Introduction: It has long been recognised that tumours are composed of a mosaic of cells and numerous methods have been developed to detect tumour heterogeneity, including in situ hybridisation, multi-regional sampling, cytological assays, and whole genome and single cell sequencing. Using these methods, heterogeneity has been observed at the genetic, epigenetic, and phenotypic level in numerous cancers. With the advent of deep sequencing technology, we now appreciate a greater complexity of distinct genotypes and phenotypes that drive the biological behaviour of cancer. Despite decades of progress in detecting tumour heterogeneity, the question remains: to what extent does it matter?Areas covered: This review explores the evidence for and against the importance of tumour heterogeneity in three main areas: prognostication, development of targeted therapeutics and tumour resistance; summarising current understanding before evaluating ongoing experimental and clinical developments.Expert opinion: Theoretical understanding and in vitro detection of intratumour heterogeneity promises much but is yet to translate into meaningful clinical benefit. However, the recent emergence of a host of technological innovations and upcoming clinical trials may soon change the landscape of this field.