Purpose Continued debate surrounds whether or not offender profiling is a valid practice. Critics have mainly contended that few studies have produced clear, quantifiable, evidence of a link between crime scene actions (A) and offender characteristics (C). Arguing that this is due to a failure to study offender actions as part of a dynamic decision‐making process, this study sought to identify action phases that are representative of the general decisions an offender must make during the commission of a sexual offence, and relate these decisions to known characteristics of the offender (C). Methods Two‐step cluster analyses were performed on data from 347 stranger sexual offences, committed by 69 serial sexual offenders, by action phase: (1) search; (2) selection; (3) approach; (4) assault; and also for an offender's (5) characteristics. Multiple correspondence analysis (MCA) was then utilized to investigate the inter‐relationship of action phase clusters and offender characteristics. Results The MCA results indicated that specific behavioural macro‐clusters formed across the various actions phase and offender characteristic clusters in a meaningful way. Additionally, the macro‐clusters themselves corresponded to the extant literature on sexual assault, revealing several points of congruence between offender crime scene actions and offender characteristics. Conclusion The results of the study demonstrate that when crime scene behaviours are interpreted within a dynamic decision‐making process (i.e., utilizing action phases), reliable and valid empirical links may potentially be drawn between an offenders behavioural actions and their characteristics.