This paper presents a novel technique for representing cost information of a candidate design. The underlying principle is to provide a knowledge-based system for cost modelling that makes the cost effects of design decisions more visible and quickly available to the designer. The development of a hybrid between hierarchical trees and object-oriented knowledge representation environment to represent cost information is shown. The development of an abstract product structure, reusable tree object libraries and an automated system to extract design information needed for estimation from computer-assisted design geometry is discussed. It is shown how the tool provides fast incremental cost fluctuations in response to changes in component geometry. This paper also considers the uncertain quantities present in cost models and analyses cost risk. A case study demonstrating the research applied to a Rolls-Royce aircraft gas turbine engine component is presented