Changeover management: running the switch between products without losing the day

Readiness is checked before the machine stops, not after

The single biggest determinant of changeover duration is whether the next job's requirements were assembled while the current one was still running. That means tooling located and confirmed serviceable, materials staged at the machine, consumables present, documentation and settings available, and the people who will do the work not committed elsewhere. A short readiness check completed some agreed time before the switch, with a named person accountable, turns a scramble into a procedure. Where readiness fails, the supervisor still has time to change the sequence rather than stopping a machine to wait for a trolley.

Who does what while the machine is down

Changeovers on a shared machine often involve a setter, an operator, sometimes maintenance and sometimes quality, and time disappears when each waits for the other. Assign the tasks explicitly, including which can run in parallel and which must follow in sequence, and make sure the people know their part before the stop begins. Watch for the common pattern where one skilled setter does everything while others stand available, which is a role design problem rather than a skill problem. Recording who was involved and where the waiting occurred gives the supervisor something specific to fix.

First-off approval is part of the changeover, not a separate event

The machine is not changed over until it is producing acceptable product, and the time spent adjusting, sampling and waiting for approval belongs inside the recorded duration. Treating it as separate hides where the time actually goes and creates an incentive to release product before it has been confirmed. Decide in advance who can approve a first-off, make sure that person is available at the expected time, and define what happens if approval fails: how many adjustment cycles before escalation, and who decides to stop and investigate rather than continuing to tune.

Sequencing the plan to reduce how many changeovers happen

A changeover that never happens costs nothing. Grouping work by tooling family, colour, grade or cleaning requirement reduces both the number of switches and their individual length, since some transitions are far more demanding than others. Where transitions have a natural direction — light to dark, low allergen to high, fine to coarse — a standard sequence is worth defining and following. This has to be balanced against delivery dates, and the balance is a scheduling decision made daily, so the rule should be written down rather than left to whoever builds the list.

Changeover is where quality escapes are made

The transition between products is when wrong components, residual material from the previous run, incorrect settings and mislabelled output are most likely. Line clearance is the control: confirming that all material, packaging and documentation from the previous product has been physically removed before the next begins, verified by someone other than the person who did the work. Regulated sectors treat this formally, and it is worth the same rigour anywhere product mix-up would be costly. A changeover recorded as fast but followed by a batch rejection has not saved anything.

Frequently asked questions

How should changeover time be measured?

From the last acceptable unit of the previous product to the first acceptable unit of the next, at rate. Measuring only the mechanical work understates the event by leaving out the run-down, the adjustment cycles and the wait for approval, which is where much of the loss sits. Use the same definition everywhere so comparisons between shifts and machines mean something, and record the elements separately so improvement effort can be aimed at the largest one.

Should we schedule changeovers at shift ends or in the middle?

Mid-shift usually works better, because the full team is present, support functions are available and the work does not get split across a handover. Changeovers begun near the end of a shift frequently finish on the next one, with the incoming team inheriting a partly reassembled machine and no record of what was done. Where the volume of changeovers makes this impractical, at least ensure a documented handover point exists at a natural break in the sequence.

Who owns changeover performance, production or engineering?

Production owns the execution and its duration; engineering owns making the changeover inherently easier through tooling, quick-release fittings, presetting and standardisation. Both are needed and they fail differently. Production alone will drill the existing method until it plateaus; engineering alone will design improvements the floor does not adopt. Reviewing recorded changeover times together, with the setters present, keeps the two connected to the same evidence.