The optimal method of cardiac resynchronization therapy (CRT) optimization is as yet unknown. We sought to investigate the responses of optimization at rest and on exercise. This 2 stage study involved 59 patients (age 65 ± 10, 75% male), who had all recently undergone successful CRT implantation. In the first stage, the 6-month response was compared between 30 individuals who underwent resting echocardiographic optimization of CRT [atrioventricular (AV delay) plus ventriculo-ventricular delays (VV delay)], compared with the 29 who did not. In the second stage, a subset of 37 patients from the original cohort were randomized (double-blind) to either resting echocardiographic optimization (n = 20) or exercise echocardiographic optimization (n = 17) and followed for a further 6 months. Clinical and echocardiographic data were collected at each stage. Patients undergoing rest optimization demonstrated improvement in almost all variables and significantly in B-type natriuretic peptide (BNP) in contrast to those without optimization. In a linear regression model, the only significant predictor of BNP change was whether an individual underwent resting optimization or not (β = 0.38, P = 0.04). In those undergoing resting optimization, the degree of change in AV delay was correlated with improvement in left ventricle (LV) end-diastolic volume (r2 = 0.33, P < 0.01). Optimization on exercise was associated with a significant decrease in 6-minute walk test compared to those randomized to rest optimization possibly due to inducing nonoptimization at rest. In conclusion, echocardiographic optimization of CRT at rest is superior to no optimization or optimization on exercise. Patients with the greatest change in AV delay to reach optimal may undergo greater LV remodeling.