The cancer stem cell hypothesis is becoming more widely accepted as a model for carcinogenesis. Tumours are heterogeneous both at the molecular and cellular level, containing a small population of cells that possess highly tumourigenic "stem-cell" properties. Cancer stem cells (CSCs), or tumour-initiating cells, have the ability to self-renew, generate xenografts reminiscent of the primary tumour that they were derived from, and are chemoresistant. The characterisation of the CSC population within a tumour that drives its growth could provide novel target therapeutics against these cells specifically, eradicating the cancer completely. There have been several reports describing the isolation of putative cancer stem cell populations in several cancers; however, no defined set of markers has been identified that conclusively characterises "stem-like" cancer cells. This paper highlights the current experimental approaches that have been used in the field and discusses their limitations, with specific emphasis on the identification and characterisation of the CSC population in epithelial ovarian cancer.