Antibody-drug conjugates (ADCs) have become important scaffolds for targeted cancer therapies. However, ADC exposure - response correlation is not well characterized. We demonstrated that intratumor payload exposures correlated well with the corresponding efficacies of several disulfide-linked ADCs bearing an DNA alkylating agent, pyrrolo[2,1-c][1,4]benzodiazepine-dimer (PBD), in HER2-expressing xenograft models. The correlation suggests that a threshold concentration of intratumor payload is required to support sustained efficacy and an ADC can deliver an excessive level of payload to tumors that does not enhance efficacy ('Plateau' effect). In contrast to tumor PBD concentrations, related assessments of systemic exposures, plasma stability, and drug-to-antibody ratio (DAR) changes of related ADCs did not consistently rationalize the observed ADC efficacies. A minimal efficacious dose could be determined by ADC dose-fractionation studies in the xenograft models. Mechanistic investigations revealed that both linker immolation and linker disulfide stability are the key factors that determine intratumor PBD concentrations. Overall, this study demonstrates how a linker design can impact ADC efficacy and that the intratumor exposure of a payload drug as the molecular mechanism quantitatively correlate with and predict the anti-tumor efficacy of ADCs.