Cardiac resynchronisation therapy (CRT) can reduce symptoms, hospitalisations, and mortality in patients with severe left ventricular (LV) systolic dysfunction and electro-mechanical dyssynchrony. Unfortunately, approximately 30 % of eligible patients fail to respond to CRT. This study prospectively compared electrocardiography (ECG), echocardiography, and cardiac magnetic resonance (CMR) imaging for the prediction of response to CRT. We performed ECG, echocardiography and CMR on 46 patients prior to planned CRT implantation. Patients were divided into predicted responder and non-responder groups using previously described criteria for each modality. Changes in indicators of CRT response were recorded 6 months post-implantation, and later for transplant-free survival. Less dyspnoea, lower levels of N-terminal pro-brain natriuretic peptide, more LV reverse remodelling, and longer transplant-free survival were observed in predicted responders compared to predicted non-responders using each of the three modalities (p < 0.05 for each comparison). Additionally, for patients with QRS duration <150 ms and/or non-left bundle branch block (non-LBBB) QRS morphology, CMR predicted both clinical response and improved longer term transplant-free survival (80 % transplant-free survival in predicted responders vs. 20 % in predicted non-responders, p = 0.04). ECG and cardiac imaging techniques predict improvements in markers of response following CRT with similar accuracy. However, for CRT candidates with shorter, non-LBBB QRS complexes, a subgroup known to derive less benefit from CRT, CMR may predict those who are more likely to gain both symptomatic and survival benefits.