Self-replication of nucleic acids in the absence of enzymes such as polymerases or ligases represents an important and poorly understood step in the origin of life. In fact, the self-replication of nucleic acids remains strongly hindered by product inhibition, even when enzymes are present. Studying one of the few successful examples of enzymatic DNA self-replication based on a simple ligation chain reaction can shed light on how this fundamental process may have originally evolved. Previously, our group reported lesion-induced DNA amplification (LIDA), which is an isothermal ligase chain reaction. It was proposed that the lesion, a destabilizing abasic group, played an important role in reducing product inhibition. However, using abasic groups in non-enzymatic ligation cycles did not yield appreciable amplification. In order to identify the unknown factors that overcome product inhibition in LIDA, we have performed a detailed kinetic and thermodynamic analysis of the process, employing isothermal titration calorimetry (ITC) and global fitting of PAGE fluorescence data to characterize the individual steps of the amplification process. We found that in the absence of an enzyme, the ligated product binds four to five orders of magnitude more tightly to the template than to the shorter unligated strands of the intermediate complex. In the presence of T4 DNA ligase, this stability gap was reduced by two orders of magnitude, such that the intermediate and product complexes were roughly equal in stability. Thus, the ligase helps overcome product inhibition by reducing the affinity difference between the DNA product duplex and the intermediate complex. Furthermore, kinetic simulations showed that the stability of the intermediate complex, as well as the rate constant of ligation, significantly impacts the rate of self-replication, suggesting that catalysts that stabilize the intermediate complex might be a route to efficient nonenzymatic replication.