Stoppage Response Stages
The Stoppage Response Stages are the five sequential, separately timed stages of responding to an unplanned stop: Operator Evaluation, Operator Resolution, Technician Response, Active Wrench and Inactive Wrench. Each stage has an owner. Timed apart, they show that most of a long repair is usually people noticing, calling and waiting, and only a few minutes of it is the repair.
What the stages are
A forty-minute breakdown is reported by standard OEE as forty minutes of Availability loss with a reason code. Inside it, a chain of human response ran before and after the repair: someone noticed the stop, tried to clear it, decided to call for help, waited for a technician to arrive, waited again for a part. Augos Assist, the operator and technician accountability layer, timestamps each handover from the Augos Operator Terminal, so the single forty-minute number becomes five timed stages, each owned by the function that can shorten it. The stages apply to unplanned stops only. Planned stops are measured against a standard instead; see Planning Efficiency Loss and Operator Discipline.
Source: Augos Factory method, Glossary: Stoppage Response Stages
The five stages
| Stage | What is timed | From → to | Owner |
|---|---|---|---|
| 1 · Operator Evaluation | Assessing the fault and deciding to call a technician | Machine stops → operator escalates | Operator |
| 2 · Operator Resolution | The operator clearing the stop alone | Machine stops → machine running, no technician called | Operator |
| 3 · Technician Response | Getting a technician to the machine | Operator escalates → technician arrives | Technician |
| 4 · Active Wrench | Hands-on repair | Technician arrives → repair complete, or waiting begins | Technician |
| 5 · Inactive Wrench | Technician on the machine but not working: a part, a permit, a second pair of hands | Waiting begins → waiting ends | Shared (Technician + supply) |
Stages 1 and 2 are alternatives rather than a sequence: a stop is either cleared by the operator (Resolution) or escalated (Evaluation, then 3 to 5). Stage 5 is shared because the technician is present but the storeroom, the permit office or a colleague holds the clock.
Two paths through a stop
Every unplanned stop takes one of two paths. On the first, the operator clears it alone: the whole stop is Operator Resolution, and what matters is how quickly it was cleared against the line's own best for that cause. On the second, the operator decides the fault needs a technician: the time up to that decision is Operator Evaluation, then Technician Response runs until the technician arrives, then Active Wrench, with Inactive Wrench cut out wherever the repair stalls. Which path a stop takes is itself a measure. A line where operators clear nine stops in ten without help is a different line from one where every fault waits for maintenance, and the Downtime Report shows the share directly as operator intervention.
A month, split into stages
The beverage packaging line used across the method pages logged 408 breakdowns in one month in 2026, 97.7 hours in 720 paid hours. The stop count, the durations and the split between stops cleared alone and stops over fifteen minutes are the line's measured figures. The division of the escalated stops into stages is modelled, because this line does not yet run Augos Assist; it is shown to make the arithmetic concrete, with the shares an Assist deployment would time directly.
| Stage | Stops | Hours | Median per stop | Owner |
|---|---|---|---|---|
| Operator Resolution (cleared alone, up to 15 min) | 312 | 33.2 | 5.2 min | Operator |
| Operator Evaluation (escalated stops, time to call) | 96 | 3.2 | 2 min | Operator |
| Technician Response | 96 | 12.8 | 8 min | Technician |
| Active Wrench | 96 | 43.9 | 22 min | Technician |
| Inactive Wrench (two stops waiting for parts) | 2 | 4.5 | 135 min | Shared |
| Breakdowns in total | 408 | 97.7 | 6.7 min |
Three things fall out of the split. First, three stops in four never reached a technician: 312 of 408 were cleared on the line, and their 33 hours are an operator coaching question, benchmarked against the operator who clears that fault fastest. Second, of the 64 hours in the 96 escalated stops, about 16 are response and waiting before any work starts; shortening the call and the walk recovers more than any spanner. Third, two stops waiting for parts cost 4.5 hours between them, more than all 96 evaluations combined; that is a storeroom conversation, and without the stage split it would read as two slow repairs.
Source: stop counts and durations from the Augos platform, Unplanned Stoppage analysis, one anonymised beverage line, a 30-day month in 2026; stage division modelled for illustration (Assist not yet deployed on this line)
Typical, Recommended, Recoverable
Each stage of each fault code is benchmarked against the plant's own history. Typical Time is the median duration of that stage for that cause: what usually happens. Recommended Time is the fastest quarter: what the plant has already shown is achievable. Recoverable Time is the difference. If the typical Technician Response on a filling fault is 8 minutes and the fastest quarter is 4, every occurrence carries 4 recoverable minutes, and the target is one the plant's own technicians have already met. A top-down target such as "MTTR should be 30 minutes" is an opinion; a Recommended Time is proof.
Source: Augos Factory method, Glossary: Typical Time, Recommended Time, Recoverable Time
MTBF and MTTR
Two reliability figures sit above the stages. MTBF, mean time between failures, is running time divided by the number of stops: higher is better. MTTR, mean time to repair, is total stop time divided by the number of stops: lower is better. They answer different questions, and a line can post a healthy MTBF with a poor MTTR or the reverse. The catch with MTTR is this page's whole point: most of it is response and waiting. The stages resolve a high MTTR into the stage that caused it, so "maintenance is too slow" becomes "Inactive Wrench is 40% of our long stops", which points at the storeroom rather than the fitter. On the beverage line, MTBF for declared stops was 81 minutes and MTTR 26 minutes in the month.
Medians, small samples, immutable timestamps
Three rules keep the stages fair. Stage durations are reported as medians, so one catastrophic stop, a machine down for days waiting for a shipped part, does not make every technician look slow; MTBF and MTTR use means because they describe the whole interval. A stage with fewer than ten events in the window is flagged as statistically fragile and not graded. And every stage timestamp is generated by the system and cannot be edited by any role afterwards, so when the operator says the call went out at ten and the technician says half past, the record settles it.
Source: Augos Factory method, outlier and median rules; Glossary: Outlier (IQR method)
What needs Augos Assist
The Stoppage Response Stages need Augos Assist, an optional extra. They need the operator interacting with the Augos Operator Terminal at each handover, which is what Augos Assist instruments. Without Assist, an unplanned stop is one number: the platform still reports total unplanned downtime and the loss breakdown by cause, but the stage columns are empty because nothing records the handovers.
Source: Augos Factory method, Glossary: tiers
Common questions
What are the Stoppage Response Stages?
The five sequential, separately timed stages of responding to an unplanned stop: Operator Evaluation, Operator Resolution, Technician Response, Active Wrench and Inactive Wrench. Each has an owner, and together they show where the minutes of a long stop actually went.
Is a slow repair a skills problem?
Rarely. Most of a long stop is response and waiting: noticing, deciding to call, the technician arriving, a part on order. Splitting Active from Inactive Wrench shows whether the loss is in the workshop or the storeroom.
Do we need more technicians to cut downtime?
Usually not. On the line above, three stops in four were cleared by operators without a technician, and a quarter of the hours in the escalated stops was response and waiting before any work started.
Why medians and not averages?
So that one extreme stop does not define a person or a shift. Means are used only for MTBF and MTTR, which describe the whole interval rather than a typical response.
Can a technician change the timestamps afterwards?
No. Stage timestamps are system-generated and immutable across all roles, which is what makes them usable in a dispute between operations and maintenance, or with a contractor's service level.