How do positive interventions work and for whom? We hypothesized that positive interventions affect both common and specific outcome variables as well as that effects depend on person-intervention fit. We randomized 432 German adults to perform either an optimism, gratitude, self-compassion, or control writing intervention in an online setting. Participants reported emotional self-awareness and trait gratitude before and affect, optimism, gratitude, and self-compassion immediately after the intervention. Path analysis results indicate strong specific effects for the optimism condition on optimism and the gratitude condition on gratitude. Unexpectedly, the optimism condition also increased gratitude and the self-compassion condition showed no effects. As expected, multiple group analysis results indicate that the gratitude intervention only provided benefits for individuals high in trait gratitude. Future studies should apply longitudinal mediation designs to identify which variables explain the effectiveness of positive interventions and whether effect mechanisms differ between subgroups of participants.