The purpose of this paper is to understand better the number of people consuming alcohol, the types of beverages chosen and the amount of alcohol consumed. Design/methodology/approach - Actual alcohol consumption collected using the covert observation method is compared with claimed alcohol consumption collected through surveys to understand the extent of the gap between claimed and actual behaviour. Findings - A notable gap between claimed and actual drinking levels was evident. A total of 70 percent more males were observed drinking alcohol at risky or high-risk levels while 49 percent more females were observed drinking at risky or high risk levels when compared to claimed behaviour data. Further, a higher proportion of people were observed drinking alcohol than claims lead one to believe. Research limitations/implications - This research used human covert observation, thus limiting episode length. Further, this study was restricted to six venues in one month of one year. Future research opportunities abound including the use of electronic devices, variation in the observation methodology employed, and extending covert observation to different venue types, locations, and times of year. Practical implications - The covert observation method can be used to critique the impact of the socially responsible programs and practices. Public policy makers may need to be mindful that alcohol may be consumed by more people in larger amounts than is currently reported in studies employing survey methodologies. Originality/value - The paper demonstrates how the covert observation method can be used to record what consumers actually do. The covert observation method can be used to extend the understanding of alcohol consumption by enabling researchers to observe behaviour in naturalistic settings.