Biomarkers may be a key foundation for the precision medicine of the future. In this article, we review current knowledge regarding biomarkers in difficult-to-treat asthma and their ability to guide the use of both conventional asthma therapies and novel (targeted) therapies. Biomarkers (as measured by tests including prednisolone and cortisol assays and the fractional exhaled nitric oxide (NO) suppression test) show promise in the assessment and management of non-adherence to inhaled and oral corticosteroids. Multiple markers of type 2 inflammation have been developed, including eosinophils in sputum and blood, exhaled NO, serum IgE and periostin. Although these show potential in guiding the selection of novel interventions for refractory type 2 inflammation in asthma, and in determining if the desired response is being achieved, it is becoming clear that different biomarkers reflect distinct components of the complex type 2 inflammatory pathways. Less progress has been made in identifying biomarkers for use in difficult-to-treat asthma that is not associated with type 2 inflammation. The future is likely to see further biomarker discovery, direct measurements of individual cytokines rather than surrogates of their activity and the increasing use of biomarkers in combination. If the promise of biomarkers is to be fulfilled, they will need to provide useful information that aids clinical decision-making, rather than being 'just another test' for clinicians to order.