AOL did not have a common vocabulary across its 56+ media brands and content platforms which made content difficult to classify and relate. To address this, they created a faceted ontology with over 6,800 subjects and 7 entity vocabularies to classify all AOL content in the same way. They had to train the classification engine by providing training data for each subject. In the future, they aim to integrate more with advertising systems, provide better recommendations to editors, and continue refining the ontology.