The main aim of this paper is to investigate speech pauses and gestures as means to engage the audience and present the humorous message in an effective way. The data consist of two speeches by the USA president Barack Obama at the 2011 and 2016 Annual White House Correspondents’ Association Dinner. The success of the message is measured in terms of the immediate audience response. The analysis of the multimodally annotated data indicates that silent speech pauses structure and emphasise the discourse, and often precede the audience response. Only few filled pauses occur in these speeches and they emphasise the speech segment which they follow or precede. We also found a highly significant correlation between Obama’s speech pauses and audience response. Obama produces numerous head movements, facial expressions and hand gestures and their functions are related to both discourse content and structure. Characteristics for these speeches is that Obama points to individuals in the audience and often smiles and laughs. Audience response is equally frequent in the two events, and there are no significant changes in speech rate and frequency of head movements and facial expressions in the two speeches while Obama produced significantly more hand gestures in 2016 than in 2011. An analysis of the hand gestures produced by Barack Obama in two political speeches held at the United Nations in 2011 and 2016 confirms that the president produced significantly less communicative co-speech hand gestures during his speeches in 2011 than in 2016.