Feedback is generally considered a beneficial learning tool, and providing feedback is a recommended instructional practice. However, there are a variety of feedback types with little guidance on how to choose the most effective one. We examined individual differences in working memory capacity as a potential moderator of feedback type. Second- and third-grade children (N = 64) solved unfamiliar math problems prior to receiving instruction. Children received verification feedback on their answers (outcome-feedback) or on their strategies (strategy-feedback). Working memory capacity moderated the effect of feedback type on procedural transfer—the ability to solve novel problems. Children with lower working memory capacity benefitted less from strategy-feedback than outcome-feedback, whereas children with higher working memory capacity benefitted similarly from the two types of feedback. Results suggest the need to consider the cognitive demands of different feedback types. Problem solving can be optimized by considering both characteristics of the learner and the learning environment.