Cyber abuse can be executed directly (e.g. by sending derogatory emails or text messages addressed to the victim) or indirectly (e.g. by posting derogatory, private or false information, documents, images or videos about the victim online). This exploratory, mixed-method triangulated study examines cyber abuse crime events with the goal of identifying factors associated with the increased risk of personal victimization from both direct and indirect methods of cyber abuse. First, in-depth qualitative interviews with cyber abuse victims (n = 12) were conducted. The interviews were analysed using thematic analysis to generate hypotheses. These hypotheses were then tested using content analysis of newspaper reports (n = 110) and victims’ posts on online forums (n = 91) describing incidents of cyber abuse. Logistic regression using Bayesian Model Averaging analysis revealed that the combination of a prior offender–victim relationship and expressive motivation best predicts the use of indirect methods of cyber abuse, while direct methods of cyber abuse are more likely to occur when the offender does not know the victim and is motivated by instrumental ends. Implications for crime prevention are also discussed.