The ability to model preferences and exploit preferential information to assist users in searching for items has become an important issue in knowledge representation. Accurately eliciting preferences from the user in the form of a query can result in a coarse recommendation mechanism with numerous results returned. The problem lies in the user's knowledge concerning the items among which they are searching. Unless the user is a domain expert, their preferences are likely to be expressed in a vague manner and so vague results (in the form of numerous alternatives) are returned. In this paper we remedy this problem by exploiting ontological information regarding the domain at hand. This ontological information can be provided by a domain expert and need not concern the user. However, we show that it can prove useful in focussing query results, providing more meaningful and useful recommendations.