The continuously increasing power of modern supercomputers renders the application of more and more accurate parameterfree models to systems of increasing complexity feasible. Consequently, it becomes possible to even treat different realstructure effects such as alloying or n-doping in systems like the technologically important transparent conducting oxides. In this paper we outline how we previously used a combination of quasiparticle calculations and a cluster expansion scheme to calculate the fundamental band gap of Mg xZn 1-xO and Cd xZn 1-xO alloys. We discuss the results in comparison to values for In 2O 3, SnO 2, SnO, and SiO 2. In addition, we discuss our extension of the Bethe-Salpeter approach that has been used to study the interplay of excitonic effects and doping in n-type ZnO. The dependence of the Burstein-Moss shift on the free-carrier concentration is analyzed.