What you are buying
Before comparing data quality results, agree on the dataset boundaries and intended use. Check what was measured, how records were selected and whether a result from one sample or period still represents the data currently being used.