Recent years have witnessed tremendous progress in the development of new materials for solution-processed solar cells. The impressive pace of experimental R&D is gradually reshaping the role and mission of materials modelling in this area. For example ever more often theorists are asked to screen broad materials libraries in order to identify promising new compounds. In this talk I will try to provide a fair assessment of where we stand in the computational prediction of new absorbers. I will start by discussing a recent attempt at designing novel perovskites. Here we wanted to address the following simple question: can we design perovskites with a given optical gap? In order to answer this question we used a hybrid rational design approach combining extensive first-principles calculations with a simple mathematical model of the metal-halide network. Using this approach we found that the gap correlates with the largest metal-halide-metal bond angle. With this information at hand we performed an extensive set of first-principles calculations, and established that the bond angles can be controlled via the size of the organic cations. This study led us to identify several new cations which would allow tuning the gap from the infrared to the visible. In a second example I will describe our work on the design of semiconductor sensitizers with optimal work function. The starting point of our investigation was stibnite (antimony sulphide), a mineral of the tetradymite family which was successfully used as a dye-replacement in sensitized solar cells. Stibnite is accompanied by three isostructural compounds, obtained by replacing sulphur by selenium and/or antimony by bismuth. Using first-principles calculations we predicted that, while bismuthinite (bismuth sulphide) should not inject electrons into titania, antimonselite (antimony selenide) should be an efficient sensitizer. These predictions were recently confirmed by experiments. Finally I will briefly discuss an example of the pitfalls of rational design. In the context of dye-sensitized solar cells significant efforts are devoted to design new dyes with minimal loss-in-potential, by tuning the frontier orbitals of the dyes in isolation. Here, by combining electronic structure calculations of titania/dye interfaces with atomically-resolved STM measurements, we established that the dye adsorption geometry plays a crucial role in the energy-level alignment. This finding implies that the optimization of dyes in isolation is bound to be ineffective. This example highlights the limitations of rational design, and stresses the importance of understanding the fundamental mechanisms underlying the operation of solution-processed solar cells.