In this paper, we compared the methods of photothermal spectroscopy used in different spatial scales, namely thermal-lens spectrometry (TLS) and thermal-lens microscopy (TLM) to enhance the performance parameters in analytical procedures. All of the experimental results were confirmed by theoretical calculation. It was proven that the design for both TLM and TLS, despite a different scale for the effect, is governed by the same signal-generating and probing conditions (probe beam diameter at the sample should be equal to the diameter of the blooming thermal lens), and almost does not depend on the nature of the solvent. Theoretical and experimental instrumental error curves for thermal lensing were coincident. TLM obeys the same law of instrumental error as TLS and shows better repeatability for the same levels of thermal-lens signals or absorbances. TLS is more advantageous for studying low concentrations in bulk, while TLM shows much lower absolute LODs due to better repeatability for low amounts. The behavior of the thermal-lens signal with different flow rates was studied and optimum conditions, with the minimum contribution to total error, were found. These conditions are reproducible, are in agreement with the existing theory of the thermal response in thermal lensing, and do not significantly affect the design of the optimum scheme for setups. TLM showed low LODs in solvent extraction (down to 10(-8) M) and electrokinetic separation (10(-7) M), which were shown to be governed by discussed instrumental regularities, instead of by microchemistry.