In many academic settings completeness is defined for cases where you expect the data to be complete. However in the real world you have too much data and never enough time to document what and when data should be complete. It is vastly more efficient to use Collibra DQ
Profile features to generate statistical process control around completeness. This will alert you to a change in slope, or in other words a drastic change in completeness, which tends to be the exact DQ events you care the most about.