The solar wind is a continual outflow of plasma and magnetic field from the Sun's upper atmosphere -- the corona -- that expands to fills the solar system. Variability in the near-Earth solar-wind conditions can produce adverse space weather that impacts ground- and space-based technologies. Consequently, numerical fluid models of the solar wind are used to forecast conditions a few days ahead. The inner-boundary conditions are supplied by observationally-constrained models of the corona. While solar eruptions -- coronal mass ejections (CMEs) -- are treated as time-dependent structures, a single coronal ''snapshot'' is typically used to determine the ambient solar-wind structure through which CMEs propagate. Thus, all available time-history information from previous coronal-model solutions is discarded and the solar wind is treated as a steady-state flow, unchanging in the rotating frame of the Sun. In this study, we use one year of daily-updated coronal-model solutions to comprehensively compare steady-state solar-wind modelling with a time-dependent method. We demonstrate, for the first time, how the SS approach can fundamentally misrepresent the accuracy of coronal models. We also attribute three key problems with current space-weather forecasting directly to the steady-state approach: (1) the seemingly paradoxical result that forecasts based on observations from 3-days previous are more accurate than forecasts based on the most recent observations; (2) high inconsistency, with forecasts for a given day jumping significantly as new observations become available and changing CME propagation times by up to 17 hours; and (3) insufficient variability in the heliospheric magnetic field, which controls solar energetic particle propagation to Earth. The time-dependent approach is shown to alleviate all three issues. It provides a consistent, physical solution which more accurately represents the information present in the coronal models. By incorporating the time history in the solar wind along the Sun-Earth line, the time-dependent approach will provide improvements to forecasting CME propagation to Earth.