This is what comes back from the free diagnosis: a short written brief you can read in five minutes and act on yourself. There are two here. One ends in something we’d build. One ends in a written procedure and a few days of training. The companies, numbers, and details in both are fictional.
Your diagnostic brief — illustrative example
Why Changeovers Take a Day and a Half
Business snapshot
A 40-person contract manufacturer running three production lines, switching each line between products several times a week.
How it works today
- 1.The production schedule is set on Friday for the following week, including which lines change over and when.
- 2.When a line changes over, one senior operator sets the tooling, dials in the first-article settings, and signs off on the first good part.
- 3.Nobody else on the floor has run a full changeover on the two newer lines. Two operators have done parts of one.
- 4.If she’s on shift, a changeover takes three to four hours. If she isn’t, the line either waits for her next shift, or another operator works from memory and the first-article check fails more often.
- 5.The settings that worked last time live in her notebook and in her head. The setup sheets in the binder date from the original line install and no longer match the tooling.
What we found
High confidence — from your own production dataThe constraint isn’t the changeover procedure, and it isn’t the equipment. It’s that the working version of the procedure exists in one person’s memory and a personal notebook, while the written version stopped matching reality when the tooling changed fourteen months ago. That makes her the only reliable path through a step that happens several times a week, and it turns every one of her scheduled days off into a capacity decision nobody is consciously making.
- –Changeovers with her on shift average 3.4 hours. Without her, 11.2 hours including wait time.
- –14 of the last 60 changeovers happened without her.
- –First-article rejection rate is 4% on changeovers she runs, 23% on the others.
- –The setup binder was last updated three years ago. The current tooling went in fourteen months ago.
What we ruled out, and why
Hypothesis — reasoned, not testedOther diagnoses we considered
The equipment or the tooling is the limit
If it were, changeover time would be roughly constant regardless of who ran it. It isn’t. The same line with the same tooling swings from 3.4 to 11.2 hours depending on who is on shift. That points at knowledge rather than hardware.
Too many changeovers on the schedule
Running longer batches would cut the total hours lost without touching the cause, and it would push you toward more finished-goods inventory to carry. Worth revisiting afterward if the economics still favor it. It isn’t what’s limiting you now.
Cheaper fixes we considered first
Hire or promote a second senior operator
This is the closest to right, which is why it’s worth being careful about. It solves coverage and leaves the knowledge problem exactly where it is: you would then have two people holding it in their heads, and the same failure the first time both are out. Worth doing, and it works far better after the procedure is written down, because then you can train against a document instead of running a second apprenticeship.
Digital work-instruction or changeover-tracking software
The software-first answer. It would give you a good place to put the procedure, which is a problem you don’t have yet, because there is no current procedure to put anywhere. The binder already exists and goes unused because it’s wrong, not because it’s paper. Buy the tool later if you still want it.
Current business impact
Known impact
Roughly 110 hours of line time a year goes to changeovers she didn’t run, before counting scrap from a 23% first-article rejection rate on those.
Impact to calculate
What an hour of line time is worth at your current margin, and the scrap cost of the extra rejections.
If this were fixed
You’d get most of those 110 hours back, and her days off would stop being a scheduling risk. The larger thing you’d get is that a retirement or a resignation stops being an emergency.
What we’d change
- –Have her run the next four changeovers with a second operator shadowing and writing down the steps as they happen, including the settings she reaches for without thinking about them. She reviews and corrects the draft. She doesn’t write it.
- –Replace the binder’s setup sheets with what comes out of that, per line and per product family, using the current tooling’s real numbers.
- –Run the two changeovers after that with the shadowing operator leading and her available but off the floor. Whatever they have to ask her is your list of what’s still missing from the document.
- –Train a third operator from the document alone, with no shadowing. If that works, the knowledge has left the one head.
What’s included and excluded
Included
- –Four shadowed changeovers, documented as they happen
- –Rewritten setup sheets matching the current tooling
- –Two supervised changeovers led by the second operator
- –A document-only training run with a third operator
Excluded
- –Any change to the production schedule or run lengths
- –New tooling, equipment, or maintenance changes
- –Software of any kind
- –Any change to her role or compensation. That one is yours to decide.
How you’d know it worked
Changeover hours when she is not on shift
Baseline: 11.2 hours average today
Target: Under 5 hours within 60 days; under 4 once a third operator is trained
If we’re wrong: If this barely moves once the document exists and two operators have led a changeover, then what’s missing isn’t the setup steps. More likely it’s judgment about the material or the machine that a document can’t carry, and the answer is a longer apprenticeship rather than better paperwork.
First-article rejection rate on changeovers she didn’t run
Baseline: 23% today
Target: Under 8% within 90 days
If we’re wrong: If rejections stay high while changeover time drops, the operators are following the document and the document is wrong somewhere specific. Fixable, but it means another pass.
People who can run a full changeover on all three lines
Baseline: 1 today
Target: 3 within 90 days
What we’d still confirm
- –Whether the two newer lines differ enough to need separate documents
- –Her own read on which parts of the job she thinks can’t be written down
- –Whether anyone has tried this before, and what happened
Recommended next step
Validate the details above in one short working session, ideally on the floor during a real changeover. The four shadowed changeovers after that need no budget and no help from us — it’s your operators and a notebook. If you’d rather we ran the documentation and the training, we’ll scope that separately after the session.
That was a fictional example
See what a real brief would say about your business.
The free diagnosis produces this same kind of brief, built from how your business actually runs.
Get Your Free DiagnosisFree. Brief or follow-up within 24 hours.