ALMA observations of line emission from planet forming discs have demonstrated to be an excellent tool to probe the internal disc kinematics, often revealing subtle effects related to important dynamical processes occurring in them, such as turbulence, or the presence of planets, that can be inferred from pressure bumps perturbing the gas motion, or from detection of the planetary wake. In particular, we have recently shown for the case of the massive disc in Elias 2-27 how one can use such kind of observations to measure deviations from Keplerianity induced by the disc self-gravity, thus constraining the total disc mass with good accuracy and independently on mass conversion factors between the tracer used and the total mass. Here, we refine our methodology and extend it to two additional sources, GM Aur and IM Lup, for which archival line observations are available for both the 12CO and the 13CO line. For IM Lup, we are able to obtain a consistent disc mass of Mdisc = 0.1 M⊙, implying a disc-star mass ratio of 0.1 (consistent with the observed spiral structure in the continuum emission) and a gas/dust ratio of ∼65 (consistent with standard assumptions), with a systematic uncertainty by a factor ≃ 2 due to the different methods to extract the rotation curve. For GM Aur, the two lines we use provide slightly inconsistent rotation curves, that cannot be attributed only to a difference in the height of the emitting layer, nor to a vertical temperature stratification. Our best fit disc mass measurement is Mdisc = 0.26 M⊙, implying a disc-star mass ratio of ∼0.35 and a gas/dust ratio of ∼130. Given the complex kinematics in the outer disc of GM Aur and its interaction with the infalling cloud, the CO lines might not well trace the rotation curve and our results for this source should then be considered with some caution.