Abstract
Wrench time is routinely presented to executives as a labour productivity figure and routinely read as evidence that maintenance crews are underworked. Both readings are wrong. Tool time measures how much of a tradesman's day the organization succeeded in making executable, which makes it a readiness metric wearing a productivity costume. This paper argues for reclassifying it, shows what the number actually varies with, and describes what changes when the measurement is filed in the right column.
Twenty-six hours, then eight
The same heavy lift took twenty-six hours on day shift with more than twenty people on it, and eight hours on graveyard with three of us. Me, another millwright, and a welder.
An outside contractor crew had been called in for the job. They walked in to find it already running.
People call bullshit on that story, and I understand why, but there is nothing clever in it and no heroics. I was not trying to prove anything. I wanted my plant running and I did not want to be bored all night.
The difference was not effort and it was not skill. It was this. The job was planned ahead instead of organized the morning of. One man was on signals instead of four people shouting. The area was cleared before anyone needed it. And the lockout was handled by one guy before the crew showed up, instead of twenty people standing in a line waiting for a permit.
Every one of those is a readiness item. Not one of them is a labour item.
Now do the arithmetic that a productivity metric would do. Twenty-plus people for twenty-six hours against three people for eight. If you file that under labour, the day crew looks like it was loafing at roughly a twentieth of the rate, and the obvious management response is pressure, supervision and monitoring.
That response would have been aimed at the wrong thing entirely, and it would have made the next one worse.
What the number actually measures
A tradesman is waiting because a part did not arrive. Because a permit is unsigned. Because an isolation is not confirmed. Because a crane is committed elsewhere. Because a drawing is obsolete. Because the job as written does not match the machine in front of him.
Every one of those is a readiness gate that failed upstream, and none of them is inside his control.
So a plant reporting low tool time has not measured its workforce. It has measured its planning, its stores, its permit system and its scheduling, and then attributed the result to the only people who happened to be physically present when the stopwatch came out.
That attribution is not a rounding error. It points the entire improvement effort at the one group in the chain who cannot fix any of it.
The blocked crew and the stuck crew
There are two different ways a crew ends up not turning wrenches, and they look identical on a report.
A blocked crew is waiting on something external. The part, the permit, the crane. That is the readiness failure described above and it is the majority of it.
A stuck crew is a different animal, and it is invisible to every system I have ever seen.
I came across a crew once beating their heads against a bearing buried up inside a drum. They were going nowhere and were convinced there was no way at it. I told them to cut the side out of the drum, pull it, change the bearing, and weld it back up.
That was the whole intervention. One sentence, and they were moving.
They had not done it because it had never occurred to them that they were allowed to. Not because they were not good enough, but because they were too close to the problem to see past the obvious approach, and because the unlock was permission to do the destructive-looking thing.
The other half of the finding came on a down day, on a deck, with a crew changing out a set of rails. Big, heavy, worn-out things. I came up at coffee and they had been there for hours and done nothing. Nothing. I asked what was going on and they told me the rails were too heavy, they could not do it, so they had just sat there and waited.
They did not call anybody. They did not try another way. They did not flag it. They downed tools and parked it while the window bled.
A crew that is stuck will not come and find you. They will stop, sometimes for hours, and the cost of a crew quietly parked is invisible until you walk up on it.
Which is the operational answer, and it is not a metric at all. On a down day you have to be on your feet, running a circuit through your crews. You carry the fresh eyes to the crew that is stuck, because they will not carry the problem to you.
Half the time it is not laziness. They are stuck and too close to it. The other half of the time they are blocked, and nothing in their power will change that either.
Why the misreading is expensive
A plant that believes it has a labour problem responds with supervision, monitoring and pressure. All three cost money. None addresses a single cause. And all three damage the relationship with the people whose informal knowledge is the plant's actual diagnostic capability.
The misreading also inverts the character of the workaround, which is the part that does real long-term harm.
A tradesman who keeps a private stock of bearings in his toolbox is not saving himself effort. He is getting the line running sooner, despite a system that would otherwise have him waiting three days on a two hundred dollar part. Same behaviour, and the word you choose decides whether he reads as lazy or as competent.
Recording that as an inventory discipline problem rather than as a readiness signal throws away the most accurate data the plant owns about where its formal system fails. The contents of the unofficial inventory are a precise list of the parts the official chain cannot deliver in time. It is free diagnostic information, and most plants respond to it with a memo.
What good looks like, and why it is not a leaner crew
The temptation, once tool time improves, is to bank the difference in headcount. That is available and it is usually the wrong prize.
I rebuilt the PM routing in a wood room once and took it from four men walking all week to one man in a day, and the same approach spread to the other departments. The saving was real. But the thing worth having was not the three men. It was that the plant could then all but guarantee that everything worth looking at got looked at at least once a week.
Efficiency was the vehicle. Daily coverage and early detection were the destination. A plant that converts every readiness gain into a smaller crew has spent the gain and kept the exposure.
Reclassifying the metric
Three changes, none of which require new data collection.
Report tool time next to the four readiness gates, not next to labour cost. Where it sits on the page determines what people do about it.
Break the non-tool time into causes rather than reporting one figure. A report saying the crew is at forty percent is only actionable as blame. A report saying that thirty percent of the lost time is unconfirmed isolation is actionable on Monday.
Separate blocked from stuck. They have completely different fixes. Blocked is a planning problem you solve upstream. Stuck is solved by somebody walking the floor, and no dashboard will ever surface it.
The automation opportunity here is narrow and honest. Attributing waiting time to a cause is a classification task over records the plant already holds, and the output is a ranked list of gates rather than a judgement about a person. It is also the kind of work where the staged adoption literature applies directly. Crespo-Márquez and Gómez Fernández find agentic AI evidence concentrated in technical phases while managerial phases such as KPI alignment remain largely unexplored, and tool-time attribution sits exactly in that unexplored gap.
Conclusion
Wrench time is one of the few maintenance metrics that is both widely collected and almost universally misfiled.
Moving it from the labour column to the readiness column costs nothing, changes no data collection, and redirects every intervention it triggers away from the people who are waiting and toward the systems that are making them wait.
The lift took eight hours with three people because the permit was done, the area was clear, and one man was on signals. It took twenty-six with twenty because none of that was true. Both crews worked. Only one of them was allowed to.
References
- Crespo-Márquez & Gómez Fernández, “Agentic AI for maintenance management: a process-centric review and a staged framework for industrial adoption,” Computers & Industrial Engineering, 2026. https://www.sciencedirect.com/science/article/pii/S0360835226004869
- Prometheus Group, “From Hours to Minutes: How AI Is Transforming Maintenance Planning and Scheduling,” June 26, 2026. https://www.prometheusgroup.com/resources/posts/from-hours-to-minutes-how-ai-is-transforming-maintenance-planning-and-scheduling
- IBM Research (O’Donncha et al.), “Evidence-Driven Reasoning for Industrial Maintenance Using Heterogeneous Data,” ACL 2026. https://research.ibm.com/publications/evidencedriven-reasoning-for-industrial-maintenance-using-heterogeneous-data