Finding an (optimal) solution to a configuration problem has exponential complexity in the worst case. Nevertheless, most practical cases can be solved sufficiently fast using techniques such as divide-and-conquer (workflows, decomposition), complexity reductions (relaxations), symmetry breaking, heuristics, and solver parametrization. In this chapter, we illustrate the exponential complexity of configuration problems to motivate the need for performance tuning, present some ideas to judge the relevant aspects of complexity, and explain in detail some important techniques to improve solving performance.