Since 2010, the Survey of Household Spending (SHS) has adopted an international data collection model, aimed at reducing the recall error in expenditure data and the recall effort of respondents. It transitioned from a 12 month recall period for all items to a varying shorter recall period based on the type of expenditures. These changes have had impacts on SHS users: the annual proportion of households with a given expenditure and the annual average spending among reporting households are no longer published. Several annualization techniques have been evaluated to produce these statistics of interest once again, thus allowing a more complete use of the SHS data as well as being more adapted to the users’ needs. The presentation will discuss the challenges associated with this objective and will describe a solution recently developed that relies on the minimization of the change in the consumers’ patterns and on modelling, through segmented regression.