The last several years have seen major progress towards the goal of translating our growing understanding of the molecular basis of cancer into drugs with improved therapeutic activity and selectivity. Tremendous advances have been made but significant obstacles remain. In this review we assess our experience in the design and development of signal transduction drugs for cancer treatment, with a specific focus on small molecule kinase inhibitors. The druggability of cancer kinome targets is exemplified by imatinib, gefitinib, erlotinib and many other emerging agents. We assess the current status of the design of potent and selective kinase inhibitors, which has benefited greatly from high throughput screening and structure-based approaches. A diverse range of kinase inhibitory scaffolds is now available based on these methods. Multi-parameter optimisation now focuses as much on pharmacokinetic and metabolic properties as it does on target potency and selectivity. Development of a molecular audit trail requiring assays to demonstrate mechanism of action in vitro and in vivo is essential. Current issues include our relatively poor ability to predict the level of kinase selectivity in the intact cell, uncertainties around the most desirable selectivity profile, and the emergence of drug resistance.