This article describes two major studies (i.e. Project MATCH and the US Department of Veterans Affair Effectiveness Study) of mediators and moderators of drinking behavior change following treatment, and considers alternatives to the dominant paradigm that has guided treatment research for more than two decades. Both studies were designed to test the “Technology Model” of therapeutic change, which postulates that patient attributes and treatment process elements, respectively, constitute mediators and moderators of change in drinking and drug use following treatment. The studies show that matching to therapeutic orientation is not an essential ingredient to substantially enhanced outcomes, as previously believed. They also indicate that the mediational mechanisms underlying several of the most popular therapies are different than what is suggested by their proponents. The inability of the Technology Model to explain, much less improve, the effectiveness of addiction treatment suggests the need for a new paradigm. It may be more fruitful to look for matching in larger populations at the level of communities or treatment systems, where a wider range of settings and therapeutic interventions can be evaluated.