Looking at photo metadata, typical features exploited for semantic enrichment and retrieval are content and basic context features extracted from the photos' Exif headers. Interestingly, so far little attention has been spent on observing and exploiting the usage of media, such as when was a photo edited and how, for which purposes has it been used, or whom was it given to. One reason is the lack of proper means to store this information together with the single photo. In this paper we give examples of exploiting photo usage metadata. We identify different kinds of usage metadata and derive a conceptual metadata usage model from this. Finally we transform this model into an extension of the XMP metadata standard.