as a reviewer I pretty much follow exactly what Heidi outlines so nicely in her message. Internal to my current library, I'm the main contributor, so I sometimes ask someone outside the library to review a heading I'm unsure of (for whatever reason), but otherwise do not have another reviewer.
I am somewhat compulsive and so generally review the NAR a day or two after it's been contributed. I catch the odd typo (usually in the 670 field) that I then fix.
At a past library, we did for awhile always review someone else's record before contribution. We discontinued doing so because, as you ask, it did not justify the time spent and the slowdown in workflow.
Head of Cataloging
Central Washington University Library
400 E. University Way
Ellensburg, WA 98926-7548
[log in to unmask]