Improving the way we measure LUR