This report describes the investigation of how the structure of the ligand influences the product distribution in a series of model Heck reactions. The results showed that the ligand had a strong influence on the selectivity in all cases. It also describes efforts to build prediction models through PLS modelling of experimental data against calculated descriptors for the ligands in silico. In addition, a subset of the data was used to investigate the ability of DFT calculations to predict the selectivity.