Natural products have long played a leading role as direct source of drugs or as a means to inspire informed molecular design. Indeed, natural products have been biologically prevalidated as protein-binding motifs by millions of years of evolutionary pressure. Despite the tailored architectures, and the ever-growing chemistry toolbox to aid access such privileged structures, identifying the modes of action by which these molecules can be harnessed as therapeutics remains a major bottleneck in discovery chemistry. Herein, an overview of cheminformatics methods applied to the identification of modes of action of natural products is given, and a discussion of successful case studies is provided. A special focus is given to machine learning methods that may help to streamline the development of natural products into drug leads.