The enormous therapeutic potential of haematopoietic stem cells may be realized if we acquire the ability to control their survival in vitro and their behaviour in vivo. While extrinsic approaches using drugs are unlikely to be developed for this purpose in the near future, altering the intrinsic pool of biological information of stem cells by somatic gene transfer is more likely to succeed. Using the first generations of retroviral gene transfer vectors, we are confronted with the first examples of both successful therapeutic interventions and severe adverse events. The latter are related to the incomplete precision of the existing technologies. Concerted safety and efficiency evaluation has been enforced to further improve the prospects of this field. This review summarizes the current state of the debate, proposing future research directions towards understanding the complex interplay of risk factors related to random transgene insertion, unphysiological transgene expression and additional contributory factors of the specific therapeutic setting.