Crossed-beam thermal lens spectrometry can be implemented using continuous-wave- (cw) laser or pulsed-laser excitation. In both cases, the signal depends on the position of the sample with respect to the probe beam waist, the size of the excitation beam, the beamsize ratio into the sample, and the power or energy of the excitation beam. However, due to differences in the rate of formation and relaxation of the thermal lens, both methods exhibit distinct key features. Optimization of the experimental setup and understanding the thermal lens signal are more complicated under cw-laser excitation than with pulsed-laser excitation. Unlike that observed under pulsed excitation, the effect of the excitation beam waist, of the sample size, and of the flow rate are closely related to the effective size of the thermal element and depend on the chopping frequency. Although the intrinsic sensitivities are almost the same, the performance can significantly differ depending on the chopping frequency or pulse repetition rate, which should be high enough to allow fast data collection and efficient signal averaging.