Restoration and rehabilitation are globally implemented to improve ecosystem condition but often without tracking treatment expenditures relative to ecological outcomes. We evaluated the cost‐effectiveness of widely conducted woody plant and herbaceous invasive plant removals and seeding treatments in drylands of the western United States from 2004 to 2018 to determine how land managers can optimize efforts. Woody plant cover decreased at a similar rate per dollar spent regardless of vegetation removal type, and the dominant invasive species was reduced by herbicide application. Relatively inexpensive herbicide application also had a large positive effect on seeded perennial grass cover that was enhanced by additional cost; while expensive woody mastication treatments had little effect regardless of additional cost. High seed cost was driven by including a large proportion of native species in seed mixes, and combined with high seeding cost, promoted a short‐term (2–3 yr) gain in perennial forb cover and species richness. In contrast, seeding and seed mix cost had no bearing on seeded perennial grass cover, in part, because relatively cheap nonnative seeded species rapidly increased in cover. Our results suggest the differential benefits of commonly implemented treatments aimed at reducing wildfire risk, improving wildlife habitat and forage, and reducing erosion. Given the growing need and cost of restoration and rehabilitation, we raise the importance of specifying treatment budgets and objectives, coupled with effectiveness monitoring, to improve future outcomes.

Woodland encroachment is a global issue linked to diminished ecosystem services, prompting the need for restoration efforts. However, restoration outcomes can be highly variable, making it difficult to interpret the ecological benefits and risks associated with woodland-reduction treatments within semiarid ecosystems. We addressed this uncertainty by assessing the magnitude and direction of vegetation change over a 15-year period at 129 sagebrush (Artemisia spp.) sites following pinyon (Pinus spp.) and juniper (Juniperus spp.) (P–J) reduction. Pretreatment vegetation indicated strong negative relationships between P–J cover and the abundance of understory plants (i.e., perennial grass and sagebrush cover) in most situations and all three components differed significantly among planned treatment types. Thus, to avoid confounding pretreatment vegetation and treatment type, we quantified overall treatment effects and tested whether distinct response patterns would be present among three dominant plant community types that vary in edaphic properties and occur within distinct temperature/precipitation regimes using meta-analysis (effect size = lnRR = ln[posttreatment cover/pretreatment cover]). We also quantified how restoration seedings contributed to overall changes in key understory vegetation components. Meta-analyses indicated that while P–J reduction caused significant positive overall effects on all shrub and herbaceous components (including invasive cheatgrass [Bromus tectorum] and exotic annual forbs), responses were contingent on treatment- and plant community-type combinations. Restoration seedings also had strong positive effects on understory vegetation by augmenting changes in perennial grass and perennial forb components, which similarly varied by plant community type. Collectively, our results identified specific situations where broad-scale efforts to reverse woodland encroachment substantially met short-term management goals of restoring valuable ecosystem services and where P–J reduction disposed certain plant community types to ecological risks, such as increasing the probability of native species displacement and stimulating an annual grass-fire cycle. Resource managers should carefully weigh these benefits and risks and incorporate additional, appropriate treatments and/or conservation measures for the unique preconditions of a given plant community in order to minimize exotic species responses and/or enhance desirable outcomes.