Introducing groups to the MADCOW annotation system solves the privacy-collaboration problem for users, but a new challenge emerges for the group joining process in terms of irrelevance, time and effort, for both users and groups’ authorisers. We introduce MADCOW domains and link them to domain ontologies. Users can then use tags derived from ontology terms to synthesize the intent and scope of their annotations. We propose the use of matching measures between domains and users’ annotations as the basis for suggestions of groups to users and of users to groups authorisers. A reduction of the mentioned problems ensues.