Using cationic dyes, such as hexaamminecobalt trichloride (CoHex), is a popular method for measuring clay minerals' cation exchange capacity (CEC), because a short single extraction followed by colorimetric measurements is sufficient to accurately measure the CEC. In addition, the analysis of the cations extracted provides additional confirmation of the CEC measurement. These cationic dyes are redox sensitive species and, surprisingly, the existing literature does not contain any information on possible biases to the method in relation to this redox reactivity. This study applies the CoHex method to three kinds of clay and clay mineral samples: (i) oxidized smectites (in their pristine state), (ii) chemically reduced smectites and (iii) clay-rock samples originating from various geological formations and preserved from atmospheric oxidation. Redox reactions occur between the Fe2+ in the clay mineral structures and the Co(III) from the cationic dye, leading to greatly overestimating the CEC values calculated from the colorimetric assays. Reduction of the cationic dyes also leads to a large release of ammonia in solution, originating from the coordination sphere of Co(III) in the CoHex. These redox reactions impede correct determination of the CEC value from the colorimetric method for samples containing clay minerals bearing reactive structural Fe2+. However, the CEC can be correctly determined from the analysis of extracted cations: i) if the soluble salts present in the solid sample can be quantified independently and ii) if the samples are not rich in ammonium