The system currently flags duplicates / matches based on last name and address. This means I spend a lot of time de-duping families that all share the same last name and live at the same address. I would choose the system to search on first name, last name and then address when assessing if records are the same.