This work revisits the idea of ground heat extraction in solar ponds by addressing various limitations associated with the conventional assumption of perfect heat transfer. For this, an overall heat transfer coefficient, U has been introduced to analyse the heat extraction taking place from the ground beneath the pond. The pond's thermal performance is analysed for different values of U using closed form solutions and significant departure is seen with respect to the conventional assumption of an infinite U. Temperature distributions across various zones are obtained which exhibit deviation from the ideal distributions. It is also observed that the optimum size of non-convective zone yielding maximum efficiency depends on the effectiveness of ground heat extraction. Further, calculations are made for the minimum salt diffusion rate required to sustain stable pond operation. It is observed that the idealized exchanger performance assumption underestimates this critical value and a design based on it may initiate instability. Finally, calculations are made for the net entropy production rate which is also seen to be under-predicted with the existing theory. The present work can therefore prove useful for making an accurate assessment of the performance and stability of solar ponds involving ground heat extraction.