Despite overall improvements in outcomes of patients with diffuse large B cell lymphoma (DLBCL), B30-40% of patients develop relapsed or refractory disease. For patients with chemo refractory disease, or recurrent disease following autologous hematopoietic SCT (auto-HCT), the prognosis is poor, with no consensus on the optimal therapy. Currently, owing to the graft vs lymphoma effect, hematopoietic allogeneic hematopoietic cell transplantation (allo-HCT) is the only potentially curative option for such patients. In addition, many patients who are considered today for auto-HCT actually have a low likelihood of benefit. For example, a patient with prior rituximab exposure who relapses within 1 year of diagnosis and has a second-line age-adjusted International Prognosis Index of 2 or 3 at relapse has a o25% chance of being cured by auto-HCT. It is possible that such patients may be better served with an allo-HCT. Unfortunately, in many cases, allo-HCT applicability is limited by patient age, comorbidities, performance status and treatment-related toxicities. Recent attempts to improve the efficacy of auto-HCT, such as incorporating radio-immunotherapy into the conditioning regimen, have not resulted in improved outcomes. However, incorporation of novel agents such as antiprogrammed death-1 antibodies as maintenance therapy after auto-HCT show promise. Allo-HCT in relapsed/refractory DLBCL patients can result in a 30-40% PFS rate at 3 years, in part due to a graft vs DLBCL effect. While reduced-intensity/nonmyeloablative conditioning is increasingly being used, certain patients may benefit from myeloablative conditioning. We present an algorithm intended to discriminate which relapsed and refractory DLBCL patients are most likely to benefit from auto-HCT vs allo-HCT. New approaches, using novel agents that target the molecular heterogeneity in DLBCL, will be an essential component of moving the field forward. Lastly, we propose a prospective registry-based study as the only feasible mechanism to define the optimal position of allo-HCT in the overall treatment strategy for DLBCL. It is hoped that this review will promote the development of prospective multicenter efforts to determine whether such patients do, in fact, benefit from earlier and/or more effective implementation of allo-HCT.