This experiment expanded the visual availability paradigm by subsuming it under the broader principle of recipient design. We varied recipient design by asking speakers to describe a picture to someone who would see a videotape of their description or only hear an audiotape. Second, speakers described pictures that varied in verbal encodability. Finally, in addition to gestural rate, we analysed the redundancy of gestures with words. The results (N = 40) confirmed our predictions that speakers gesture at a higher rate and use a higher proportion of nonredundant gestures when their recipient would see their videotape; that they also use more nonredundant gestures when describing a picture for which they have a poor vocabulary; and that these two factors interact to produce the strongest effects when vocabulary is limited and the recipient would see the videotape. These effects support the hypothesis that speakers design their gestures to communicate to recipients.