Ideally all the books in the API should stay in the store even if they haven’t been included on any of the named shelves in the last Goodreads scrape. When a new scrape is run it would add any books that don’t appear, shift any books that are in the store but don’t appear in the latest scrape to a no-shelf status (representing books I know about but have no relation with, I guess). All of that is much easier if Goodreads has a persisting book ID. If the data persists I can add my own columns to the data too, like whether I own the books.