Production Changeover Loss

Many factories make more than one product on a line, and switching can involve cleaning, tooling, settings, material changes and quality checks during which the equipment cannot produce good units. A fictional bottling line stops at noon to switch from juice to water and resumes making approved bottles at 12:40, so under that line's rule the changeover stop is 40 minutes.

The measurement depends on what counts as planned production time, because if a line was never scheduled to run that interval is not an availability loss in the usual OEE model, and a fictional team that cleans equipment on a holiday shutdown does not add the entire shutdown to its changeover loss. Lean Production's OEE guide puts changeover stop time under availability loss when it interrupts planned production, while its OEE measure also separates slow running and quality rejects, so a plant should keep those categories distinct.

A fictional line takes 20 minutes to restart and then runs slowly for another ten, so the first interval is a stop and the slower run may be a performance loss under its definitions. The start event, such as the last good unit of product A, and the end event, such as the first approved unit of product B, must be used consistently, and a fictional operator who changes a die in 15 minutes but needs ten more to produce a conforming sample shows why a team must state whether both intervals count in its changeover clock, even though another organisation may choose a different boundary.

Count each changeover, not merely the total number of hours per month, because frequent short switches can consume more capacity than one long switch, so product mix matters. Record reasons, since tool search, cleaning, calibration and waiting for materials require different fixes and a single setup code hides useful information.

A fictional manager who sees 30 minutes lost to missing labels during every changeover finds that pre-positioning approved labels is a better fix than changing machinery. Lean Production's SMED approach distinguishes internal work that needs a stopped machine from external work possible while the line runs, so preparing tools beforehand can shorten downtime provided safety and quality checks are preserved.

A fictional crew stages clean tooling and verifies the next job before the machine stops, without bypassing lockout or inspection to save minutes. Changeover loss can be converted into illustrative capacity using a validated production rate, so a fictional line that saves 20 minutes and can make 100 units per hour estimates roughly 33 units of extra theoretical capacity but, because this is an opportunity estimate and not guaranteed sales or profit, confirms demand and material availability before claiming revenue.

Avoid rushing to shorten a necessary sanitation or regulatory process, because a faster change that increases defects or safety risk is not an improvement; a fictional food plant that cuts wash time and later fails hygiene checks restores the required cleaning and redesigns other steps instead. Product sequence can matter, since grouping similar products may reduce cleaning and tooling work but larger batches can increase inventory or delay customers, so a fictional bakery schedules allergen-containing products last where approved procedures support that sequence while still performing required cleaning between runs.

Track the human side, since skills, handoffs and clear instructions affect time and training improves repeatability more safely than informal shortcuts, and equipment condition matters too, because worn fixtures or inconsistent settings can lengthen setup. Cost reporting should not confuse lost time with an actual cash expense, because staff pay and equipment cost may continue during the stop while the value of missed output depends on the bottleneck and demand.

Record a baseline and test improvements across similar products and shifts, since one unusually easy switch is not proof of a lasting gain, and communicate the rule to operators, because teams that feel punished for recording real stop time make the data unreliable. Production changeover loss is the scheduled capacity used by a switch, so it should be measured honestly and avoidable work reduced without sacrificing safe, conforming production.