Chimeric antigen receptor (CAR) T-cell therapies have shown significant benefits in the treatment of hematologic malignancies, such as B-cell acute lymphoblastic leukemia (B-ALL) and B-cell lymphoma. Despite the therapeutic advances offered by these innovative treatments, failures are still observed in 15% to 40% of patients with B-ALL and >50% of patients with B-cell lymphoma. Several hypotheses have emerged including CD19-negative or -positive relapses, low CAR T-cell activation and/or expansion in vivo, or T-cell exhaustion. To date, in the European Union, CAR T cells granted with marketing authorization are autologous and thus associated with a strong heterogeneity between products. Indeed, the manufacturing of a single batch requires cellular starting material collection by apheresis for each patient, with variable cellular composition, and then challenging pharmaceutical companies to standardize as much as possible the production process. In addition, these cost and time-consuming therapies are associated with a risk of manufacturing failure reaching 25%. Thus, there is a growing need to identify early risk factors of unsuccessful production and/or therapeutic escape. Quality of the apheresis product, pathology progression, as well as previous treatments have been reported as predictive factors of the variability in clinical response. The aim of this review is to report and discuss predictive factors that could help to anticipate the manufacturing success and clinical response.