During the first half of the twentieth century, psychotherapy research was synonymous to single case research. Research and practice were highly integrated in this era, but to be considered full-fledged scientific research, the case descriptions lacked methodological rigor. From 1970 onwards, the experimental Randomized Controlled Trial (RCT) design gained momentum in the field of psychotherapy research and the single case paradigm was marginalised. In this article, it is argued that the classical RCT design is ethically troublesome, created a dramatic gap between research and practice, fails to yield the promised objective evaluation of the efficacy of psychotherapy, and systematically disadvantages the types of therapy that prove to be most effective in everyday clinical practice. For several reasons, returning to the classical single case paradigm, however, is not an option. As an alternative, a research program based on naturalistic effectiveness studies and empirical single case studies is put forward. It is argued that, compared to RCT research, this type of research: (1) is methodologically superior; (2) is more informative towards clinicians; (3) is a more reliable basis for anticipating cost-effectiveness of psychotherapy.