For Semantic Web Service discovery, matchmaking techniques are used to compare descriptions of semantically annotated services issued by service requesters and providers. Such semantic descriptions are typically expressed in ontology languages to capture the functionality of a service in a meaningful and machine-interpretable way. Among various other techniques proposed in the literature, intersection-based matchmaking rates two service descriptions as compatible if they agree on some concrete instantiation of service parameters. We report on an intersection-based matchmaking framework which we have applied to an industrial logistics case study. We particularly focus on the how-to of designing semantic service descriptions, providing modelling guidelines and patterns, an aspect that is often neglected in matchmaking literature. By an example from our logistics case study that involves sophisticated ontology reasoning, we show how some of the limitations of this matching technique can be overcome by an appropriate modelling.