The quantity and frequency of alcohol consumption are crucial both in risk assessment as well as epidemiological and clinical research. Using the Munich Composite International Diagnostic Interview (M-CIDI), drinking amounts have been assessed in numerous large-scale studies. However, the accuracy of this assessment has rarely been evaluated. This study evaluates the relevance of drink categories and pouring sizes, and the factors used to convert actual drinks into standard drinks. We compare the M-CIDI to alternative drink assessment instruments and empirically validate drink categories using a general population sample (n = 3165 from Germany), primary care samples (n = 322 from Italy, n = 1189 from Germany), and a non-representative set of k = 22503 alcoholic beverages sold in Germany in 2010–2016. The M-CIDI supplement sheet displays more categories than other instruments (AUDIT, TLFB, WHO-CIDI). Beer, wine, and spirits represent the most prevalent categories in the samples. The suggested standard drink conversion factors were inconsistent for different pouring sizes of the same drink and, to a smaller extent, across drink categories. For the use in Germany and Italy, we propose the limiting of drink categories and pouring sizes, and a revision of the proposed standard drinks. We further suggest corresponding examinations and revisions in other cultures.