We investigate granular flows generated by collapse of an initially fluidized column into a horizontal channel in order to evaluate the factors controlling the efficiency of fluidization in increasing the run-out distance of pyroclastic density currents. This configuration is analogous to flows generated by collapse of volcanic domes or lava flow fronts. We use an incompressible two-phase numerical model able to simulate dam-break experiments, and we compare the numerical results with experimental data. This model permits us to describe depth-dependent variations of flow properties and the effect of pore pressure on the rheology of the granular material. We show that the interplay between timescales of column collapse and of flow front propagation plays a primary role in determining the effective influence of fluidization on run-out distance. For columns with high aspect ratio (i.e. initial height/initial width), the collapse velocity decreases abruptly after reaching its peak, a significant portion of the collapse has occurred when the flow front has travelled a long distance from the reservoir and, importantly, the decrease of basal pore pressure with time in the reservoir translates into a reduced velocity of the granular material entering into the propagation channel during final phases of collapse. Thus, at some point, the collapsing material is not able to significantly affect the flow front dynamics. This behaviour contrasts with that of low aspect ratio collapsing columns. These results are consistent with complementary analogue experiments of high aspect ratio collapsing columns, which show that the granular material at the front of the deposit originates from lower levels of the column. Comparison with new experimental data also reveals that the effective pore pressure diffusion coefficient in the propagating flow is an increasing function of column height, and it can be considered as proportional to a weighted average of flow thickness during propagation. This is consistent with experiments on static defluidization columns, but had not been tested in dam-break experiments until this study. Considering this type of dependency, under our experimental and simulation conditions, the non-dimensional run-out distance presents a relative maximum for initial aspect ratios between 1 and 2, and beyond this critical range the non-dimensional run-out distance decreases abruptly.