||Construction of buildings is a very special kind of production because each product is generally produced only once, as an individual. Therefore, there is no means of an empirical assessment of cost as a function of resource deployment, time, and sequence of tasks, by way of experimentation as in the case of industrial mass production. On the other hand, buildings are very expensive, so that there is a great need for reliable cost estimation, estimate updates during construction, and cost analysis. Due to the lack of known, reliable functional relations between the key cost-relevant quantities, very coarse simplifications of the system need to be introduced. The contribution discusses ways to develop such simplified cost models, without rendering valueless the results obtained by such a model. We assume that the whole cost estimation process will be implemented into an overall schedule planning system, without introducing insurmountable overhead.