There is inherent variation in the responses to drugs used in medicine. This can affect both efficacy and toxicity, which in some individuals leads to a poor benefit-risk ratio of drugs. Strategies to improve the benefit-risk ratio of drugs by choosing the right drug and the right dose for individual patients which leads to more predictable drug responses is an aim that has been espoused by many researchers, but unfortunately rarely achieved. Drug response by its vary nature is highly complex, and is determined by the nature of the disease (and its sub-phenotypes), genetic factors and environmental factors. It is important to take into account all of these factors to optimise drug therapy. There has been a lot of interest in genomic variation for many decades, and the recent advances in technologies for human genome wide analysis, including sequencing technologies, provides us with unprecedented capabilities to interrogate individual genomes and relate this to variation in drug responses. Most of this work is still focused on discovery, but this is important as it will uncover tractable variants which can then be taken further along the translational pathway to clinical implementation. Those discoveries which are further along this pathway have struggled to push into clinical practice for many reasons including the lack of clear guidance as to type of evidence that is needed, and poor study designs. Some of these concepts will be explored using key examples to draw upon strategies that have been adopted, where these have succeeded, and if not, why success has not occurred and what could be done to improve this.