Placebo controls in clinical trials are usually employed to filter out undesired, psychological or non-specific effects from "true" therapeutic effects. Although this is useful in the context of clinical trials for the purpose of proving pharmacological efficacy, it is misleading to generalize this strategy to therapy as a whole. This would imply that placebo effects are irrelevant. Precisely this opinion is criticized in this paper. After some consideration about an appropriate notion of placebo, some empirical evidence is reviewed which underlines the importance of placebo effects. It is likely that the unspecific effects due to the individual meaning of an intervention are an important factor of any therapeutic approach. It would be therapeutically desirable to maximize these factors. This point of view would also have consequences on how interventions have to be evaluated.