Despite the overall success of CBT for Major Depression, it has become increasingly clear that the outcomes for CBT have hit a plateau, and that alternative models of care are beginning to approximate those of CBT. While the dropout rates for CBT for depression are only about 10%, suggesting that the treatment is highly acceptable, the overall success rates of about 60 - 67% have not improved since the approach was first introduced. This keynote address offers a framework to understand our failure to improve outcomes. In particular, it will be noted that Major Depression is a heterogeneous disorder, which likely has several types of genesis, and complex risk and resiliency profiles. It will be argued that our CBT models need to formally reflect the diverse forms of depression, and that the field needs to adapt our models to recognize these variations in risk and resilience. A novel model, that incorporates different etiological considerations for depression will be presented, as will a concomitant research program to evaluate this model.