Responses to exposure therapy vary across individuals, emphasizing the need for a deeper understanding of its underlying mechanisms. This study examined two key processes during exposure that serve as readouts of different clinical rationales: (1) within-session fear reduction (measured as the decline from peak to end fear within an exposure exercise) and (2) threat expectancy processes (assessed via expectancy violation, expectancy change, and learning rate). Data from 516 patients with anxiety disorders who completed at least 10 exposure exercises in a clinical trial were analyzed. Results showed that expectancy measures and fear reduction were only weakly correlated within exposure exercises. While no significant differences were found in their time courses, both readouts independently predicted treatment success. Specifically, a higher learning rate and greater relative fear reduction were associated with better outcomes. These findings highlight the clinical relevance of monitoring fear reduction and expectancy-related readouts as indicators of two distinct exposure rationales — the fear reduction rationale and the threat expectancy rationale. Although it remains unclear whether they reflect separate mechanisms of change or different aspects of a shared mechanism, addressing both rationales may help optimize and personalize exposure therapy.