Figure out how much money bad map data costs the company, then figure out how much money it takes to license good map data (clearly this exists somewhere). If licensing costs hurt more than bad maps, then continue with the current product and make slow improvements. Otherwise get out the checkbook.