It is unknown how co-rumination, or perseverating on problems or feelings with another person, unfolds in the daily lives of romantic couples. Using a variance decomposition procedure on data from a 14-day dyadic diary, we assessed how much variance in co-rumination was attributable to temporally stable and varying factors, as well as whether co-rumination is better measured as a couple-level or individual-level process. Within-person, within-couple fluctuations in co-rumination contributed most (~33%) to the total variance and summary scores based on these fluctuations were reliable. Stable between-couple differences accounted for ~14% of the total variance and could also be reliably assessed. However, within-couple agreement in co-rumination was low, such that the reliability at the level of within-couple change was inadequate. Research is needed to understand these divergent perceptions of co-rumination and potential downstream consequences. We conclude by considering how these results inform theory and can be applied to similar dyadic constructs.