1. Analyses of flower-visitor interaction networks allow application of community-level information to conservation problems, but management recommendations that ensue from such analyses are not well characterized. Results of modularity analyses, which detect groups of species (modules) that interact more with each other than with species outside their module, may be particularly applicable to management concerns.

2. We conducted modularity analyses of networks surrounding a rare endemic annual plant, Eriogonum visheri, at Badlands National Park, USA, in 2010 and 2011. Plant species visited were determined by pollen on insect bodies and by flower species upon which insects were captured. Roles within modules (network hub, module hub, connector and peripheral, in decreasing order of network structural importance) were determined for each species.

3. Relationships demonstrated by the modularity analysis, in concert with knowledge of pollen species carried by insects, allowed us to infer effects of two invasive species on E. visheri. Sharing a module increased risk of interspecific pollen transfer to E. visheri. Control of invasive Salsola tragus, which shared a module with E. visheri, is therefore a prudent management objective, but lack of control of invasive Melilotus officinalis, which occupied a different module, is unlikely to negatively affect pollination of E. visheri. Eriogonum pauciflorum may occupy a key position in this network, supporting insects from the E. visheri module when E. visheri is less abundant.

4. Year-to-year variation in species' roles suggests management decisions must be based on observations over several years. Information on pollen deposition on stigmas would greatly strengthen inferences made from the modularity analysis.

5. Synthesis and applications: Assessing the consequences of pollination, whether at the community or individual level, is inherently time-consuming. A trade-off exists: rather than an estimate of fitness effects, the network approach provides a broad understanding of the relationships among insect visitors and other plant species that may affect the focal rare plant. Knowledge of such relationships allows managers to detect, target and prioritize control of only the important subset of invasive species present and identify other species that may augment a rare species' population stability, such as E. pauciflorum in our study.