Understanding of the role of environmental change in the decline of endangered species is critical to designing scale-appropriate restoration plans. For locally endemic rare plants on the brink of extinction, frugivory can drastically reduce local recruitment by dispersing seeds away from geographically isolated populations. Dispersal of seeds away from isolated populations can ultimately lead to population decline. For localized endemic plants, fine-scale changes in microhabitat can further limit population persistence. Evaluating the individual and combined impact of frugivores and microhabitat heterogeneity on the short-term (i.e. transient) and long-term (i.e. asymptotic) dynamics of plants will provide insight into the drivers of species rarity. In this study, we used four years of demographic data to develop matrix projection models for a long-lived shrub, Cyrtandra dentata (H. St. John & Storey) (Gesneriaceae), which is endemic to the island of O'ahu in Hawai'i. Furthermore, we evaluated the individual and combined influence of a non-native frugivorous bird, Leiothrix lutea, and microhabitat heterogeneity on the short-term and long-term C. dentata population dynamics. Frugivory by L. lutea decreased the short-term and long-term population growth rates. However, under the current level of frugivory at the field site the C. dentata population was projected to persist over time. Conversely, the removal of optimum microhabitat for seedling establishment (i.e. rocky gulch walls and boulders in the gulch bottom) reduced the short-term and long-term population growth rates from growing to declining. Survival of mature C. dentata plants had the greatest influence on long-term population dynamics, followed by the growth of seedlings and immature plants. The importance of mature plant survival was even greater when we simulated the combined effect of frugivory and the loss of optimal microhabitat, relative to population dynamics based on field conditions. In the short-term (10 years), however, earlier life stages had the greatest influence on population growth rate. Synthesis and applications. This study emphasizes how important it is to decouple rare plant management strategies in the short versus long-term in order to prioritize restoration actions, particularly when faced with multiple stressors not all of which can be feasibly managed. From an applied conservation perspective, our findings also illustrate that the life stage that, if improved by management, would have the greatest influence on population dynamics is dependent on the timeframe of interest and initial conditions of the population.