Modern search engines aggregate information from a variety of sources (e.g. images, videos) and return this information to users, merged into a single results page. Current aggregation techniques are limited to merging blocks of heterogeneous content into organic result rankings. We propose a new approach to search aggregation that takes into account result semantics and explicit searcher preferences in the form of result composition. Our findings suggest that result composition can be an effective search paradigm and can positively impact search behaviour in certain contexts.