Over the last three decades housing economists have shown increasing interest in the analysis of the length of time a property spends on the market. This literature provides useful insights into the operation of local housing markets. It is clear, for example, that an appreciation of the factors influencing time on the market (ToM), and the signals associated with marketing duration, are important influences on selling strategies and price setting behaviour. It also seems that ToM will be related to property type and location and may be sensitive to submarket-specific demand and supply conditions. This issue provides the focus for our analysis. This paper seeks to explore the extent to which time on the market varies between spatial and structural housing submarkets and over time. The empirical part of the paper uses data on properties sold by the Solicitorsí Property Centre in Aberdeen, Scotland. The database contains details of more than 60,000 properties marketed between 1984 and 2002. Empirically, ToM is examined using survival analysis and consideration is given to the underlying distribution of the hazard model. Results for Weibull, lognormal and log-logistic models are presented. The results show that houses that sell for more than the asking price have a significantly increased hazard rate (i.e not surviving on the market). Length of time on the market also varies across housing submarket and house type.