The Streaked Horned Lark (Eremophila alpestris strigata) is an endangered songbird endemic to prairie and open coastal habitats west of the Cascade Mountains. Its persistence is handicapped by dramatic habitat loss, human disturbance and predation. Fewer than 1,000 individuals remain in four isolated subpopulations across Washington and Oregon. By combining demographic data and information on size and distribution of local populations, I developed a stage- and space-structured demographic model to analyze E. a. strigata’s viability in Washington State. Importantly, simulations that include variation and correlation in survival rates yield variable, yet qualitatively consistent, forecasts of population growth. The model predicted a continuing statewide population decline and near certain risk of extinction over the next 25 years. To determine where conservation efforts and data collection are best focused, I performed a series of perturbation analyses in which the effects of changing vital rates were quantified. I found that population growth is most sensitive to the survival of adults. Under a modest scenario, increasing the survivorship of adults by 10% was sufficient to lengthen the median time to extinction by more than 5 years. A more optimistic improvement (20%) yielded a viable Coastal subpopulation. In contrast, I found that some well-meaning management activities, in particular efforts solely targeting fecundity, are unlikely to be either cost effective or biologically sound. Although many anthropogenic impacts threaten E. a. strigata, the subspecies’ future could be bright, provided that research and management focus on biologically significant aspects of its life history. However, I demonstrate that complacency is ill-advised; decisive action to quickly improve demographic rates is needed, given the consistent qualitative output across models and the inherent uncertainty in predictions of future population trends.