Changeover: the lost hours between one job and the next

Changeover is the time between the last good unit of one job and the first good unit of the next. It is not downtime in the sense of a breakdown — nothing is wrong — which is exactly why it goes unmeasured for years. A line losing ninety minutes four times a day is losing six hours of capacity daily, and that capacity is usually bought back by buying a second machine.

Measuring changeover honestly

• Time from last good unit to first good unit, which includes the trial runs and adjustments people leave out.
• Measured on the floor rather than estimated in a meeting — the estimate is reliably half the real figure.
• Recorded per changeover type, since not all are equal: similar product to similar product is a different job from a full tooling change.
• Including the waiting: fetching tools, finding the paperwork, locating somebody with the authority to sign off the first article.
• Separated into internal work, which requires the machine stopped, and external work, which does not.

The split that does the work

Almost every changeover contains work that did not need the machine stopped — tools collected, next materials staged, paperwork prepared, settings written down. Moving that work outside the stopped window is the single largest and cheapest improvement available, and it typically removes a third of the time without buying anything. The remaining internal work is what deserves engineering attention, and only after the easy half has been moved.

Standardise the sequence before trying to speed it up. A changeover done four different ways by four operators cannot be improved, because there is no baseline to improve against and every measurement reflects who happened to be on shift.

Why batch sizes follow from changeover

Long changeovers force long runs. If switching costs two hours, nobody switches for a small order, so batches grow, stock grows, and the business ends up holding inventory to compensate for a setup problem. That is the hidden cost: the inventory on the shelf is the price of the time lost at the machine. Halving changeover time does not just add capacity — it makes smaller runs economic, which is what lets a small manufacturer quote short lead times without carrying everything in stock.

Getting started without a programme

1. Film one changeover end to end and watch it with the people who did it; the waste is obvious on video and invisible in person.
2. List every step, mark each as internal or external, and move everything external out of the stopped window first.
3. Prepare a changeover kit per job so nothing is fetched while the machine waits.
4. Write the settings down rather than rediscovering them by trial, which is where the adjustment time goes.
5. Record the time for every changeover for a month, since the average matters less than the spread.
6. Re-time it after each change, because improvements that are not measured get quietly abandoned.

Ettex Sheets is enough to hold the log — job pair, type, start, first good unit and who ran it — which is all the data an improvement needs. Where the result changes what can be promised, it feeds straight into the production schedule, since shorter changeovers are what make a tighter sequence realistic.