Dilatometry is a useful technique to obtain experimental data concerning transformation kinetics in ferrous alloys. This technique is commonly used in cooling experiments to study the austenite decomposition of hypo-eutectoid steel grades. In the standard analysis of the dilatation signal there are two factors that are normally neglected. During the pro-eutectoid ferrite formation the austenite enriches in carbon, resulting in a non-linear temperature dependence of the specific austenitic volume. Furthermore, the specific volume of the formed ferrite is considerably different from that of the formed pearlite. In total not taking into account these two effects can lead to an error in the determined fraction ferrite of up to 25%. A method is presented that takes into account the two above-mentioned factors. In order to determine both the fraction ferrite and the fraction pearlite, in the analysis the temperature range of the transformation is divided into a ferrite-formation range and a pearlite-formation range. Two possible criteria for this division are discussed, and it is shown that the choice does not have an essential influence on the results.