetitive operations is increasing. More physical "variation" is commonly believed to be a remedy against musculoskeletal disorders in such jobs. One aim of the present paper was to shortly review the validity of this conviction. An examination of the available epidemiologic literature pointed out that the effectiveness of initiatives like job rotation or more breaks is weakly supported by empirical evidence, and only for short-term psychophysical outcomes. Only a limited number of studies have been devoted to physical variation, and concepts and metrics for variation in biomechanical exposure are not well developed. Thus, as a second objective, the paper proposes a framework for investigating and evaluating aspects of exposure variation, based on explicit definitions of variation as "the change in exposure across time" and diversity as "the extent that exposure entities differ". Operational methods for assessing these concepts are also discussed.