Understanding the causes and consequences of pollinator declines is a priority in ecological research. However, across much of the globe we have a poor understanding of pollinator assemblages, population trends and the ecological and economic importance of particular pollinators, due to a marked geographic bias in research effort. Here, we show that almost half the data cited in thirteen recent meta-analyses, which ask important and diverse questions in pollination ecology, were collected in just five countries: Australia, Brazil, Germany, Spain and the USA. In contrast, the entire continent of Africa contributed only 4% of the data. We believe that the consequences of this geographic bias are severe. Foremost, pollinator assemblages (and possibly their sensitivity to ecological drivers) can greatly vary among these regions. In addition, many communities that rely on pollinators, bees in particular, for food security and wealth generation are in geographic regions where our understanding of pollination is poor. Collecting accurate information on pollinator populations in data deficient areas will allow us to identify vulnerable populations and species and so better target conservation measures. Moreover, it will help us to determine if our current understanding of pollinator losses, based on data collected in a few locations and on the species that predominate in those regions, is representative of the wide diversity of ecosystems. We propose means of collecting such data given socioeconomic constraints.