1. Providing non-crop flowering resources in agricultural landscapes is widely promoted as a strategy to support arthropods that deliver pollination and pest control services. However, management options have largely been developed separately for pollinators and natural enemies, whereas possible effects on insect herbivores, such as crop pests, have often been overlooked. A first critical step for design and implementation of multifunctional plantings that promote beneficial arthropods while controlling insect pests is to identify suitable plant species to use. 2. We aimed to identify California native plants and, more generally, plant traits suitable for the coordinated management of pollinators (wild bees and honey bees), insect herbivores and arthropod natural enemies (predators and parasitic wasps). We established 43 plant species in a common garden experiment and sampled arthropods by weekly netting (wild bees), observations (honey bees) or vacuum sampling (insect herbivores, arthropod predators and parasitic wasps) during peak bloom of each plant species over two years. 3. Plant species differed in attractiveness for each arthropod functional group. Floral area of the focal plant species positively affected honey bee, predator and parasitic wasp attractiveness. Later bloom period was associated with lower numbers of parasitic wasps. Flower type (actinomorphic, composite or zygomorphic) predicted attractiveness for honey bees, which preferred actinomorphic over composite flowers and for parasitic wasps, which preferred composite flowers over actinomorphic flowers. 4. Across plant species, herbivore, predator and parasitic wasp abundances were positively correlated, and honey bee abundance correlated negatively to herbivore abundance. 5. Synthesis and applications. We use data from our common garden experiment to inform evidence-based selection of plants that support pollinators and natural enemies without enhancing potential pests. We recommend selecting plant species with a high floral area per ground area unit, as this metric predicts the abundances of several groups of beneficial arthropods. Multiple correlations between functionally important arthropod groups across plant species stress the importance of a multifunctional approach to arthropod habitat management.