OBJECTIVE To test whether reducing the standard serving size of alcoholic beverages would reduce voluntary alcohol consumption in a laboratory (study 1) and a real-world drinking environment (study 2). Additionally, we modelled the potential public health benefit of reducing the standard serving size of on-trade alcoholic beverages in the United Kingdom. METHODS Studies 1 and 2 were cluster-randomized experiments. In the additional study, we used the Sheffield Alcohol Policy Model to estimate the number of deaths and hospital admissions that would be averted per year in the United Kingdom if a policy that reduces alcohol serving sizes in the on-trade was introduced. METHODS A semi-naturalistic laboratory (study 1), a bar in Liverpool, UK (study 2). METHODS Students and university staff members (study 1: n = 114, mean age = 24.8 years, 74.6% female), residents from local community (study 2: n = 164, mean age = 34.9 years, 57.3% female). In study 1, participants were assigned randomly to receive standard or reduced serving sizes (by 25%) of alcohol during a laboratory drinking session. In study 2, customers at a bar were served alcohol in either standard or reduced serving sizes (by 28.6-33.3%). METHODS Outcome measures were units of alcohol consumed within 1 hour (study 1) and up to 3 hours (study 2). Serving size condition was the primary predictor. RESULTS In study 1, a 25% reduction in alcohol serving size led to a 20.7-22.3% reduction in alcohol consumption. In study 2, a 28.6-33.3% reduction in alcohol serving size led to a 32.4-39.6% reduction in alcohol consumption. Modelling results indicated that decreasing the serving size of on-trade alcoholic beverages by 25% could reduce the number of alcohol-related hospital admissions and deaths per year in the United Kingdom by 4.4-10.5% and 5.6-13.2%, respectively. CONCLUSIONS Reducing the serving size of alcoholic beverages in the United Kingdom appears to lead to a reduction in alcohol consumption within a single drinking occasion.