The Tool Nobody Uses and What It Cost

Why this matters

The purchase is the smallest part of what an idle tool costs you. The real bill is the work that kept getting done the slow way for two years while the fast way sat on a shelf, plus the rental you paid twice for a machine you already owned, plus the bay it occupied. Owners look at an unused tool and see a bad buying decision. Usually it was a fine buying decision followed by a failure to finish - the tool arrived and nobody was ever put on it.

What follows is one shop's investigation, worked through the way they actually worked it, including the two hypotheses that were wrong.

The signal

An eleven-tech shop runs its annual tool audit. One line stands out: a specialty machine in the crib with a checkout log showing three entries in twenty-six months.

The owner's first reaction is the common one, and it is wrong: sell it. It was a mistake, cut the loss, move on. His crib keeper pushes back with one question - what job type did we buy it for, and how often did that job type run? Nobody knows offhand. That question is what turns a disposal into an investigation.

Hypothesis 1: the work went away

The cheapest thing to check, so check it first. If the job type the machine serves has dried up, everything else is moot and the tool goes.

They pull completed jobs for the same twenty-six months and count jobs of that type: 22 of them. Steady across the period, no trend down.

Eliminated. The work is there. Machine used on 3 of 22, an attach rate of about 14%. That number reframes the whole problem: this is not an unused tool, it is a tool that loses to something 6 times out of 7.

Hypothesis 2: it is broken

Second cheapest. A machine that failed once and was quietly shelved will look identical to an unwanted one in a checkout log, and nobody will have reported it, because the tech who found it broken solved his own problem another way and moved on.

They run it on the bench. It starts, runs, holds settings, no faults. Consumables are worn but present.

Eliminated, with one useful finding attached: the consumable that the machine needs had not been restocked in over a year. Not enough to stop it working today, but a tech reaching for it eight months ago might have found nothing on the shelf and given up. Note that and keep going.

Hypothesis 3: it is slower than the manual method

This is the hypothesis most shops assume without testing, and it is the one that most deserves a test, because if it is true the correct action is to sell the machine and stop feeling bad about it.

They time both methods on the same task, same operator, back to back. Machine method: 1.9 hours. Manual method: 2.6 hours. Run twice to make sure the first machine run was not a fluke, and the second run comes in slightly faster as the operator settles.

Eliminated, but the test itself nearly misled them, and the way it nearly did is worth carrying. The manual method was timed by the crib keeper, who is the fastest person in the shop at that particular manual task, having done it for years. The machine was run by the same person on his second-ever go. So the comparison was the shop's best manual operator against a near-novice machine operator, and the machine still won. A typical tech doing the manual method would have been slower than 2.6 hours, which means 0.7 hour is the conservative edge of the real gap, not the middle of it. When you time two methods, write down who ran each one and how experienced they are at it; a timing test with that field missing can be read whichever way you already wanted to read it.

The machine is about 0.7 hour faster per job, roughly 27% off the manual time. Against the 22 jobs in the window, the shop had been running the slow method on 19 of them. That is 19 jobs at 0.7 hour, about 13.3 hours of extra field time over twenty-six months, a little over half an hour a month. Real, but notice how unspectacular it looks spread across two years. This is exactly why idle tools survive audits: the loss never shows up as a single visible event.

What was actually true

They ask a plain question at the next huddle: who here has run this machine? One hand goes up out of eleven, the crib keeper's. Checking back through the records, exactly one other person had ever run it, and he left the shop four months ago. So at the time of the audit, one tech in eleven could operate it, and he accounted for all three checkouts himself.

That is the answer. The machine was never a purchase problem. It was a purchase that never finished. It arrived, the tech who advocated for it learned it, and the shop assumed that constituted adoption.

Two contributing factors made it stick:

Nobody knew it existed. In interviews, 5 of the 10 current techs who had never run it did not know the shop owned one. It lived in a case at the back of the crib and it was not on the job template for the job type it served.

The consumable was out. The one tech who could run it had reached for it, found the consumable shelf empty on at least one occasion, and used the manual method instead. Nobody logged that as a stockout because he did not think of it as one.

The cost, counted properly

The owner had been thinking of the cost as the purchase. Here is what it actually was over twenty-six months:

• 13.3 hours of avoidable field time, from running the slow method on 19 jobs at 0.7 hour each.
• Two rentals of the same class of machine. The crib keeper found them in the records while pulling job history. Two different leads, on two different jobs, rented a machine the shop already owned because they did not know it was there. This is the item that stings, and it is common: an unknown tool does not just sit, it gets re-acquired.
• A bay position in the crib occupied by a case nobody opened, in a shop that was already short on shelf space and had been talking about renting storage.
• The purchase itself, which is real but is the one cost that is already sunk and cannot be recovered by any decision made now.

Note which of these are recoverable. The rentals and the field hours stop the day the crew can use the machine. The purchase does not come back whatever you decide, so it should carry no weight in the decision - a point owners routinely get backwards, holding onto a tool because of what it cost or dumping it for the same reason.

The fix, and what it took

A 40-minute check-out session on the machine, run twice to catch both shifts, covering seven techs. Total trainer time about 1.3 hours plus setup.

Three other changes, none of them large:

1. The machine went on the job template for the job type, as a named line. That single line does more than any reminder, because it appears at the moment of decision rather than in a meeting.
2. The consumable went on the crib's reorder list with a minimum quantity, so an empty shelf triggers an order rather than a workaround.
3. The machine moved from the back of the crib to a position at eye level near the door. Findability is physical before it is digital.

How they confirmed it stuck

They re-pulled the attach rate at twelve months: 17 of 19 jobs of the type used the machine, about 89%, against 14% before. The two exceptions were both access-limited jobs where the machine could not be brought to the work, which is a legitimate exception and worth knowing.

They also checked the thing that would have proved the fix cosmetic: rentals of that machine class in the following twelve months came back at zero.

And they added a permanent question to the annual audit that costs nothing and would have caught this two years earlier: for every crib item, how many current techs have run it? A tool that only one person can operate is one resignation away from being idle, whatever its usage log says today.

The four reasons a tool goes idle, and how to tell them apart

Having worked one case, the general shape is worth carrying. There are four causes, they need different fixes, and the tests that separate them are cheap:

Every one of these fixes is cheap, which is the uncomfortable part: the cost was never the fix, it was the months nobody ran the test. Two of them are worth singling out. The last row is the most common cause by a distance, and the first row is the only one where selling is the right answer rather than a surrender.

Run them in that order. Each is cheap, and each one you eliminate makes the next more likely, which is what keeps the investigation honest rather than jumping straight to the conclusion you already had.

The single highest-yield question in the whole set is the last one. In a shop under fifteen techs, an idle tool is far more often an untrained crew than a bad purchase, because the shop is small enough that one enthusiastic tech can drive a purchase and large enough that his knowledge does not automatically spread.

The pattern to watch for going forward

An idle tool almost always has a birth defect: the purchase was approved and the adoption was never scheduled. The tool arrives, the person who wanted it uses it, and the shop books that as done.

The correction is to treat a tool purchase as unfinished until three things are true: more than one current tech can operate it, it appears by name wherever the work is planned, and its consumables are on the reorder list. Until all three, you own an asset that depends on one person's memory, and memory leaves the shop when he does.

References

• Trade-standard practice for annual tool and equipment audits in small contracting shops
• Manufacturer documentation for consumable specifications and operator competency requirements
• See related: The Tool Utilization Question, How to Train a Tech on a Tool Before First Use, The Tool and Equipment Audit Checklist, How to Decide What a Specialty Tool Has to Earn