Virtual fencing is an emerging animal management technology that uses auditory cues followed by a mild electrical cue instead of physical barriers to contain animals. Virtual fencing has long been conceptualized as a tool to help land managers achieve livestock production and land management goals, yet little research has focused on factors that influence virtual fence performance. We evaluated the effects of stocking density, the quantity of forage inside the paddock, and differences between forage quantities inside and outside the paddock on the effectiveness of a commercially available virtual fencing system. We tested the virtual fencing system at stocking densities ranging from 2 to 8 animals ha^–1 and measured stubble height inside and outside the paddocks as a proxy for forage quantity. Predictability and controllability of the electrical cue have been identified as key animal welfare considerations associated with virtual fences, but have not been specifically defined, quantified, and analyzed. Therefore, we quantified predictability and controllability and evaluated the extent to which they were affected by stocking density, stubble height, and the difference in stubble heights inside and outside the paddock. We found that neither stocking density, forage quantity, nor the difference in forage quantity inside and outside the paddock influenced the effectiveness of virtual fences or the predictability and controllability of the electrical cue. This implies that virtual fencing systems are likely to be reliable tools for livestock management in productive settings and for stocking densities up to 8 animals ha^–1 even when stubble heights are at or below common management targets for riparian plant communities.