Augos Factory / Learn / Best operator, trainer and trainee
Learn · operator variation
Your best operator already sets the target.
In short
Operators doing the same task on the same line rarely take the same time. Augos ranks them on their measured stops, takes the best operator's typical time as the benchmark, and adds up the time everyone else spends above it. The best operator becomes the trainer; the operators furthest behind on a task become its trainees.
The Best Operator Benchmark is the line's best plausible day, synthesised from your own best operators' times for each stop cause. It is an empirical maximum from your own floor rather than a textbook target. The gap between it and today's OEE is capacity the line has already reached, with the equipment, products and people it has now.
What it is
What it is
Standard OEE sees a machine; it never sees the person running it. Yet the operator decides how fast a stop is noticed, how many are cleared without help, and whether a changeover keeps to its allocated time. Across a shift those decisions are the difference between a good day and an ordinary one. The Best Operator Benchmark measures that difference. For every stop cause on the line, it takes the fastest operator's typical time, combines them, and asks what the line's OEE would have been if every stop had been handled the way the best operator handles that cause. The answer is a benchmark nobody can argue with, because a colleague has already met every part of it.
Source: Augos Factory method, Glossary: Best Operator Benchmark
Why your own best beats a textbook
A benchmark such as 85% OEE describes discrete manufacturing somewhere else, calculated by someone else's formula. It cannot say whether your changeovers are tight or your cleaning regime is sensible, and a crew told to hit it has every reason to doubt it. A benchmark built from the line's own best runs has none of those problems. It was achieved on this equipment, with these products, by people in this room. It is also honest about ceilings. A line whose best operator still loses forty minutes a shift to raw-material waits has a benchmark that includes those waits. The time account puts them with planning. What is a good OEE? covers the published benchmarks and where they come from.
Why operators
Site gaps are measured. Operator gaps are where you can start
Variation between sites needs group operations to move a method across a business. Variation between operators sits on one line, under one supervisor, this week. More experienced and disciplined operators get better results, and they show what the line can already do. Variation and standard work
How it is built
Four steps, all of them from measured events.
| Step | What happens |
|---|---|
| 1 · Attribute | Every stop is tied to the operator logged in at the Augos Operator Terminal. Stops with nobody logged in fall into an unassigned pool and are not attributed to anyone. |
| 2 · Clean | For each stop cause, extreme events are removed by the inter-quartile range rule, so one catastrophic stop does not set anyone's baseline. |
| 3 · Benchmark | For each cause, the median duration is the Typical Time, the median minus the median absolute deviation is the Recommended Time, and the best operator's median is the Best Operator time. |
| 4 · Synthesise | The line's actual stops are re-timed at the best operator's time for each cause. The result is the Best Operator day; the difference from the actual day is the recoverable gap, in minutes and in OEE points. |
Source: Augos Factory method, outlier and benchmark rules; Glossary: Typical Time and Recommended Time, Outlier (IQR method)
The ranking
How the best operator is found
| Step | What happens |
|---|---|
| Score every planned stop | Under the target time scores 2. Within the allocated time scores 1. Over the allocated time scores minus 0.5. An outlier, beyond the line's normal spread, scores minus 2. |
| Weight by paid time | Each operator's average score is multiplied by their share of the line's paid time, so a few lucky shifts cannot top the ranking. |
| Set the benchmark | The top-ranked operator's median time on planned stops becomes the benchmark: a time a colleague has already achieved. |
| Size the gap | Every stop longer than the benchmark adds its excess time. Stops faster than the benchmark add nothing. The gap is in hours; any rand figure is your own estimate. |
Ranking score = average stop score × (operator's paid time ÷ line's paid time). Benchmark = median planned-stop time of the rank 1 operator. Training gap = the sum of (stop time − benchmark), for every slower stop by every other operator.
Source: Augos Factory method, defensible calculations (stop scoring and paid-time weighting)
Trainer and trainee
Trainer and trainee
We suggest a simple first rule. An operator within 15% of the best operator's typical time on a task is at the benchmark. An operator more than 15% slower marks an opportunity. The best operator becomes their benchmark, and can mentor them or share how they do the task. Formal training is one way to close the gap, not the only one. The gap is in minutes, so every hour spent closing it has a number beside it. Plants can set a tighter rule once the first gaps close.
Changeover time by operator (illustrative):
| Operator | Typical time | Against the best | Role for this task |
|---|---|---|---|
| A, most paid time, rank 1 | 5.5 min | Best | Trainer |
| B | 5.8 min | +5% | At the benchmark |
| C | 6.1 min | +11% | At the benchmark |
| D | 6.6 min | +20% | Trainee |
| E | 7.2 min | +31% | Trainee |
| F | 7.9 min | +44% | Trainee |
Illustrative operators, not a customer. Typical time is each operator's median for the same changeover on the same line.
Trainer and trainee sign off, task by task
The trainer is the best operator for the task. The trainee does the task while the trainer watches, and both sign. After 10 or more events, the trainee's typical time is checked against the trainer's. Still more than 15% slower: the opportunity stays open, so keep sharing how the best operator does the task. How operators learn the terminal
Task by task
The same person can be a trainer and a trainee
Rankings are kept per task, so nobody carries a blanket label. One operator may be fastest on changeovers and slow to clear interruptions. That operator coaches changeovers and is coached on interruptions.
Planned stops
Changeovers, cleaning and set-ups. Each operator's time is compared with the allocated time and with the best operator's typical time. Coaching starts with the method and the tools.
Unplanned stops
Operator Evaluation and Operator Resolution: how fast a stop is noticed, and how many are cleared without a technician. Each has its own typical and recommended time.
On the record
Three months of planned stops, set against the line's own best
One beverage line, four operators, every planned stop. Each row compares the allocated time, the line's typical time and the best operator's typical time. Time above the best is the coaching opportunity, task by task.
| Task | Events | Allocated | Line typical | Best operator | Within allocated | Above the best | Trainees |
|---|---|---|---|---|---|---|---|
| CIP | 82 | 170 min | 170.1 min | 154.9 min (B) | 49% | 33.7 h | D |
| Start-up | 83 | 20 min | 20.8 min | 19.3 min (B) | 31% | 10.0 h | D |
| Tank change | 55 | 10 min | 5.7 min | 5.1 min (A) | 80% | 1.3 h | C |
| Quality check | 162 | 2 min | 3.5 min | 2.5 min (C) | 22% | 3.2 h | A, B, D |
| Hot rinse | 26 | 80 min | 83.7 min | Not graded | 46% | ||
| Nitrogen tank change | 19 | 10 min | 10.0 min | Not graded | 53% |
Source: Augos platform, Planned Stoppage Analysis. One anonymised beverage line, three months in 2026, operators A to D. Typical time is the median. Only operators with 10 or more events of a task are graded. Time above the best excludes outlier stops.
Two rows say as much about the standard as the people. Tank changes run well inside their 10 minutes, so the allocation has slack. Quality checks run past their 2 minutes on most stops, so the allocation may need review before anyone is coached.
One fault, four operators
On the beverage packaging line used across the method pages, one recurring fault at a single station halted the line 50 times in one month in 2026. The line's Typical Time for that cause was 6.1 minutes; the fastest quarter of stops cleared in 4.0. Four operators handled it.
| Operator | Stops handled | Median clearance | Graded |
|---|---|---|---|
| A | 40 | 5.5 min | Yes |
| B | 5 | 12.8 min | No: fewer than 10 events |
| C | 4 | 11.6 min | No: fewer than 10 events |
| D | 1 | 3.0 min | No: fewer than 10 events |
Operator A is the benchmark for this cause: 5.5 minutes, from forty events. Operator D's three-minute clearance is faster but is a single event and is not graded. Operators B and C clear the same fault in roughly twice A's time, but on four and five events each the platform withholds a grade and flags the cells; a median of five is not a verdict. What it does show the supervisor is where to look: if B and C take this fault more often next month and the gap holds, a specific, teachable difference exists in how A clears it. At the Recommended Time of 4.0 minutes the fifty stops would have taken 3.3 hours instead of 7.7.
Source: Augos platform, Unplanned Stoppage analysis, one anonymised beverage line, a 30-day month in 2026
The gap on a whole line
Repeated over every cause, the same arithmetic gives the line's gap. In the month, this line's Operator Resolution across all declared stops had a Typical Time of 7.1 minutes and a Recommended Time of 3.4 minutes; 51 hours of the month sat above the Recommended Time. The platform's Best Operator calculation put the recoverable gap at 69.7 hours, 9.7 points of paid time, on a line that produced for 41.5% of it. Across the Augos fleet the average gap between a line's current OEE and its Best Operator Benchmark is about 18 OEE points.
Source: Augos platform, Performance Overview, one anonymised beverage line, a 30-day month in 2026; fleet gap, Augos Factory method, Glossary: Best Operator Benchmark
Fair by design
Keeping it fair
The benchmark is a coaching tool, and four rules keep it one. Medians rather than means: a difficult shift caused by a failing machine does not define an operator's typical time. The ten-event rule: a metric with fewer than ten events in the window is suppressed and greyed out rather than graded. Peer masking: by default an operator sees their own numbers and the line benchmark rather than named colleagues' scores; supervisors and managers see the full picture. And ranking is weighted by time on the line, so an operator who covered a third of the paid hours is not outranked by one who covered a single shift. Operators are measured only on what they control: adherence to the allocated time and speed of response.
| Rule | Why it matters |
|---|---|
| The median, not the average | One bad shift with a faulty machine does not define anyone. |
| Outliers removed | Stops beyond the line's normal spread are scored separately and kept out of the typical time. |
| At least 10 events | Fewer than 10 stops of a type and the cell is greyed out, not graded. |
| The line's own best | The target is a time a colleague has achieved, not a textbook standard. |
| Same line, same task | Operators are only compared on work they all do. |
| Peer names masked | Operators see their own numbers and the benchmark. Supervisors see the names. |
Every stop sits in a timestamped event log, so a disputed result can be checked stop by stop. The shop-floor charter
Source: Augos Factory method, Operator Analysis rules (median statistics, N under 10 suppression, peer masking); Glossary
Consistency as a force
The benchmark also reports consistency, as a coefficient of variation: how widely an operator's times for a cause scatter around their median. Two operators with the same median can be very different colleagues to plan around. On the beverage line the four operators who shared the planned stops had compliance between 21% and 45% and coefficients of variation between 1.2 and 1.9; the operator with the highest compliance also had the lowest variation. A low number means the line behaves the same way every shift, which is worth more to a planner than an occasional heroic one. Consistent operator behaviour is a powerful force on a line; inconsistency, shift to shift, quietly destroys output, and the benchmark is how it becomes visible without anyone being singled out.
Source: Augos platform, Planned Stoppage analysis, one anonymised beverage line, a 30-day month in 2026; CV definition, Glossary
What needs Augos Assist
The Best Operator Benchmark and the operator dashboards are in the standard fee. The response-stage medians behind it, Operator Evaluation and Operator Resolution split apart, need Augos Assist, an optional extra. Without Assist, an unplanned stop is one number.
Source: Augos Factory method, Glossary: one offer
Questions
Common questions
What is the Best Operator Benchmark?
The line's best plausible day, synthesised from the best operator's typical time for each stop cause. The gap between it and the actual day is the capacity the line has already shown it can reach.
Isn't scoring operators just surveillance?
The method measures what operators control, uses medians so one bad shift does not define anyone, withholds grades below ten events, and masks peers' scores by default. Its purpose is to find who to learn from and what to coach.
How is this different from standard OEE?
Standard OEE reports that Availability was low. It cannot tell you that one operator clears a fault in half the time of another, or that the line has already run a better day than its target. The benchmark names both.
What makes one operator better than another?
Mostly how they handle a stop: noticing it quickly, clearing more without help, keeping to the allocated time on planned work, and doing it consistently. All of it is learnable, which is why the best operator is the training benchmark.
How big is the gap usually?
It depends on the line. On one anonymised beverage line it was 9.7 points of paid time in one month in 2026. Across the Augos fleet the average gap between current OEE and the Best Operator Benchmark is about 18 OEE points.
Will the unions accept individual measurement?
Objections are usually about opinion, not measurement. The median, the minimum of 10 events and masked peer names give one consistent rule for everyone, set by a time a colleague has already achieved.
Can operators see each other's results?
No. Operators see their own numbers and the benchmark. Supervisors and plant managers see the named comparison.
What if the best operator just gets the easy shifts?
The ranking is weighted by paid time and compares stops of the same type on the same line. Read it over three months or more, because person-level figures are noisy on short windows.
Do operators need to log in?
Yes. Each operator signs in on the Augos Operator Terminal with a PIN. Stops with nobody signed in fall into an unassigned row, which shows how much time still needs an owner.
Related
What is being benchmarked
The stages of a stop and the standards for planned work.
Stoppage Response Stages ›Planning Efficiency Loss and Operator Compliance ›See who sets the bar on your line.
Thirty minutes with a manufacturing specialist: how the best operator is found, and how big the gap is on a line like yours.
- Best operator per task
- Trainers and trainees, in minutes
- Prices in rand, ex VAT
A manufacturing specialist replies within two business hours to set a 30-minute call.