The diffusion of technology artifacts is often marked by abrupt events and incremental evolutionary moves, resulting in shifts in diffusion parameters as well as the underlying mechanics. In this paper, we model the diffusion of Android and iOS based handsets, where new models and operating system versions are released periodically. We relax a common assumption in IT diffusion studies, of holding diffusion parameters constant, and find that there are clear breaks in their values at specific points in time. Using the system dynamics methodology, we then develop and calibrate a causal model of the underlying mechanics. Significant events during evolution of the two platforms are matched temporally with the observed breaks, and the changing mechanics of diffusion across the breakpoints are identified using this causal structure. We find that iOS and Android handset diffusion patterns, although superficially similar, were driven by different mechanics. Our study contributes to the IT diffusion literature by (i) establishing the need to test for, and model, shifts in diffusion parameters over the horizon of interest (ii) offering a method to identify changes in diffusion mechanisms accompanying these shifts and (iii) demonstrating that similar temporal diffusion patterns need not imply similar underlying mechanics.