Bell (1991) argued that utterances are designed primarily with their audience in mind. Audience design, tailoring an utterance to accommodate an addressee’s needs, is only successful if the speaker possesses an accurate knowledge of an audience’s requirements (Traxler & Gernsbacher, 1993). Experience of a task may enhance this knowledge, building detailed representations of addressees’ perspectives. We investigated whether direct experience and experience through observation (vicarious experience) increased audience design, and allowed it to transfer more easily between tasks. Personality was addressed, investigating whether certain traits are more adept at using audience design. A referential matching paradigm was used with two tasks to assess transfer. Directors instructed their matchers to retrieve shapes from a cabinet or box set-up and place them on a grid to replicate theirs. The set-up of both tasks was such that an instruction could be helpful or unhelpful. Experience was manipulated at three levels, directors having either: taken part in a practice game (direct); watched a game on a video (vicarious); or had no previous experience of the task. Four selecting games and four checking stages, where the director read out the completed grid, were played. Ninety-six responses were recorded per director and deemed helpful or unhelpful. The IPIP (Goldberg, 1999) questionnaire collected five scores for personality traits. Two one-way ANOVAs were conducted using either the number of helpful instructions or the difference between helpful instructions in the selecting and checking tasks as a measure for audience design against level of experience. Both produced marginally significant results between no experience and direct experience. This supports the beneficial qualities of direct experience and the mean patterns were promising for the advantages of vicarious learning. Correlations of helpfulness with personality traits were significant between the box selecting task and openness. Transfer and personality results were interesting, but showed little support for the hypotheses. Limitations and further research are discussed.