The thin-layer chromatographic (TLC) retention of 35 model compounds has been investigated with ten screening mobile phases on six normal-phase and seven reversed-phase adsorbents. The retention factors formed two cubes with dimensions 35 × 10 × 6 and 35 × 10 × 7, respectively, which enabled three-way analysis by PARAFAC. A one-component PARAFAC model was optimum in both cases and two-component models performed worse. The one-component model explained 78.8% of the variance in normal-phase chromatography and 94.2% of the variance in reversed-phase chromatography. These results showed that the major variability of the retention factor (RF) can be modelled as the product of three factors related to the substance itself, the mobile phase, and the adsorbent. RF modelling was substantially better than using k or RM (rate mobility) values.