Ultraviolet photoelectron spectroscopy (UPS) is the most widely used technique to determine the ionization energy (IE) of electronic materials, as this parameter is critically important for the energy level alignment in electronic and optoelectronic devices. For organic semiconductor IE assessment, molecules are typically evaporated and polymers spin-coated onto a conductive substrate, and then measured by UPS. For substrates that possess a constant work function over large area, the determination of IE from the measured UPS data is straight forward. However, if the substrate is heterogeneous (intentionally or unintentionally) in local work function, the conventional method to determine IE yields erroneous results and a more careful data evaluation is necessary. While the secondary electron cutoff (SECO) can exhibit area-averaged values, the valence levels are split in energy according to the local work function. Here, we demonstrate the possible pitfalls of heterogeneous substrates by employing well-controlled model systems, and show how appropriate data analysis can still yield correct IE values.