Modeling the correct electronic and geometrical configurations of polarons is a challenge with common electronic structure methods, such as density functional theory (DFT). There is no guarantee that the 'correct' wavefunction with a polaron will occur when using DFT. Delocalized, non-polaronic solutions may readily occur. The problem is even more difficult for semiconductors that may have several polaronic states that are nearly degenerate, or that have polaronic states not limited to single atomic sites. Methods such as DFT+U and hybrid functionals often provide a better description of polarons. Still, even these methods may not lead to desired polaronic states, and many hours of wasted simulations may occur. Strategies for controlling and enabling polaron formation exist, such as starting with initial geometries or wavefunctions that mimic polarons. Results using these strategies are compared for several common semiconductors: TiO2, HfO2, BiVO4. This work shows that by carefully modification of input geometries, faster formation of polarons will occur during simulation and that formation of polarons at specific sites can be controlled. These results will aid those modeling polarons, leading to more systematic simulation of polarons, and also pave the way for automation of polaron modeling.