Diagnosis and management of chronic kidney disease (CKD) will be characterized in the future by an increasing use of biomarkers—quantitative indicators of biologic or pathologic processes that vary continuously with progression of the process. “Classical” biomarkers of CKD progression include quantitative proteinuria, the percentage of sclerotic glomeruli or fractional interstitial fibrosis. New candidate biomarkers (e.g., urinary proteomic patterns) are being developed based on both mechanistic and “shotgun” approaches. Validation of potential biomarkers in prospective studies as surrogate endpoints for hard clinical outcomes is often complicated by the long lag time to the ultimate clinical outcome (e.g., end-stage renal disease). The very dense data sets that result from shotgun approaches on small numbers of patients carry a significant risk of model overfitting, leading to spurious associations. New analytic methods can help to decrease this risk. It is likely that clinical practice will come to depend increasingly on multiplex (vector) biomarkers used in conjunction with risk markers in early diagnosis as well as to guide therapy.