In the last ten years electron cryotomography has made it possible to visualize large macromolecular assemblies inside intact cells in a near-native, "frozen-hydrated" state in 3-D to a few nanometers resolution. Increasingly, atomic models of individual proteins and smaller complexes obtained by X-ray crystallography, NMR spectroscopy, or other methods can be fit into cryotomograms to reveal how the various pieces work together inside cells. A few good pictures is therefore sometimes all that is really needed to distinguish between competing models. To illustrate these points, I will present several case studies from our recent work in bacterial cell biology, including new images and mechanistic insights into bacterial chemoreceptor arrays, secretion systems, and the Type IV pilus. Current limitations and future prospects will be described, including the possibility of correlated super-resolution fluorescence cryomicroscopy and electron cryotomography of FIB-milled eukaryotic samples.