Despite the costs of killing civilians in intrastate armed conflicts, in terms of effects on domestic support and international condemnation, civilians are often the direct targets of violence by both governments and rebel groups. In this paper, I propose that one way of understanding this violence is by looking at the bargaining process of war, and how targeting of civilians may sometimes be used in addition to the fighting on the battlefield for strategic purposes. Conflict actors target civilians in order to improve their bargaining position, and there are two aspects of this argument. First, by targeting civilians that are believed to be potential supporters of the other side, conflict actors try to secure territorial control, which in turn gives them bargaining leverage. Second, actors target civilians to signal their resolve and commitment to continued conflict, regardless of their capabilities of fighting. Five factors that affect conflict actors’ propensity for targeting civilians are suggested and tested using new data on direct and deliberate killings of civilians by all actors involved in an intrastate armed conflicts from 1992 to 2004. The results show that battlefield intensity is a strong predictor of violence, but at the same time actors are less likely to target civilians when the other party is already employing such a strategy. Moreover, governments tend to be more likely to target civilians when there is an external audience in the form of a third party present in the conflict.