Hyper-herbivory following predator removal is a global issue. Across North America and Europe, increasing deer numbers are affecting biodiversity and human epidemiology, but effectiveness of deer management in heterogeneous landscapes remains poorly understood. In forest habitats in Europe, deer numbers are rarely assessed and management is mainly based on impacts. Even where managed areas achieve stable or improving impact levels, the extent to which they act as sinks or persist as sources exporting deer to the wider landscape remains unknown. We present a framework to quantify effectiveness of deer management at the landscape scale. Applied across 234 km2 of Eastern England, we assessed management of invasive Reeve's muntjac (Muntiacus reevesi) and native roe (Capreolus capreolus), measuring deer density (using thermal imaging distance transects 780 km/year), fertility, neonatal survival, and culling to quantify source-sink dynamics over 2008–2010. Despite management that removed 23–40% of the annual population, 1,287 (95% CI: 289–2,680) muntjac and 585 (454–1,533) roe deer dispersed annually into the wider landscape, consistent with their ongoing range expansion. For roe deer, culled individuals comprised fewer young deer than predicted by a Leslie matrix model assuming a closed population, consistent with age-dependent emigration. In this landscape, for roe and muntjac, an annual cull of at least 60% and 53%, respectively, is required to offset annual production. Failure to quantify deer numbers and productivity has allowed high density populations to persist as regional sources contributing to range expansion, despite deliberative management programs, and without recognition by managers who considered numbers and impacts to be stable. Reversing an unfavorable condition of woodland biodiversity requires appropriate culls across large contiguous areas, supported by knowledge of deer numbers and fertility. © 2013 The Wildlife Society.