When individuals try to predict their future memory performance, they do not have direct access to it. Instead, they rely on probabilistic cues. Previous research suggests that when people form judgments of learning (JOLs) and multiple cues are available, they integrate them in a compensatory manner rather than focusing on a single cue. However, past experiments, which aggregate data from multiple trials, do not allow to address potential dynamic changes in cue usage. Our study introduces a new experimental design where one cue (the context cue) remains constant across sequences of trials, while another cue (the target cue) changes randomly from trial to trial. We investigate whether the number of trials since the last context cue switch influences how both cues affect JOLs. Our findings reveal that the influence of the context cue decreases over successive trials when it remains constant. In our first experiment, we found no interaction between the target cue and the number of trials since the last context cue switch. This study supports the idea that people use multiple cues when judging their future memory performance and shows that there are dynamic changes in cue influence over time.