This study examined the relation between both implementation fidelity and quality and the outcomes of two different anti-bullying interventions targeting distinct processes involved in bullying: moral disengagement and social norms. In total, 34 French-speaking Belgian teachers from six elementary schools were trained to deliver either the moral disengagement or the social norms intervention to their Grade 4-6 students (N = 747, 50.4% boys) in a randomized controlled trial. Students reported their moral disengagement, perceived injunctive class norm toward bullying, and bullying behaviors in the fall and spring of the 2018-2019 school year. Teachers' implementation fidelity and quality were assessed through direct observation in each class by two independent raters with a satisfying interrater reliability. A multiplicative moderated mediation model using latent change scores revealed that both greater fidelity and quality amplified students' decrease in moral disengagement, which was accompanied by a decrease in bullying. Importantly, when quality was sufficient, fidelity did not matter. However, higher fidelity could compensate for a lack of quality. Contrary to expectations, there was no significant change in students' perceptions of the class injunctive norm, even under ideal implementation conditions. Alternative mediators should thus be considered. Because implementation by teachers inevitably fluctuates, anti-bullying program components should be evaluated under various implementation conditions to determine their cost-effectiveness ratio and to gain insights into how anti-bullying programs work.