Ran the CSV through our deduper first. Connect rate jumped noticeably once we filtered by category. Saved us hours of manual cleanup.