Changeover time is the elapsed time from the last good unit of one production run to the first good unit of the next. The practical formula is first-good timestamp minus last-good timestamp. This page shows how to compare standard and actual time, how three extra minutes repeated three times a day become annual capacity hours, and what to record by product, crew, and condition.
What changeover time is in manufacturing
The Lean Enterprise Institute defines changeover time as the time between the last piece in the completed run and the first good piece after the changeover. That quality boundary matters. A machine can restart before the process is stable enough to produce the first good unit.
Setup time is part of the wider changeover event. Depending on the line, the event can include cleanup, tooling or material changes, adjustments, verification, and startup. Define the boundary once, then use it across products, crews, and shifts.
This is a general manufacturing method. Food and beverage examples are relevant because sanitation, QA release, packaging, and scheduling can make the boundary visible, but the worked numbers below are illustrative, not plant data or an industry benchmark.
Changeover time formula, from last good unit to first good unit
Use this formula:
Changeover time = first-good timestamp - last-good timestamp
For example, if the last good unit of Product A is recorded at 10:00 and the first good unit of Product B is recorded at 10:33, the measured changeover is 33 minutes. The timestamp at 10:20 when the machine restarts is not the end of the event if the line needs adjustment or verification before it produces a good unit.
The record should carry the context beside the two timestamps:
| Field | What to record |
|---|---|
| Previous run | Product, line, and last good unit time |
| Next run | Product, line, and first good unit time |
| Event boundary | The agreed rule used by the plant |
| Result | Changeover minutes from the two timestamps |
| Context | Crew, shift, condition, and any startup quality issue |
Vorne’s manufacturing guidance on changeover measurement also recommends defining a standard and applying it consistently over time and across equipment. Those sources support the measurement boundary. The formula above is the direct subtraction needed to use it.
Standard versus actual changeover time
A standard is a reference for a defined product, line, and condition. Actual changeover time is what the event record shows. The useful comparison is the variance between them, not a single average that mixes different conditions.
| Record | Example | What it tells you |
|---|---|---|
| Standard changeover | 30 minutes | The approved reference for this defined context |
| Actual changeover | 33 minutes | What the last-good-to-first-good record observed |
| Variance | 3 minutes over standard | The gap to investigate, not an automatic cost claim |
| First-good confirmation | Recorded after verification | Whether the new run was actually ready |
Compare like with like. A product that needs sanitation should not be compared with a product that does not. A crew working with different tools or an equipment condition that changed during the shift should not disappear inside the same average.
There is no universal good changeover time to copy from another plant. The right reference depends on the equipment, product, changeover sequence, quality release, and operating conditions. A stable definition and a visible variance are more useful than a borrowed target.
Worked example: 3 extra minutes, 3 times a day
The video brief includes a simple operating example: a changeover that should take 30 minutes averages 33 minutes, and the plant performs three changeovers per day. The extra time is 3 minutes per event.
The arithmetic is:
- 33 minutes actual - 30 minutes standard = 3 minutes overrun per changeover.
- 3 minutes x 3 changeovers per day = 9 minutes per day.
- 9 minutes x 250 operating days = 2,250 minutes per year.
- 2,250 / 60 = 37.5 hours of capacity per year.
The 250 operating days is an explicit illustration. Replace it with the plant’s actual operating calendar. The result also changes if the number of daily changeovers changes. The transcript describes the accumulated effect as roughly 100 hours over a year, but that outcome depends on the frequency and operating days used. Do not copy that figure as a benchmark. Put the plant’s own values into the formula.
Capacity hours are not automatically lost sales. If the line is not capacity constrained, the consequence may be idle machine time, labour timing, utilities, cleaning, or a schedule shift rather than displaced product. If the line is constrained, the plant can investigate which order or product was displaced, while keeping that assumption visible.
Tracking changeover variance by product and crew
Start with a small event log. It should be detailed enough to reveal a pattern, not so large that the crew stops using it.
| Field | Example entry | Why it matters |
|---|---|---|
| Event date and ID | 2026-09-30, CO-014 | Lets production, quality, and engineering discuss the same event |
| Previous and next product | Product A to Product B | Shows whether the variance follows a product pair |
| Line and equipment | Line 2, filler 1 | Keeps comparisons like for like |
| Crew and shift | Crew B, night shift | Shows whether the method or handoff varies by team |
| Condition | Normal sanitation, tool missing, material late | Gives the variance a possible explanation |
| Standard minutes | 30 | Defines the reference |
| Actual minutes | 33 | Preserves the observed event |
| Variance minutes | +3 | Makes the investigation sortable |
| Last good and first good times | 10:00 and 10:33 | Proves the event boundary was applied |
| Startup scrap or rework | Quantity and disposition, if any | Prevents a speed gain from hiding a quality loss |
| Owner and next action | Production engineering, verify tool staging | Turns the record into a decision |
The pillar article on measuring the hidden cost of manufacturing waste and rework makes the same broader point: a local metric can look better while labour, rework, or lost production moves into another report. For changeovers, the control is to keep time, product, crew, condition, and first-good confirmation in the same event record.
What to do with the variance
First validate the boundary on one recurring changeover. Observe the event, separate work that can happen while the previous run is still active from work that requires the line to stop, and identify the step that creates the variance. Standard work or SMED can then address the measured problem.
SMED is a method for separating external work from internal work and streamlining the remaining steps. It is not a reason to skip measurement, and its under-ten-minute phrase is not a universal target for every product or process. The Vorne changeover guide is useful context for the method, but the first decision is still to agree on the event record.
Next step for food and beverage scheduling
Once the plant has a consistent changeover record, use the Food and Beverage Production Scheduling Toolkit to organize product sequence, crew context, and schedule decisions. If a rushed restart is also creating scrap or rework, compare the changeover record with First Pass Yield vs Final Yield: Formula and Examples. The two pages answer different questions, but they can expose the same startup event from time and quality views.
The useful sequence is simple: define the boundary, measure the event, preserve the variance, then choose the improvement method.
