<p>Cancer-related cognitive impairment is common and distressing. Research has primarily focused on differences in mean-level performance on laboratory cognitive tasks, but recent work has investigated within-person fluctuations, or intraindividual variability (IIV), in performance on laboratory-based tasks. The present study of ambulatory cognitive assessments in daily life evaluated whether cancer survivors differed from a comparison sample in mean-level performance, IIV, and within-group heterogeneity, as well as change in mean performance and IIV across the study. Breast cancer survivors (<i>N</i> = 47, <i>M</i>_age = 52.9) and a comparison sample (<i>N</i> = 105, <i>M</i>_age = 51.8) completed intensive longitudinal design protocols involving smartphone-based cognitive assessments up to 5 times per day for 2 weeks to measure spatial working memory, working memory updating, and processing speed. Participants were female and in the United States. Bayesian mixed-effect location-scale models were conducted. Survivors had better mean-level performance on cognitive tasks, and both groups showed improvements across the study. Survivors showed greater IIV in processing speed and working memory updating, but there was no credible evidence for a difference in spatial working memory. IIV also changed across the study, such that both groups became less consistent in processing speed and working memory updating, and survivors became more consistent in spatial working memory. There was little evidence of group differences in within-group heterogeneity. Our findings indicate that IIV, as a potential marker of instability or sensitivity to contextual factors, may be an important index in the study of cancer-related cognitive impairment. These results highlight opportunities to use intensive longitudinal designs to consider indices beyond individual differences in mean-level performance.</p>