Despite decades of work on climate change biology, the scientific community remains uncertain about where and when most species distributions will respond to altered climates. A major barrier is the spatial mismatch between the size of organisms and the scale at which climate data are collected and modeled. Using a meta-analysis of published literature, we show that grid lengths in species distribution models are, on average, ca. 10 000-fold larger than the animals they study, and ca. 1000-fold larger than the plants they study. And the gap is even worse than these ratios indicate, as most work has focused on organisms that are significantly biased toward large size. This mismatch is problematic because organisms do not experience climate on coarse scales. Rather, they live in microclimates, which can be highly heterogeneous and strongly divergent from surrounding macroclimates. Bridging the spatial gap should be a high priority for research and will require gathering climate data at finer scales, developing better methods for downscaling environmental data to microclimates, and improving our statistical understanding of variation at finer scales. Interdisciplinary collaborations (including ecologists, engineers, climatologists, meteorologists, statisticians, and geographers) will be key to bridging the gap, and ultimately to providing scientifically grounded data and recommendations to conservation biologists and policy makers.