While the last decade has seen a growing academic interest in how states count and classify Indigenous peoples in the national census, most research has been limited to small case studies. In contrast, this study examines key patterns of Indigenous enumeration worldwide spanning the period 1985 to 2014. This comparative perspective is valuable because it theorises practices of ethnic counting and classification as a phenomenon positioned with a broader social context rather than a parochial practice that can only be explained by unique historical or political factors within states. Two key questions illuminate this study: how widespread is Indigenous enumeration in national censuses globally over the focal period (1985 to 2014); and, how have forms of classification of Indigenous peoples in the census changed. To explore these questions I utilise data from a unique time-series database from the Ethnicity Counts? Project which combines information about civic and ethnic questions asked in national censuses, with data on countries’ social, economic and political characteristics. The research findings show the number of states employing some form of Indigenous enumeration in censuses has increased over the focal period. Furthermore strategies of enumeration have shifted, with ethnicity type questions becoming more prevalent over time. Despite the observed increase in Indigenous enumeration over the focal period, the majority of Indigenous peoples are not counted in their national census. Ultimately this study provides new empirical findings regarding patterns of state enumeration of Indigenous populations which highlight the need for improved coverage and quality of data on Indigenous peoples.