Dynamically scaled experiments and numerical analyses are performed to study the effects of the wake from an upstream wind turbine on the aerodynamics and performance of a downstream wind turbine. The experiments are carried out in the dynamically scaled wind turbine test facility at ETH Zurich. A five-hole steady-state probe is used to characterize the cross-sectional distribution of velocity at different locations downstream of the wake-generating turbine. The performance of the downstream wind turbine is measured with an in-line torquemeter. The velocity field in the wind turbine wake is found to differ significantly from the velocity field assumed in numerical wake models. The velocity at hub height does not increase monotonically up to the freestream velocity with downstream distance in the wake. Furthermore, the flowfield is found to vary significantly radially and azimuthally. The application of wake models that assume a constant axial velocity profile in the wake based on the measured hub-height velocity can lead to errors in annual energy production predictions of the order of 5% for typical wind farms. The application of wake models that assume an axisymmetric Gaussian velocity profile could lead to prediction errors of the order of 20%. Thus modeling wind turbine wakes more accurately, in particular by accounting for radial variations correctly, could increase the accuracy of annual energy production predictions by 5%–20%.