To better understand the drivers of differences in U.S. Navy officer retention and promotion across demographic groups, this analysis identifies features associated with differences in predicted promotion outcomes across race/ethnicity and sex-based groups of Navy line officers commissioned as O-1s in 2001-2018. Using administrative military personnel data, we train two machine learning (ML) models: one predicting retention and another promotion. We separately calculate the effect of each feature on each individual’s likelihood of retention or promotion, and compare across demographic groups to identify differences in which and how much features matter for each group. Findings suggest that retention differences may be structural in nature and offer insights worth exploring in future research.