The aim of the paper is to investigate the listing behaviour of sellers and agents. To this end, we measured the impact on the list price of the characteristics posted in sales advertisements - the starting point of the selling process. We considered two different measures: the traditional R2 measure to assess the strength of linear dependence between characteristics and price, and a general R2-like measure to assess the proportional reduction in variation of prices explained by the characteristics. Comparing the results was a first step towards investigating the way sellers and agents incorporate house characteristics into listing prices. We performed the analysis on a database of 4240 houses listed in the city of Turin -Northern Italy- in a time period from 2007 to 2012. Results show that the characteristics posted in advertisements explain almost all the price variation. We found empirical evidence that the linear relationship between house characteristics and price explains almost entirely the strength of their association: sellers and agents incorporate characteristics in prices using a linear model. Moreover, the characteristics most associated with prices are location, number of rooms and quality of the building, although only high levels of quality of the building are incorporated into listing prices. In contrast, we found empirical evidence that the other characteristics posted in advertisements, such as unit condition, type of building 'apartment building, detached house' - have a low association with listing prices.