Earth's mantle undergoes convection on million-year timescales as heat is transferred from depth to the surface. Whilst this flow has long been linked to the large-scale horizontal forces that drive plate tectonics and supercontinent cycles, geologists are increasingly recognising the signature of convection through transient vertical motions in the rock record, known as "dynamic topography". A significant component of topography is supported by lithospheric isostasy, and changes in lithospheric thermal structure are sometimes included in the definition of dynamic topography. An additional component arises from active flow within the underlying convecting mantle, and this process causes dynamic topography that has lengthscales varying from 10,000 km down to 500 km and typical amplitudes of ±1 km. Transient uplift and subsidence events are often slow, but might evolve at rates as fast as 500 m/Myr over cycles as short as ~3 Myr, leading to periodic overwriting of the geological record that results in complex interpretational challenges. Despite these difficulties, a growing number of observational and computational studies have highlighted the important role of dynamic topography in fields as diverse as intraplate magmatism, sedimentary stratigraphy, landscape evolution, paleo-shorelines, oceanic circulation patterns, and ice sheet stability. This review provides a brief overview of our current understanding of the topic and explores some basic insights that can be gained from simple three-dimensional numerical simulations of mantle convection under different convective regimes. We summarise a suite of observational techniques used to estimate dynamic topography, and finish by laying out some key unanswered questions to stimulate debate and inspire future studies.