Research suggests that although teachers’ targeted interventions can stop bullying, they still fail in about one-fourth of cases. Yet, most studies to date have not considered how targeted interventions were implemented, leaving open the possibility that improper implementation contributed to these failures. To address this gap, we examined the extent to which school personnel implementing the KiVa® antibullying program in Finland adhered to the program-recommended targeted intervention procedure when addressing bullying cases, and whether modifications to the procedure, influenced intervention effectiveness. We further tested the specific effects of two types of modifications – adaptations and omissions – on effectiveness. Data were collected using ecological momentary assessment, with school personnel documenting in a mobile application the steps they took when addressing bullying cases. The sample included 341 cases involving 396 victimized students (53% female, Mage = 12.39 SD = 2.08) and 733 bullying students (13% female, Mage = 12.52 SD = 1.96) from 22 primary and secondary schools. The results indicated that adherence to procedure varied considerably across intervention steps, and adherence to the full procedure was low. Interventions were, however, more effective when school personnel adhered to the procedure than when they made modifications. Moreover, interventions were least effective, when steps were omitted, whereas adaptations did not significantly reduce effectiveness compared to full adherence, though the trend was in the same direction as with omissions. These findings suggest that closer adherence to evidence-based procedures tends to lead to better outcomes in targeted bullying interventions.