Personalized medicine for oestrogen receptor-α (ERα)-positive breast cancer requires predictive biomarkers for broad endocrine resistance as well as biomarkers capable of predicting resistance to a specific agent. In addition, biomarkers could be used to select patients that might benefit from the addition of treatments that do not target ERα. However, biomarker identification studies seem to be far from consistent and identified biomarkers seldom face an introduction into clinical practice. Importantly, most of the studies that seek to identify biomarkers have been performed using material from consecutive series of patients treated with tamoxifen (the most commonly prescribed ERα antagonist). Consequently, the predictive value of any biomarker identified is confounded by its prognostic value. Another important issue is the lack of differentiation between premenopausal and postmenopausal patients with breast cancer. The hormonal environment of a tumour in patients who are premenopausal is intrinsically distinct from those arising in postmenopausal women. Biomarkers of different biological mechanisms might enable the prediction of either broad endocrine resistance or resistance to a specific agent in each of these patient subtypes. Ultimately, improvements to study design are needed to establish the clinical validity of the most promising biomarkers to predict benefit from endocrine therapy.