Perovskite photovoltaics is one of the fastest growing research fields in materials science. Apart from photovoltaics, a wide range of energy-related research fields have gained substan-tial new momentum owing to the fantastic optoelectronic properties of hybrid lead halide per-ovskites. Starting with solar cells showing a mere 3% power conversion efficiency in 2009, researchers have been able to increase this value to over 22% in 2017, comparable to the best monocrystalline silicon solar cells. These improvements are chiefly due to the compositional tuning and mixing of the ABX3 perovskite structure. However, this strategy increases not only the efficiency but also in the same time the complexity of the perovskite formulation. Differ-ent phases, from the nano- to macroscale, can potentially form. In order to rationalize the effi-ciency progress, one needs to understand and characterize the complex perovskite composi-tion and crystal structure in great detail. The main focus of this thesis lies in the in-depth compositional and phase analysis of perovskite thin films as well as full perovskite solar cells, trying to rationalize the consequences of A cation and X anion tuning and mixing. In the first part, it is shown that fabricating solar cells from a non-stoichiometric perovskite precursor composition, namely one with PbI2 excess, leads to increased grain size, enhanced cristallinity, reduced recombination as well as an improved TiO2/perovskite interface. As op-posed to lead iodide phases resulting from film degradation, which is hampering device per-formance, unreacted excess PbI2 phases originating from an excess of PbI2 in the precursor solution are very beneficial for the final solar cell efficiency, reproducibility and stability. The next scientific challenge adressed in this thesis is related to the nano- and microscale composition of the record-breaking mixed cation/mixed anion perovskite composition. Mac-roscale investigations have previously suggested that this formulation results in a single ho-mogenous phase. However, using nanoscale elemental and charge carrier distribution mapping as well as microscale structural and optical film analysis, it is here found that in high efficient solar cells partial phase segregation does take place at the micro- and nanoscale. Moreover, it is shown that mixed cation/anion formulations allow the formation of never re-ported 3D hexagonal lead halide perovskite polytypes, namely 4H and 6H. These polytypes are shown to play a key role during the crystallization process, which is also revealed for the first time here. Indeed, mixed cation/mixed anion perovskite films are fabricated via the crys-tallization sequence 2H4H6H3R(3C). It is demonstrated that the complex crystalliza-tion via these defect-prone hexagonal intermediates can be by-passed by the incorporaton of low amounts of Cs+ cations into the structure. This can explain the improved stability as well as the increased power conversion efficiency and device reproducibility. It gives for the first time a rational explanation as to why the halide perovskite community rapidly adopted the incorporation of Cs+ in mixed perovskites. The last results presented in this thesis are related to the hole-transporting materials (HTM) used in perovskite solar cells. It is shown that carbazole-based small molecule HTMs are a cheap and efficient alternative to the much costlier commercial spiro-OMeTAD.