Galaxy evolution is usually measured through probing scaling relations and distribution functions and determining if and how they evolve through cosmic time. While this method is effective for the characterization of the bulk of the galaxy population it does not reveal how individual galaxies may have formed, nor does it allow us to trace directly the formation processes in galaxies. Furthermore, because galaxies grow through mergers and star formation over time, selecting similar galaxies between epochs based on measurements of luminosity or even stellar or total mass is fraught with biases that makes any inferred evolution highly suspect. One approach for understanding this problem is to use abundance matching, whereby galaxies are compared between different epochs based on their relative number densities. If galaxies retain their rank ordering in some property, such as stellar mass or halo mass, then this would be an effective method for determining how different galaxy populations evolve through time. In this talk I will discuss using simulation results to quantify how well we can utilise the abundance matching technique, and within what limits the assumptions of a stable rank ordering of galaxy masses remains valid up to z=3. I will then discuss the application of this method using data from the GNS, UDS, and CANDELS surveys to show how we can effectively use abundance matching to determine how the processes of galaxy mergers, in-situ star formation and gas accretion from the intergalactic medium are driving the formation of galaxies at z < 3.