Wind tunnel experiments were conducted to understand the effect of building array size (N), aspect ratio (AR), and the spacing between building (WS) on the mean structure and decay of their wakes. Arrays of size 3×3, 4×4,and 5×5, AR = 4, 6, and 8, and WS = 0.5WB, 1WB, 2WB and 4WB (where WB is the building width) were considered. Three different wake regimes behind the clusters were identified: near-, transition-, and far-wake regimes. The results suggest that the spatial extent of these wake regimes is governed by the overall array width (WA). The effects of individual buildings are observed to be dominant in the near-wake regime (0 < x/WA < 0.5) where individual wakes appear behind each building. These wakes are observed to merge together in the ‘transition wake region’ (0.5 < x/WA < 1.5), forming a combined wake in which the individual contributions are no longer apparent. In the far-wake regime (x/WA > 1.5), the wake characteristics and scales of building clusters wakes are similar to those downstream of a single isolated building. Accordingly, new local and global scaling parameters in the near- and far-wake regimes are introduced. The decay of the centreline velocity deficit is then modelled as a function of the three parameters considered in the experiment.