--- - |2 Abstract While we often engage in conversational reminiscing with loved ones, the effects of these conversations on our memory performance remain poorly understood. On the one hand, Wegner's transactive memory theory predicts that intimate groups experience benefits from remembering together. On the other hand, research on collaborative recall has shown costs of shared remembering in groups of strangers—at least in terms of number of items recalled—and even in intimate groups there is heterogeneity in outcomes. In the current research, we studied the effects of particular communicative features in determining the outcomes of collaborative recall in intimate groups. We tested 39 older, long‐married couples. They completed a non‐personal recall task (name all the countries in Europe) and a personal recall task (name all your mutual friends), both separately and together. When they collaborated, we recorded their conversation. We coded for specific “communication variables” and obtained measures of “conversational style.” Overall, we found two clusters of communication variables positively associated with collaborative success: (a) cuing each other, responding to cues, and repeating each other; and (b) making positive statements about memory performance and persisting with the task. A negative cluster of behaviors—correcting each other, having uneven expertise, and strategy disagreements—was associated with less interactive, more “monologue” style of collaboration, but not with overall recall performance. We discuss our results in terms of the importance of different conversational processes in driving the heterogeneous outcomes of group remembering in intimate groups, suggesting that a focus on recall output alone limits our understanding of conversational remembering. - Topics in Cognitive Science, EarlyView.