Night shift pack-out: lighting, checks and error rates

The finding that surprised the plant was not that the night shift made more errors. It was that it made fewer. Night ran 1.9% rejects against day’s 4.1% over a full quarter, on the same line, the same packs and a younger, less experienced crew. Management assumed the day figure was the baseline and the night figure was a measurement problem. It was neither. Night had two things day did not: uninterrupted runs, and a supervisor with nothing else to do.

Night shifts are usually treated as a staffing problem. They are better understood as a different operating environment, with different failure modes, and the interventions that work are not the ones that work at 10:00.

Lighting changes what people can see

A pack-out room lit to 300 lux at bench level is fine for reading a label and marginal for judging whether a pack is fully frozen. Frost and condensation look similar under warm light, and a pack that is still soft in the middle looks identical to a hard one from above. Bump bench-level lighting to 500 lux and add a cool-white task light at the inspection point, and the number of soft packs found before sealing goes up sharply.

The colour of the light matters as much as the level. Warm 3,000 K lighting makes a pale gel and a frosted film look the same. A 5,000 K task light shows the difference in surface texture, which is what an operator is actually reading when they decide a pack is ready.

Checks that survive a tired brain

At 02:40, a checklist with fourteen items gets three ticks and a signature. The failure mode is not refusal, it is automatic compliance: the hand ticks while the eyes are somewhere else. The fix is to make the check physical rather than documentary. A spacer bar that must fit between two packs. A weight check that must produce a number written down. A count that must match a physical stack of empties.

Three physical checks per carton is the practical maximum at night. Choose the three that catch your most expensive failure and let the rest be covered by sampling. On a chilled lane those three are usually: pack hard at the corner when pressed with a thumb, count against the stack of empty slots, and carton closed within the two-minute window after the lid goes on. All three take under fifteen seconds and none of them require reading anything.

The handover is where the errors are born

Most night-shift quality failures we have seen begin in the twenty minutes around shift change, not at 03:00. The outgoing supervisor summarizes verbally, the incoming one is still putting on a coat, and the thing that was about to run out, or the pallet that was quarantined, gets lost in the transition.

Replace the verbal handover with a one-page sheet that both supervisors sign: line status, packs available, packs quarantined, open deviations, and the one thing the next shift must not do. Five fields, two signatures, four minutes. The night shift will fill it in because there is nobody else to talk to.

The counter-intuitive conclusion

Do not try to make the night shift behave like the day shift. Night is quieter, and quiet is an asset: fewer interruptions, fewer changeovers, a straight run of six hours on one configuration. That is why the error rate is lower. Protect it. Schedule the complex, mixed-SKU work for the day and the long single-configuration runs for the night, and both shifts get better numbers.

The exception is anything requiring a judgement call: a pack that looks marginal, a carton that has been dropped, a deviation that needs a decision. Those should be pushed to the day shift, because a tired decision at 04:00 tends to be the permissive one.

If you are writing a night-shift procedure, the inspection points in our ice pack handling guide are specified as visual and tactile checks rather than gauge readings, which is what works under task lighting at three in the morning.

Walk your pack-out room at 03:00 this week with a light meter and a blank handover sheet. Both will tell you something the daytime walk never will.