The overall aim of this study was to examine how the characteristics of two salient stimuli -music and message- of an audio advertisement influence the psychological state of consumers and how such a state subsequently determines their cognitive and affective responses to the advertisement. In achieving this aim, this study was guided by a combination of two cognitive resource utilisation theories, Limited Capacity Model of Motivated Mediated Message Processing (Lang, 2000) and Resource-Matching Hypothesis (Anand & Sternthal, 1989). In particular building upon inconsistency and load theories, this study proposed that certain stimulus characteristics prompted certain states of a consumer’s cognition. These two stimulus characteristics were the congruence of musical stimulus and the complexity of the message stimulus. The model then predicted the potential effect of these characteristics on certain psychological states (Psychological Discomfort and Cognitive Load) leading to affective (Attitude towards Advertisement) and cognitive (encoding, storage, and retention) responses. To empirically examine this model, an online experiment (using a 2 x 2 between-subject x 2 with-in subject mix design) was conducted, in which a mixed sample of 284 subjects was exposed to a set of audio advertisements especially designed for this study. Unfamiliar music in conjunction with a fictitious brand was used and the exposure level was maintained at low. ANCOVA, MANCOVA, two-stage hierarchical regression analysis, and Repeated-measures MANCOVA were administered to test the hypotheses presented in the conceptual model. Among major findings were that the multiple informational structures in a complex message positively influenced cognitive load, while congruent music was capable of attenuating the level of cognitive load. Incongruent music, on the other hand, was capable of generating a dissonance state experienced as psychological discomfort that in turn increased the level of cognitive load as a result of listener’s trying to resolve such a state. Both dissonance and cognitive load negatively influenced attitude towards advertisements, and the affect primacy of attitude formation appeared to be more applicable. Though high cognitive load clearly undermines encoding, storage, and retrieval processes, no evidence was found to support the Resource-matching Hypothesis. Furthermore, the findings suggested that the cognitive load offset by the congruent music would increase advertisement effectiveness by enabling its message to carry more information and by generating more favourable attitudes.