Although the importance of cognitive feedback in digital serious games (DSG) is undisputed, we are facing some major design challenges. First of all, we do not know to which extend existing research guidelines apply when we stand the risk of cognitive feedback distorting the delicate balance between learning and playing. Unobtrusive cognitive feedback has to be interspersed with gameplay. Second, many effective solutions for providing cognitive feedback we do know might simply be too costly. To face both challenges, this study offers an efficient approach for providing unobtrusive and retrospective cognitive feedback (RCF) in DSG. This approach was applied onto a game where feedback messages were triggered via simple rules about learners' questioning behavior on four dimensions. We found the experimental condition including such RCF to yield better learning outcomes while maintaining similar motivation.