There is currently a resurgence of interest in estimating numbers of illicit drug users in Australia. Defining why numbers are needed, who is to be counted and how estimates should be derived are vital steps in the production of useful, valid estimates. We present a range of estimates of heroin users in the Australian Capital Territory which were developed as part of an investigation of the feasibility of prescribing heroin to dependent users. These produced estimates ranging from 433 to 1251 users. We conclude that household surveys, capture-recapture, and multipliers derived from nomination techniques have serious and often unrecognised limitations. Capture-recapture estimates, in particular, are unlikely to be useful, except at a local level. The best way forward for the derivation of national estimates for the National Drug Strategy is a three-pronged approach; national surveys, validated multipliers and monitoring of key indicators by drug-related services and agencies.