Three steps, done on one roll, not on a whole pallet.
The pilot. 200 to 500 labels in real conditions: same printer, same energy setting, same operator. Then the trials: cold room, 40 °C, rubbing, cleaning, 48 hours of waiting. What passes in the meeting room often fails on the loading bay.
The measurement. A verifier gives a grade to ISO/IEC 15416 for linear symbols and ISO/IEC 15415 for 2D. The minimum grade depends on the application and on the customer specification: it is read in the applicable specification, not in an average found online. A smartphone, even with a good app, does not measure a grade, it only says "it reads".
The real cost. Compare the cost per label actually scanned: facestock + ribbon + machine time + rework. A 210 m ribbon feeds roughly 2,000 prints on a 100 mm long label, which is a few thousandths of a euro per label. The real cost centre is not there: it is in the rework.
For pilot runs and variable-content requirements, going through a personalised label makes it possible to test the right format before committing to stock.