In this paper we review some of the considerations and potential sources of error when conducting Incident Photon to Current Efficiency (IPCE) measurements, with focus on photoelectrochemical (PEC) cells for water splitting. The PEC aspect introduces challenges for accurate measurements often not encountered in dry PV cells. These can include slow charge transfer dynamics and, depending on conditions (such as a white light bias, which is important for samples with non-linear response to light intensity), possible composition changes, mostly at the surface, that a sample may gradually undergo as a result of chemical interactions with the aqueous electrolyte. These can introduce often-overlooked dependencies related to the timing of the measurement, such as a slower measurement requirement in the case of slow charge transfer dynamics, to accurately capture the steady-state response of the system. Fluctuations of the probe beam can be particularly acute when a Xe lamp with monochromator is used, and longer scanning times also allow for appreciable changes in the sample environment, especially when the sample is under realistically strong white light bias. The IPCE measurement system and procedure need to be capable of providing accurate measurements under specific conditions, according to sample and operating requirements. To illustrate these issues, complications, and solution options, we present example measurements of hematite photoanodes, leading to the use of a motorized rotating mirror stage to solve the inherent fluctuation and drift-related problems. For an example of potential pitfalls in IPCE measurements of metastable samples, we present measurements of BiVO4 photoanodes, which had changing IPCE spectral shapes under white-light bias.