The Tool Utilization Question
Why this matters
Every shop over about five techs owns tools it cannot justify and does not know which ones they are. The owner has a feeling - "we never use that thing" - and the feeling is wrong about half the time, because the tool he is thinking of gets quietly used by one guy on Tuesdays and the one he forgot about has not moved in two years. Utilization is the number that replaces the feeling. But a single utilization percentage applied across every tool in the shop will get you to sell the one item that saves somebody's life, so the number has to be read per class, against the right denominator, with one class exempted entirely.
The denominator is the whole argument
Utilization is uses divided by opportunities. The mistake is using calendar days as the denominator for everything. A specialty tool that runs on every job of its type is at 100% of its real opportunity even if it sits 340 days a year.
| Tool class | Right denominator | What the number means | What a low number tells you |
|---|---|---|---|
| Everyday personal tools (drivers, hand tools, meters) | None worth tracking | Nothing useful | Do not measure these; measure availability instead |
| Shared shop tools (crib items several techs draw on) | Working days in the window | Checkout-days per working day | Genuine idleness, or nobody knows it exists |
| Specialty tools (bought for one job type) | Jobs of that type in the window | Attach rate | The tool is losing to another method, or nobody is trained |
| Insurance tools (rescue, retrieval, backup, emergency) | Not applicable | Nothing | Nothing. Near-zero use is the design intent |
That last row is the one that saves you from an expensive mistake. A confined-space retrieval winch, a spare submersible pump, a backup generator, a second set of tester leads: these exist because the day you need them you need them immediately. Selling one because it scored 0% is not a cost saving, it is buying a delay at the worst possible moment.
The rule, with its gates
A shared crib tool that is checked out on fewer than 1 working day in 10 over a rolling 12-week window goes on the review list, AND only if it is not in the insurance class. Review list, not sale list. The review is a five-minute conversation that decides which of four things is true, and three of the four are fixable without selling anything.
Units matter here: checkout-days per working day, on a rolling 12 weeks, per tool. Not hours, because most checkout logs record a day. Not a rolling year, because a 12-month window hides a tool that died six months ago inside a strong first half. Twelve weeks is long enough to survive a slow month and short enough to catch a change.
For a specialty tool the rule is different: attach rate below 60% of the jobs it was bought for, over at least 8 jobs of that type. Below 8 jobs the sample is too thin to act on, and you should extend the window instead of ruling.
How to instrument this without a barcode system
You do not need an asset system to answer the utilization question. Three methods, in order of how much they cost you:
The checkout log you already run. If tools leave the crib against a name, you already have the data. Count distinct days a tool appears in the log over the window. Cross-reference: the log has to be actually used, which is its own problem.
The sticker method. Put a strip of masking tape on the case and have whoever takes it write the date. Twelve weeks later you have a usage history in one glance and it cost you a roll of tape. This is the right method for a shop of three to eight that has no crib discipline yet. It fails on tools that live on a truck rather than in the shop.
The photo audit. Photograph the crib shelf on the same weekday every week for twelve weeks. Anything present in all twelve photos has not moved. This catches nothing about frequency and everything about total idleness, which for a first pass is most of the value.
Whatever you use, capture the tool's location too. Half of what looks like idleness is a tool that is on a truck and being used constantly by one person while the crib log shows nothing.
Worked example: reading a 12-week crib log
An eight-tech shop with a 40-item crib pulls its checkout log for a 12-week window. That window contains 60 working days, so the 1-in-10 threshold is 6 checkout-days. Four items are worth walking through:
- Item A, a coring rig: 34 checkout-days of 60, about 57%. Well above the gate, no action.
- Item B, a pipe-freezing kit: 9 checkout-days of 60, 15%. Above the 6-day threshold, so it does not flag. Note that 15% sounds terrible and passes the rule anyway, which is the point of having a rule instead of a reaction.
- Item C, a specialty cutting machine: 2 checkout-days of 60, about 3%. Below the threshold, flags for review.
- Item D, a confined-space retrieval winch: 0 checkout-days. Below the threshold, but the second gate excludes it: insurance class, no review, stays.
The review on Item C. It is a specialty tool, so the shared-tool denominator was the wrong lens in the first place. The right question is attach rate. The job-type it serves ran 5 times in the same 12 weeks, and the machine went out on 2 of them, so attach rate is 2 of 5, 40%. That is under the 60% gate but on only 5 jobs, below the 8-job minimum, so the honest call is to extend the window rather than rule. They pull a 24-week view: 11 jobs of the type, machine used on 4, attach rate about 36%. Now the sample clears the minimum and the rate is genuinely low.
What the other 7 jobs did instead. Three were subcontracted out. Four were done with a slower manual method by techs who did not know the machine was in the crib. So the tool is not underperforming - the shop's own dispatch and training are. The corrective action is a 30-minute check-out session for four techs and a line on the job template naming the machine, not a sale.
The confirmation. They re-pull the attach rate at 12 weeks after the training: 6 of 7 jobs of the type, about 86%, which clears the 60% gate comfortably. Had it come back at 40% again with everyone trained, the conclusion would have flipped: the techs are choosing the manual method for a reason, and the next step is to watch one job and find out what that reason is before selling anything.
For everyday tools, measure availability instead
The first row of the table says do not measure utilization on drivers, hand tools, and meters, and that deserves an alternative rather than a shrug. A tech's daily tools are at effectively 100% utilization by definition; counting their use tells you nothing you can act on. The question that matters for that class is the opposite one: how often is the tool a tech needs not working or not there?
Availability = the share of working days on which a tech started the day with a complete, functioning kit. Capture it with one question at the end of the day, asked of two techs at random: did anything in your kit fail, go missing, or run out today? A yes is one availability miss.
Read it as misses per tech per month. A shop running under 1 miss per tech per month has a healthy kit standard. Above about 2 per tech per month, look at three causes in order: consumables not restocked (most common by a wide margin), a replacement cycle running too long so tools are failing of old age, and kit standards that were never actually defined so two techs carry different things and neither is complete.
The reason to separate these metrics rather than force one number: utilization is an argument about what to own, and availability is an argument about what to spend on upkeep. A shop with a great utilization picture and terrible availability is buying the right tools and letting them rot.
Three ways utilization lies
It counts checkouts, not work. A tool signed out and returned unused looks identical to a tool that ran all day. If a class of tool tends to go out speculatively (staged on the truck in case), utilization overstates it. The fix is to sample: ask on three random returns whether the tool was actually used.
It punishes seasonality. A twelve-week window that lands entirely inside your slow season will flag half the crib. Either run the window across a season boundary or accept that the flag list from a January pull needs a July re-check before anyone acts.
It cannot see the tool that enables the sale. Some tools are used rarely but their presence is why you can quote a job type at all. If losing the tool means declining the work, the tool's utilization is beside the point and the right measure is the revenue-earning job count it gates. Treat it like an insurance tool for review purposes and re-decide only if the job type itself disappears.
What to do with a genuine flag
The four outcomes, in the order to test them:
- Nobody knows it exists. Most common in shops that grew past about six techs. Fix with a crib walkthrough and a line in the job template. Cheapest fix available.
- Nobody is trained on it. Second most common. A tool a tech has never run is a tool he will not choose under time pressure, whatever it saves.
- It genuinely loses to another method. Time both on the same task before you believe this. If it loses, sell it and record why, so the next person who wants to buy one finds the note.
- The job type went away. The only case where selling is the obvious first answer, and it is the rarest of the four.
How to know you got this right
Six months after your first pass, three things should be true. The flag list should be shorter, and the items that dropped off it should have dropped because usage rose, not because you sold them. Every item you did sell should have a one-line reason recorded against it. And nothing in the insurance class should have moved at all. If you sold something from that class and have not needed it yet, you have not been proven right - you have been lucky for six months.
References
- Manufacturer documentation for expected service life and duty ratings on specialty equipment
- Trade-standard practice for tool crib checkout logging in small contracting shops
- See related: How to Decide What a Specialty Tool Has to Earn, The Tool Nobody Uses and What It Cost, How to Build a Tool Inventory That Stays Current, The Tool Check-Out System That Actually Gets Used