Several papers have been published about this subject, but I think should be interesting to open a discussion between the RG users.
In my opinion, the first step is to make transparent the algorithm used to calculate this parameter and to pay more attention to the scientific productivity evaluation. This should be differentiated according to the disciplinary sector. It is also important to appreciate both quantity and quality of the publications, particularly taking into account the research areas. In certain studies is possible to have spendable results only after several years.
RG score use principally the JIF system based on Thomson Reuters WoK database, but we know that to many journals are not ranked by TR nor some particular types of scientific contributions (conference papers, books, chapters, etc.). In these aspects RG could and should improve too much.