Since 1997 New Labour administrations have seen homelessness as a manifestation of social exclusion. Initially, ministerial attention focused on curbing street homelessness, exemplified most prominently by the Rough Sleepers Initiative. Subsequently, since 2003, the focus of the policy has widened to the promotion of a more pro-active, preventative approach in local authorities’ homelessness work. This fits with the ‘activist state’ approach which underlies much of New Labour’s social policy. Similarly, both phases of policy have been framed in terms of tightly defined statistical targets and timescales. While the definition and quantification of homelessness has been the subject of perennial interest, the commitment to homelessness prevention has introduced an important new angle to the debate. The significance of this issue is apparent from the fact that since the introduction of a new generation of prevention activity, official homelessness figures have fallen sharply. While the focus on prevention, rather than crisis response, is welcomed, this shift calls for robust mechanisms to monitor the effectiveness of the mechanisms concerned. Drawing on recently completed research this article demonstrates weaknesses in current methods for enumerating homelessness, both in understanding the true scale of the problem, and in monitoring the extent to which ‘prevention activities’ help to solve housing problems as opposed to impeding applicants’ chances of securing a social rented tenancy. Whilst Central Government has now introduced a statistical measure aimed at quantifying effective homelessness prevention activity the slipperiness of the whole concept raises questions as to practicality of this exercise and the value of the data to be generated.