Global movement patterns of migratory birds illustrate their fascinating physical and physiological abilities to cross continents and oceans. During their voyages, most birds land multiple times to make so‐called ‘stopovers’. Our current knowledge on the functions of stopover is mainly based on the proximate study of departure decisions. However, such studies are insufficient to gauge fully the ecological and evolutionary functions of stopover. If we study how a focal trait, e.g. changes in energy stores, affects the decision to depart from a stopover without considering the trait(s) that actually caused the bird to land, e.g. unfavourable environmental conditions for flight, we misinterpret the function of the stopover. It is thus important to realise and acknowledge that stopovers have many different functions, and that not every migrant has the same (set of) reasons to stop‐over. Additionally, we may obtain contradictory results because the significance of different traits to a migrant is context dependent. For instance, late spring migrants may be more prone to risk‐taking and depart from a stopover with lower energy stores than early spring migrants. Thus, we neglect that departure decisions are subject to selection to minimise immediate (mortality risk) and/or delayed (low future reproductive output) fitness costs. To alleviate these issues, we first define stopover as an interruption of migratory endurance flight to minimise immediate and/or delayed fitness costs. Second, we review all probable functions of stopover, which include accumulating energy, various forms of physiological recovery and avoiding adverse environmental conditions for flight, and list potential other functions that are less well studied, such as minimising predation, recovery from physical exhaustion and spatiotemporal adjustments to migration. Third, derived from these aspects, we argue for a paradigm shift in stopover ecology research. This includes focusing on why an individual interrupts its migratory flight, which is more likely to identify the individual‐specific function(s) of the stopover correctly than departure‐decision studies. Moreover, we highlight that the selective forces acting on stopover decisions are context dependent and are expected to differ between, e.g. K−/r‐selected species, the sexes and migration strategies. For example, all else being equal, r‐selected species (low survival rate, high reproductive rate) should have a stronger urge to continue the migratory endurance flight or resume migration from a stopover because the potential increase in immediate fitness costs suffered from a flight is offset by the expected higher reproductive success in the subsequent breeding season. Finally, we propose to focus less on proximate mechanisms controlling landing and departure decisions, and more on ultimate mechanisms to identify the selective forces shaping stopover decisions. Our ideas are not limited to birds but can be applied to any migratory species. Our revised definition of stopover and the proposed paradigm shift has the potential to stimulate a fruitful discussion towards a better evolutionary ecological understanding of the functions of stopover. Furthermore, identifying the functions of stopover will support targeted measures to conserve and restore the functionality of stopover sites threatened by anthropogenic environmental changes. This is especially important for long‐distance migrants, which currently are in alarming decline.