A rich body of literature exists on deterrence, yet little is known about how deterrence messages are communicated through social networks. This is an important gap in our understanding, because such communication gives rise to the possibility that social institutions can utilize the vicarious effect of the threat of punishment against one individual to reduce the rate of reoffending amongst their criminal associates. To test this, we identified criminals with an extensive offending history (prolific offenders) and their co-offenders using social network analysis and then conducted a randomized controlled trial to measure the effect on both prolific offenders and their cooffenders of delivering a “specific deterrence” message. The treatment— preemptive engagements with prolific offenders by a police officer offering both ‘carrots’ (desistance pathways) and ‘sticks’ (increased sanction threat)—was applied to the prolific offenders, but not to their co-offenders. The outcomes suggest that a single officer–offender engagement leads to a crime suppression effect in all comparisons, with 21.3%, 11.0%, and 15.0% reductions for specific, vicarious, and total network deterrence effects, respectively. The findings suggest that (a) social network analysis based on in-house police records can be used to cartographically understand social networks of offenders, with an aim of preventing crime; (b) deterrence messages promulgated by the police have the capacity to reduce crime beyond what was previously assumed, as the cascading of threats in cooffending relationships carries a vicarious crime reduction impact; (c) unlike “reactive specific deterrence” (i.e., a threat of punishment following a specific and detected crime) which can have perverse effects on certain offenders, preventative specific deterrence is a promising crime policy.