The Event Horizon Telescope (EHT), a global very long baseline interferometry array operating at millimetre wavelengths, is spatially resolving the immediate environments of black holes for the first time. The current observations of the Galactic center black hole, Sagittarius A* (Sgr A*), and M87 have been interpreted in terms of either geometric models (e.g., a symmetric Gaussian) or detailed calculations of the appearance of black hole accretion flows. The former are not physically motivated, while the latter are subject to large systematic uncertainties. Motivated by the dominant relativistic effects of Doppler beaming and gravitational lensing in many calculations, we propose a geometric crescent model for black hole images. We show that this simple model provides an excellent statistical description of the existing EHT data of Sgr A* and M87, superior to other geometric models for Sgr A*. It also qualitatively matches physically predicted models, bridging accretion theory and observation. Based on our results, we make predictions for the detectability of the black hole shadow, a signature of strong gravity, in future observations.