Shift changeover losses and how to win them back

You lose production at shift change in four places: the outgoing crew winds down early, the handover talk happens away from a running machine, the incoming crew starts late, and the first part waits for its check. Measure the loss as the extra stopped minutes in the last and first 30 minutes of each shift, compared with the middle, then win it back with an overlapping handover at the machine, a written handover card and the first part ready.

This guide is for plant managers running two or three shifts. It works one hypothetical mill through a week, computes the weekly loss two ways and ends with which fix to try first. Comparing whole-shift output is a different question, covered in day shift vs night shift.

What is a shift changeover loss?#

A shift changeover changes the crew, not the part. A part changeover swaps a mold, a die or a fixture, and you time it from last good part to first good part, as measuring changeover time sets out. That method treats a shift changeover as one more type. This guide measures the minutes around the handover itself.

The loss hides because no single stop looks wrong. A machine stopped 12 minutes at 2:00 PM reads like any other 12-minute stop. Only the pattern shows it: the same stops, at the same clock times, every day.

Where do the minutes go at a handover?#

The minutes go to four routines on either side of the boundary. Each leaves a different mark on the machine record, so the record tells you which one you have.

Early wind-down. The outgoing operator finishes a cycle at 1:48 PM and doesn't load the next part, because it wouldn't finish before 2:00. Cleanup and paperwork fill the gap.

The handover talk. The two operators meet at the time clock or the supervisor's desk, not at the machine. The machine sits through the whole conversation.

Late start. The incoming operator reads the job, finds the gauge and checks the tool before the first cycle. On a night-to-day handover, a morning meeting often adds 10 minutes here.

First-part check. The machine runs one part, then waits for an inspector to approve it. The stop sits inside the shift, a few minutes after the boundary, so it is easy to log under a generic reason.

How do you see handover loss in machine data?#

You see it by comparing the stopped minutes near each shift boundary with the stopped minutes in the middle of the shift. Handover loss shows as stops clustered at the boundaries, on the same clock times, day after day.

The mid-shift rate is the average stopped minutes per 30 minutes over the rest of the shift. It is the machine's normal stop rate: tool changes, gauging and jams. Anything above it in a handover window is handover loss.

1. Fix each boundary to one instant

   Write the shift boundaries down as clock times, for example 6:00, 14:00 and 22:00. Crews can overlap on the floor, but each boundary in your records is one instant.

2. Collect two weeks of stops

   Use a machine signal or a stop log with a start and end time for every stop, including stops of a minute or less. Two weeks covers each shift and weekday at least twice.

3. Split each stop at the boundary

   Count each stop's minutes into 30-minute buckets by clock time. A stop from 13:52 to 14:11 puts 8 minutes in the outgoing window and 11 in the incoming one.

4. Take out planned windows

   Remove breaks and part changeovers scheduled at the boundary from the windows and from the mid-shift rate. A planned mold change at 6:00 is a changeover, measured on its own.

5. Subtract the mid-shift rate

   For each window, subtract the mid-shift rate from the stopped minutes. The rest is handover loss for that window.

6. Split by handover and weekday

   Sum the loss by handover pair and by day of the week. A loss that belongs to one handover, or to one weekday, points to one cause.

Check the window length against the data. If the first window is stopped for more than 25 of its 30 minutes, the late start runs past it: widen both windows to 60 minutes and recompute the mid-shift rate without them.

How much does one machine lose in a week?#

Run the method on one machine and the loss becomes minutes, parts and a share of scheduled time. The machine below is hypothetical, with round numbers.

Both ways give 455 minutes, so the window sums are consistent. Against 168 scheduled hours, 7.6 hours is 4.5% of the week, lost to three routines that each look small. At 5 minutes a housing, it's 91 parts.

Split the 455 minutes two more ways before choosing a fix. The 6:00 handover carries 203 minutes, or 44.6% of the week. The incoming windows carry 105 + 70 + 84 = 259 minutes against 196 for the outgoing ones, so the larger share of the loss comes after the boundary: the late start and the first part.

How do you win the minutes back?#

You win them back by keeping the machine running while the crews change. Three fixes cover the four losses, and each targets a different side of the boundary.

The UK Health and Safety Executive's handover guidance is written for safety, and it fits production too. It asks for handover to be face to face, two-way, verbal and written, and given as much time as it needs.^[1]

Overlap the handover at the machine#

The incoming operator arrives before the boundary, for example 15 minutes early, and takes the handover at the running machine. The outgoing operator keeps loading parts until relieved. The talk and the wind-down both happen while the machine cuts.

The overlap costs paid minutes. One 15-minute overlap at the 6:00 handover costs 7 × 15 = 105 operator minutes a week, against the 203 machine minutes that handover loses.

Write a handover card#

A handover card is one page at the machine that the outgoing operator fills in during the last cycles, with the machine still running. It replaces the conversation's checklist, so the talk covers only what's unusual.

Keep the fields fixed so the incoming operator reads the card in the same order every time. Ask the incoming crew what they had to look for last week, and add that field.

Have the first part ready#

The first-part check costs the most when the inspector, the gauge or the program isn't ready at the boundary. Schedule the inspector's round to reach the machine 10 minutes after each boundary. Leave the gauges at the machine, and confirm the program before the outgoing crew leaves.

Where your quality plan allows it, the outgoing operator's check on the last part can stand in for a first-part check when nothing changed across the handover: same job, same tools, same program. Write that rule down with quality before anyone relies on it.

Measure handover loss automatically#

IoTFlows downtime tracking detects every stop from a sensor on the machine, and your shift schedule buckets every number by shift. Shifts don't overlap in the schedule, so each boundary stays one instant while crews overlap on the floor.

The Start of Shift and End of Shift rules put their own reason on a stop that runs across a boundary. Stops inside the shift, such as a first-part check, wait for an operator. The Downtimes report's CSV carries each stop's start and end, so you can bucket stops by minutes from the boundary.

Datwyler's injection molding plant runs both rules and reviews downtime week over week. Its Tuesday shift transitions turned out to add 8 to 10 minutes of downtime, every week. Datwyler cut shift-changeover downtime by 50%, as Datwyler's story describes.

Which fix should you try first?#

Try first the fix that matches the side of the boundary losing the most minutes. Read the split from your own windows, then pick from the table.

Start with the worst handover, not all three. On the mill, fixing only the 6:00 handover addresses 44.6% of the loss with one crew pair and one inspector round.

Then prove it. Keep the same windows, the same mid-shift rate and the same weekdays, and compare four weeks before with four weeks after. If the hourly output of the first shift hour rises too, the minutes came back as parts; the shift goal guide shows how to read that first hour against its goal.