When comparing aggregate metrics, e.g. average ratings per type of user, I would like to be able to view the margin of error around these, so that I'm communicating whether the differences between groups are meaningful or whether the variation is ...