The costs of purchased goods are typically one of the biggest parts of a company's expenses. Hence, the data created from these transactions can and should guide the management of the business relationships between a company and its suppliers, e.g. finding the best supplier for a certain type of goods, negotiating terms and conditions, etc.

When analyzing the respective purchasing transaction data, several negative business consequences are induced by the lack of consistent and integrated conceptual structures that would properly represent the rich semantic relations between item data, e.g. at the level of functional
similarities.

This thesis presents an approach of how Semantic Web technologies, in particular ontologies, can be used to make procurement data accessible at a conceptual level. As a key contribution, the OntoSpend prototype was developed, which provides (1) a repository for storing, cleansing, augmenting spend data at a knowledge level, (2) infrastructure for classifying item data according to products or services ontologies, (3) a Web interface that allows executing arbitrary SeRQL queries against the data, taking into account the domain knowledge contained in available ontologies and mapping data, and demonstrates the potential of this approach by a function for an extended ABC analysis, which considers item similarity beyond part identifiers.

The OntoSpend approach helps to discover previously unavailable findings on expense data, which may help to reach significant business benefits
and thus to lower costs of procurement.