Granular aluminium is a high kinetic inductance thin film superconductor which, when formed into nanowires can undergo an intrinsic electromigration process. We use a combination of experimental and computational approaches to investigate the role of grain morphology and distribution in granular aluminium thin films, when formed into nanowire constrictions. Treating the granular aluminium film as a network of randomly distributed resistors with parameters motivated by the film microstructure allows us to model the electrical characteristics of the nanowires. This model provides estimates of the dependence of sheet resistance on grain size and distribution, and the resulting device to device variation for superconducting nanowires. By fabricating a series of different length nanowires, we study the electromigration process as a function of applied current, and then compare directly to the results of our computational model. In doing so we show that the electromigration is driven by the formation of quantum point contacts between metallic aluminium grains.