How would you systematically drill down into data to find the underlying reason for a sudden drop or spike in a metric at Conduent?
