Group contribution, an important approach for estimating reaction equilibrium constants, is continually improving in accuracy and scope, but still has persistent issues that limit the accuracy of predictions. The quality and coverage of the thermodynamic data used in the method to estimate thermodynamic properties create uncertainty in estimates that particularly impact certain metabolic subsystems such as lipid metabolism. Accounting for reaction-specific uncertainty in Keq estimations explicitly is key to the effective use of Keqs in thermodynamic analysis. The group contribution approach itself has methodological limitations; most notably an inability to represent the nuanced differences of the same chemical groups in their context within different metabolites. Promising developments of thermodynamic parameter estimation methods may focus on the incorporation of new data types and additional data curation, algorithmic improvements, addition of molecular properties, and more sophisticated molecular modeling.