Exploring key demographic parameters for establishment and growth of shrub populations enables us to identify the life stages or plant organs that browsers need to consume to effectively influence shrub population dynamics. However, there is a lack of concrete knowledge on how browsing can efficiently meet this goal. We therefore tested the impact of different levels of browsing intensity on a key demographic parameter (survival of juveniles) of a major colonizing shrub species (broom, Cytisus scoparius) in order to control population growth, and designed a browsing management strategy focused on this target stage. Three browsing intensities, representing pertinent management practices, were simulated on juveniles (2 years old) in a broom population. Juvenile broom plants were either left untouched (control treatment; n ¼ 126) or had 50% or 90% of their total edible stem biomass removed for “light-intensity browsing” (n ¼ 127) and “high-intensity browsing” (n ¼ 77) treatments, respectively. Survival and fecundity data were collected over 6 years. Standard matrix modeling was used to analyze the impact of browsing on changes in population growth rate (l), and the results were 6.34 (control), 2.26(light browsing) and 0.85 (heavy browsing). Therefore, the natural expansion of broom populations may be slowed by light browsing or even reversed by heavy browsing (l < 1). This multi-year survey confirms that focusing browsing on juveniles is an efficient strategy for controlling broom dominance. Shrub control strategies should therefore target early-growth-stage populations and repeat the browsing strategy at the same intensity over several years to achieve cumulative effects.