There are many ways in which the provision of social support can be ineffective. Recent research suggests that the benefits of support may be maximized when it is provided invisibly. What remains unknown, however, is whether invisible support reflects the skillful behavior of support providers or recipients' blissful unawareness, as well as how invisible support is delivered during spontaneous social interactions. We hypothesized that both providers' skillful behavior and recipients' unawareness are necessary for invisible support to be effective, and we sought to document what effective invisible support looks like. Eighty-five couples engaged in a videotaped support interaction in the lab. Support recipients whose partners provided more invisible practical and emotional support (coded by observers) but who reported receiving less support experienced the largest preinteraction-to-postinteraction declines in negative emotions. In the case of practical invisible support, the combination of more support and less awareness of that support also predicted increases in self-efficacy. These results indicate that invisible support is a dyadic phenomenon.