This is similar to https://jira.nypl.org/browse/SIMPLY-519 and inspired by similar user needs.

When you import an OPDS feed, you may know the intended audience of the works (e.g. children or researchers), even though the OPDS feed may not contain that information. It should be possible to configure a collection with a default audience, so that books imported from the OPDS feed end up with the right audience.

Unlike with the quality measurement, it's not clear to me whether we should add the audience classification even if the OPDS feed does include audience information. If the OPDS feed has audience information, and you specify a "default audience" anway, are you filling in the gaps, or are you saying that you disagree with the information in the OPDS feed?