Ever since aerodynamics became an essential element of the automobile design process, the principal development tool for the vehicle aerodynamicist has been the full-scale wind tunnel. In the absence of a reliable alternative, it is expected that this will continue for many years.As a true simulation of the conditions on the road the conventional full-scale wind tunnel has limitations. The ground is fixed relative to the vehicle allowing an unrepresentative boundary layer to develop, the wheels of the test vehicle do not rotate and there is some uncertainty over the influences imposed by the tunnel walls. In addition, the aerodynamic data obtained from different wind tunnels shows a degree of scatter and even configuration changes do not necessarily produce consistent effects.With particular regard for aerodynamic drag, the aerodynamicist should ensure that gains obtained in the wind tunnel generate real benefits on the road. This paper describes work undertaken to enhance the understanding of differences between road and wind tunnel derived data. To assist this process, an advanced coastdown technique has been employed to determine aerodynamic drag on the road. In contrast with previous investigations, the principal vehicle tested was systematically reconfigured to provide a range of drag and lift characteristics. Saloon and fastback vehicle shapes have been tested with a range of front and rear spoilers. In addition, the influence of underfloor smoothing and the effects of wheel design and tyre size were studied.Comparison of track results with data derived from one wind tunnel shows that the wind tunnel under-predicts drag and that differences are configuration dependent. In certain cases a relationship with lift is established.