During the past decade, a stem-cell-like subset of cancer cells has been identified in many malignancies. These cells, referred to as cancer stem cells (CSCs), are of particular interest because they are believed to be the clonogenic core of the tumour and therefore represent the cell population that drives growth and progression. Many efforts have been made to design therapies that specifically target the CSC population, since this was predicted to be the crucial population to eliminate. However, recent insights have complicated the initial elegant model, by showing a dominant role for the tumour microenvironment in determining CSC characteristics within a malignancy. This is particularly important since dedifferentiation of non-tumorigenic tumour cells towards CSCs can occur, and therefore the CSC population in a neoplasm is expected to vary over time. Moreover, evidence suggests that not all tumours are driven by rare CSCs, but might instead contain a large population of tumorigenic cells. Even though these results suggest that specific targeting of the CSC population might not be a useful therapeutic strategy, research into the hierarchical cellular organisation of malignancies has provided many important new insights in the biology of tumours. In this Personal View, we highlight how the CSC concept is developing and influences our thinking on future treatment for solid tumours, and recommend ways to design clinical trials to assess drugs that target malignant disease in a rational fashion.