Early withdrawal from higher education (HE) programmes can be detrimental for the students and institutions involved. Quantitative research has often concentrated on demographic and social antecedents (e.g. gender, prior education). Other factors may be more open to intervention e.g. students' academic experiences in HE. Using data from an institutional survey (N = 1170), logistic regression tested a range of academic experiences, regarding their relationship to contemplation of withdrawal ('COW': a recognised marker for actual withdrawal). COW was associated with student perceptions of low one-to-one contact with staff; non-traditional delivery methods; low peer-interaction; and high assessment load. Interestingly, COW was not associated with overall contact hours, large classes, or personal tutoring. The contributing factors explained 5.1%-8.6% of variance in COW, suggesting they may be meaningful levers for optimising retention. The paper discusses links to existing literature, future research directions, and applied implications for institutions.