Physical inactivity is a health problem that has not been largely affected by mass communications. An interpersonal counseling intervention, THE COACH APPROACH, was tested with 2380 adults (overall M age = 40.9 yr., SD = 10.8) initiating exercise programs in 2003, 2004, and 2005, January through March, at nine YMCA facilities in the Atlanta, GA area. Contrasts of comparison (2002) and treatment condition cohorts suggested that the protocol, based on tenets of social cognitive and self-efficacy theories, resulted in significant reductions in exercise program drop out over 6 months. Changes in dropout were not significantly correlated to participants' income, percentage of full-time exercise counselors administering the intervention, or proportion of female counselors. Although the field design did not permit randomization of participants, results of conservative statistical testing suggested that THE COACH APPROACH may be effective for settings which allow for large scale dissemination of the intervention.