The bracket's final pass belongs to a different drawing revision

Quality says the locating face used for the final check changed, so our apparent before-and-after pair is using two different references, which rather spoils its career as a correction measurement.

Then remove the calculated movement, but retain both original checks with their own references so the history does not become another unexplained blank.

Was the physical bracket corrected as well? Different references explain the bad comparison but don't tell you whether someone still had work to do between those entries.

Fabrication confirms correction happened, Bruno, but cannot recover an amount or active time against that bracket, so I can describe the route without putting a number on it.

Who is deciding what current work to observe before that number returns to the estimate?

Use the current drawing and fixture set for that observation, with the first released check recorded before correction. Keep incoming fit-up separate from the fixture seating check as well.

MinaBrown, process specialist and quality chose the current set, and Jane's first-release entry is on its traveller rather than a note someone is supposed to remember later.

I'd keep waiting separate from active correction time too. The gap between inspection entries can be mostly queue, even when somebody really did correct the bracket.

First observations now separate those stages: some passed at first release, others needed correction, and the long waits were mostly for the return check, so the old revision-mixed calculation is out of our cost sheet.

That gives you a current baseline. Has the proposed fixture been compared with it yet, or is that the next stage?