The quantitative determination of substance concentration in the stratum corneum (SC) is important for profiling of the skin barrier function-related parameters and for investigating drug delivery through the SC. Hereby, confocal Raman microscopy (CRM) is a widely used method, which can be applied noninvasively and in vivo. Most analysis methods are based on the assumption that keratin is distributed homogeneously throughout the SC. In this study, the depth profiles of keratin-related Raman peaks (1,003, 1,450, 1,650, and 2,935 cm −1 ) are investigated in detail. It was shown that each of these Raman peaks can be used for the precise determination of the skin surface position, which is important in case of overlapping of these with Raman bands associated to a formulation for treated skin. For correction of the depth-dependent signal attenuation, that is, normalization, the 1,450, 1,650, and 2,935 cm −1 peaks can be used, whereas the 1,003 cm −1 peak cannot. Further analysis shows that the keratin concentration is nonhomogeneous throughout the SC and decreases towards its bottom due to water increase. Thus, the concentration of certain substances in the SC, such as urea, lipids, water, or natural moisturizing factor, determined using normalization on the keratin peak intensity, is overestimated in the deep SC layers starting from 30% SC depth. The correction coefficients were calculated for all four Raman peaks and for all SC depths with 10% increments. It was determined that at the bottom of the SC, the values obtained using keratin normalization should be multiplied by 0.88 (for 1,003 or 1,450 cm −1 ), 0.94 (for 1,650 cm −1 ), and 0.92 (for 2,935 cm −1 ) in order to compensate nonhomogeneous distribution of keratin in the SC. Thus, original values are overestimated at the SC bottom by 6-12%. The presented correction mechanism is proposed to be used for further correction.