In recent years, optical nanothermometers have seen huge improvements in terms of precision as well as versatility, and several research efforts have been directed at adapting novel active materials or further optimizing the temperature sensitivity. The signal-to-noise ratio of the emission lines is commonly seen as the only limitation regarding high precision measurements. The role of re-absorption caused by a population of lower energy levels, however, has so far been neglected as a potential bottleneck for both high resolution and material selection. In this work, we conduct a study of the time dependent evolution of population densities in different luminescence nanothermometer classes under the commonly used pulsed excitation scheme. It is shown that the population of lower energy levels varies when the pump source fluctuates in terms of power and pulse duration. This leads to a significant degradation in temperature resolution, with limiting values of 0.5 K for common systems. Our study on the error margin indicates that either short pulsed or continuous excitation should be preferred for high precision measurements. Additionally, we derive conversion factors, enabling the re-calibration of currently available intensity ratio measurements to the steady state regime, thus facilitating the transition from pulse regimes to continuous excitation.