The Indigenous Enumeration Strategy (IES) of the Australian National Census of Population and Housing has evolved over the years in response to the perceived 'difference' of the Aboriginal and Torres Strait Islander populations. Its defining characteristics are the use of locally recruited, mostly Indigenous collector interviewers, and the administration of a modified collection instrument in discrete Indigenous communities, mostly in remote Australia. The research reported here is unique. The authors, with the assistance of the Australian Bureau of Statistics, were able to follow the workings of the IES in the 2006 Census from the design of the collection instrument to the training of temporary census field staff at the Northern Territory's Census Management Unit in Darwin, to the enumeration in four remote locations, through to the processing stage at the Data Processing Centre in Melbourne. This allowed the tracking of data from collection to processing, and an assessment of the effects of information flows on the quality of the data, both as input and output. This study of the enumeration involved four very different locations: a group of small outstation communities (Arnhem Land), a large Aboriginal township (Wadeye), an 'open' town with a majority Aboriginal population (Fitzroy Crossing), and the minority Aboriginal population of a major regional centre (Alice Springs). A comparison between these contexts reveals differences that reflect the diversity of remote Aboriginal Australia, but also commonalities that exert a powerful influence on the effectiveness of the IES, in particular very high levels of short-term mobility. The selection of sites also allowed a comparison between the enumeration process in the Northern Territory, where a time-extended rolling count was explicitly planned for, and Western Australia, where a modified form of the standard count had been envisaged. The findings suggest that the IES has reached a point in its development where the injection of ever-increasing resources into essentially the same generic set and structure of activities may be producing diminishing returns. There is a need for a new kind of engagement between the Australian Bureau of Statistics and local government and Indigenous community-sector organisations in remote Australia. The agency and local knowledge of Indigenous people could be harnessed more effectively through an ongoing relationship with such organisations, to better address the complex contingencies confronting the census process in remote Indigenous Australia.