The Hidden Cost of Under-Training a Tech

Why this matters

Nobody in a small shop has ever looked at a profit and loss statement and seen a line called "under-training." That is the whole problem. The cost is real, it is large, and it is distributed across six other lines that each look like a different problem: rework, low throughput, parts consumption, senior tech time, missed work, and turnover. Because it never shows up where it was caused, the training line stays at zero and the six symptom lines get managed one at a time, forever, by people who have no idea they are all the same expense.

Under-training is a capacity problem wearing six costumes

The single sharpest way to hold this: an under-trained tech does not cost you a training budget, they cost you a fraction of a tech. Two people on the same payroll, the same truck, and the same paid hours can produce output that differs by a third. That difference is capacity you are paying for and not receiving, and it never appears anywhere you would look for it.

The six places it lands:

1. Diagnostic time inflation. The largest and most invisible. An under-trained tech takes longer on the same call, and nobody notices, because the job got done and the customer was satisfied. There is no incident, no complaint, no line item. There is just a schedule that fits fewer jobs than it should.

2. Rework and callbacks. The one shops do see, though usually only as a count and rarely as hours. Every callback is a full drive plus a visit that cannot be invoiced, and the customer relationship takes a hit that does not show up in any ledger.

3. Parts consumption. A tech who cannot confidently isolate a fault replaces parts to find it. Watch the parts-to-labor ratio per tech. A ratio that runs meaningfully above the shop's median on the same job mix is a diagnostic signal before it is a cost signal.

4. The interruption tax. Every call from the field to the lead tech costs two people's time, and it costs the lead more than the minutes, because it breaks their own work. This is a genuinely tricky one to read, since early in a tech's development calls going up is a good sign. It is the tech who is two years in and still calling four times a week that you are measuring.

5. Work not seen and not sold. The under-trained tech services the complaint and walks past the failing component next to it, not out of laziness but because they do not recognize it. Nobody can count what was never noticed, which is why this line is usually left out entirely and is often the biggest one.

6. Turnover. People leave jobs where they feel incompetent, and they leave faster than people who are underpaid. The under-trained tech is more likely to resign, and the replacement starts the whole cycle again at the bottom.

Why the cost stays invisible

Three attribution failures keep this hidden even in shops that watch their numbers:

It never appears on the line that caused it. The callback lands under warranty or rework. The parts land under materials. The lead tech's lost hours land nowhere at all, because senior techs rarely code the time they spend answering the phone. Nothing routes back to the training line.

It looks like a personality. "He's just slower." "She's careful." Slow is an outcome, not an explanation, and the difference between careful and uncertain is precisely the thing a competency check is designed to distinguish. Labeling it as temperament ends the inquiry.

The comparison is never made. Most shops do not compute average hours per completed call per tech, so nobody knows the spread exists. The moment you compute it on the same job mix, the conversation changes from "he's fine" to "why is he taking 40 percent longer."

Worked example: two techs, identical paid hours

Both techs work the same job mix, same territory, same truck stock. Assume roughly 4 dispatched jobs a day across 230 working days, giving each about 1,472 paid field hours a year.

The trained tech. Averages 1.6 hours per completed call. That is 1,472 divided by 1.6, or about 920 completed calls a year. Callback rate 8 percent, so about 74 return visits. Calls the lead tech about 0.5 times a week.

The under-trained tech. Averages 2.3 hours per completed call. That is 1,472 divided by 2.3, or about 640 completed calls a year, which is roughly 70 percent of the trained tech's throughput on identical paid hours. Callback rate 19 percent, so about 122 return visits on fewer jobs. Calls the lead tech about 4 times a week.

Read the headline first: 280 fewer completed calls a year, a 30 percent capacity loss, on a fully paid tech. That is the number nobody sees, because it never generates an event. Nothing goes wrong. There is just less.

Now the secondary lines, all of them additional to that gap:

• Return visits: 122 against 74 is 48 extra return visits. At about 1.5 unbillable hours each that is roughly 72 hours consumed, and note it is consumed out of the same 1,472 hours, which is part of why the completed-call figure is already so low.
• Interruption tax on the lead: 4 calls a week at about 12 minutes each is 0.8 hours a week; across 46 working weeks that is about 37 hours of the lead tech's year. The trained tech's 0.5 calls a week comes to about 4.6 hours. The difference is roughly 32 hours of senior capacity spent on one person, which is most of a working week of your most expensive resource.
• Parts: a parts-to-labor ratio of 1.4 against a shop median of 0.9 is about 56 percent more parts consumed per labor hour, and since the under-trained tech completes fewer jobs with those parts, the per-completed-job figure is worse still.
• Work not seen: uncounted, and structurally uncountable, but it moves in the same direction as everything above.

Now price the fix in the same units. Suppose 40 hours of structured training across a year - bench-rig practice, paired dispatch, cold demonstrations - moves the under-trained tech from 2.3 hours per call to 1.8. Pay for the training out of field time honestly: 1,472 minus 40 leaves 1,432 field hours. At 1.8 hours per call that is about 796 completed calls, up from 640, a net gain of 156 completed calls after fully absorbing the training hours.

That is roughly 4 additional completed calls for every hour of training invested, before counting the callback reduction, the parts normalization, or the 32 hours of lead-tech time returned. Run this on your own two numbers - average hours per completed call and callback rate, per tech, on the same job mix - before you believe any of it, because your spread may be narrower or wider. The point is not the specific figures, it is that the arithmetic is available to any shop that keeps job records, and almost nobody runs it.

How to size this for your own shop, in an afternoon

Four measurements, all from records you already have:

1. Average hours per completed call, per tech, filtered to one common job type so the mix is comparable. The spread between your best and worst is the capacity you are missing.
2. Callback rate per tech, as a share of that tech's jobs, not as a raw count, or your busiest tech will always look worst.
3. Parts-to-labor ratio per tech over a quarter. Compare to the shop median, not to a target.
4. A one-week tally of calls from the field to the lead, marked with who called. One week is enough to see the shape.

Do not attempt a precise total. The purpose is to establish that the gap is large, which it almost always is, and to identify which tech and which skill to attack first. A rough number you act on beats a precise number you spend a month building.

What changes the answer

• A tech under six months in. The whole framing is wrong for them. High hours per call, frequent calls to the lead, and cautious parts behavior are all correct and expected during ramp-up, and treating them as under-training costs is how a shop panics on a new hire who is doing fine. Apply this reading from roughly the one-year mark.
• Job mix contamination. If one tech gets all the difficult or unfamiliar work, their hours per call will be higher for reasons that have nothing to do with skill. Filter to a common job type before comparing, every time.
• A shop where the schedule is not full. If there is not enough work to fill the trained tech's 920 calls, then the throughput gap is not costing you revenue today, it is costing you the ability to grow without hiring. Real, but a different argument, and worth being honest about rather than overstating.
• A tech who is slow because they are thorough. They exist. The distinguishing test is the callback rate: slow and low-callback is a pace question and possibly a pricing question. Slow and high-callback is a competence question. Do not confuse the two, because coaching the thorough tech to hurry is how you buy their callbacks.
• Where the constraint is truck stock, not skill. A tech who cannot complete calls because parts are not on the truck reads identically to an under-trained tech in the throughput number. Sort the incompletes by cause before concluding anything.

The three false economies

"We can't afford to train right now." The comparison being made is training hours against zero. The real comparison is training hours against the capacity gap, and in the worked example above 40 hours was set against a 280-call shortfall. The shop that cannot afford to train is generally the shop that most needs to, because the capacity loss is what is making things tight.

"We'll train them when things slow down." Deferral to a window that arrives with its own cash problem. This is why the slow-season plan has to be built and dated in advance rather than intended.

"Why train them just so they leave." The two flaws in this are that the untrained tech leaves too, and leaves sooner, because feeling incompetent every day is a stronger push than pay; and that a replacement runs at a fraction of full output for months, so a departure costs a multiple of the training hours that might have prevented it. The version of this worry worth taking seriously is different: train people and then give them nowhere to go. That is a progression problem, not a training problem, and the answer is a visible path with sign-offs attached, not less training.

Verifying you are reading this correctly

• Recompute after a change and check the spread narrowed, not just that the average moved. An average that improved because one strong tech got stronger has told you nothing about the under-trained one.
• Check the callback rate and the hours-per-call moved together. Hours per call falling while callbacks rise means you taught speed, not competence, and you have made things worse in a way that will take a quarter to show up.
• Check the lead tech's recovered hours actually went somewhere. If the interruption tax fell by 32 hours and the lead's output did not change, the hours were absorbed rather than recovered, and the business case is weaker than you think.

References

• U.S. Bureau of Labor Statistics, job openings and labor turnover data on separations in construction and repair occupations
• U.S. Small Business Administration, workforce productivity and retention guidance for small employers
• Association for Talent Development, research on the relationship between structured training and employee retention
• See related: Measuring Whether Your Training Is Working, Build a Skills Matrix: Who Can Do What, How to Train During the Slow Season on Purpose, How to Decide When Someone Is Ready to Work Alone