Dye-sensitized solar cells (DSSCs) are photoelectrochemical solar cells. Their function is based on photoinduced charge separation at a dye-sensitized interface between a nanocrystalline, mesoporous metal oxide electrode and a redox electrolyte. They have been the subject of substantial academic and commercial research over the last 20 years, motivated by their potential as a low-cost solar energy conversion technology. Substantial progress has been made in enhancing the efficiency, stability, and processability of this technology and, in particular, the interplay between these technology drivers. However, despite intense research efforts, our ability to identify predictive materials and structure/device function relationships and, thus, achieve the rational optimization of materials and device design, remains relatively limited. A key challenge in developing such predictive design tools is the chemical complexity of the device. DSSCs comprise distinct materials components, including metal oxide nanoparticles, a molecular sensitizer dye, and a redox electrolyte, all of which exhibit complex interactions with each other. In particular, the electrolyte alone is chemically complex, including not only a redox couple (almost always iodide/iodine) but also a range of additional additives found empirically to enhance device performance. These molecular solutes make up typically 20% of the electrolyte by volume. As with most molecular systems, they exhibit complex interactions with both themselves and the other device components (e.g., the sensitizer dye and the metal oxide). Moreover, these interactions can be modulated by solar irradiation and device operation. As such, understanding the function of these photoelectrochemical solar cells requires careful consideration of the chemical complexity and its impact upon device operation. In this Account, we focus on the process by which electrons injected into the nanocrystalline electrode are collected by the external electrical circuit in real devices under operating conditions. We first of all summarize device function, including the energetics and kinetics of the key processes, using an "idealized" description, which does not fully account for much of the chemical complexity of the system. We then go on to consider recent advances in our understanding of the impact of these complexities upon the efficiency of electron collection. These include "catalysis" of interfacial recombination losses by surface adsorption processes and the influence of device operating conditions upon the recombination rate constant and conduction band energy, both attributed to changes in the chemical composition of the interface. We go on to discuss appropriate methodologies for quantifying the efficiency of electron collection in devices under operation. Finally, we show that, by taking into account these advances in our understanding of the DSSC function, we are able to recreate the current/voltage curves of both efficient and degraded devices without any fitting parameters and, thus, gain significant insight into the determinants of DSSC performance.