When to Send Someone to Outside Training

Why this matters

Outside training is the most visible training spend a shop makes and the least measured. A tech comes back with a binder and a certificate, everybody feels good, and six months later nothing about how the shop works has changed. Meanwhile the capability that actually needed building was one a senior could have taught in four afternoons.

The reverse error is just as expensive and harder to see: a shop that refuses to send anybody out propagates one person's habits, including the wrong ones, through every tech who ever works there. Nobody in that shop can tell you which of their standard practices are correct and which are just what the owner has always done.

The decision is not about budget. It is about whether the capability exists inside your building at the standard you need, and whether your constrained people have the hours to move it.

The question that decides it

Ask it in this exact form: can somebody here teach this, to the standard we need, with hours we actually have?

All three have to be true for in-house to be the right call. Shops get this wrong by checking only the first - somebody here knows how, so we will teach it here - then discovering that the person who knows how is the person the shop cannot spare.

Which condition failed tells you what kind of outside training to buy:

• Nobody here knows it points to instruction: a class, a school, a manufacturer session.
• Somebody knows it but not to the standard points to validation: a certification or a proctored assessment that produces a credential a third party will accept.
• Somebody knows it to the standard but has no hours points to time-shifting: buy the instruction, keep the assessment in house.

That third case is the one shops handle worst. They cancel the training rather than recognizing that they can buy the teaching and still do the signing off themselves.

Four conditions that make outside training clearly right

A credential is required and only an outside body can issue it. Licensing, refrigerant handling, lift or rigging operation, confined space, respirator fit testing, lead-safe work practices. There is no build option. Attempting one is not thrift, it is exposure.

The capability is genuinely absent and the gap is durable. Your skills gap audit shows a task family with zero qualified people and recurring demand. You cannot teach yourself out of zero.

The manufacturer holds information you cannot get otherwise. Factory sessions on a specific equipment generation, warranty procedures, support tiers requiring a certified tech on staff. Some is genuinely gated and some is marketing dressed as training; the test is whether the session unlocks something you cannot otherwise obtain.

Your internal practice needs an outside check. If your shop has never had anyone trained anywhere else, you have no way to know where your standard practice has drifted. Sending one person out every year or two is calibration, and the value comes back in the teach-back, not the certificate.

Four conditions that make it clearly wrong

The capability exists in-house and the constraint is only motivation. Sending somebody out will not fix an unassigned training function. It buys three days of enthusiasm and changes nothing structural.

Nobody has defined what should be different afterward. "General professional development" produces no change and no way to detect that it produced none. If you cannot state the task the person should be able to perform on their return, do not buy the seat.

The class teaches theory the shop will never practice. A session on equipment you do not service and are not planning to service is entertainment. It may still be worth it for retention reasons, but call it what it is and budget it as such.

It is being used as a substitute for pay. Techs read that correctly, and the person you were trying to retain now knows what the shop would not do instead.

The options, compared

The cost that is not on the invoice

Register the full bill before you decide, in hours, because the fee is usually the smallest line.

A three-day session is not three days. It is three class days plus travel on either end, so budget four days off the board per attendee. Then add the hole: the calls that person would have run get absorbed or pushed, and in a busy stretch pushed calls become lost calls.

Then add the part nobody budgets, which is the return. Training that does not get applied within about the first month decays fast, so the real bill includes the field time to get the new capability onto real jobs while it is still fresh. If your board has no room for that in the four weeks after the class, you have bought a certificate. The honest comparison against in-house is therefore not fee-versus-free, it is total hours consumed, weighted by whose hours they are.

Choosing who goes, and what they owe on return

Send the person who will still be here and who can teach. Those two filters matter more than who wants to go most.

Attach a teach-back obligation before they leave, in writing, with a date:

• Within two weeks of returning, a session for the crew: what changed, what we do differently now, and what we were doing wrong.
• A one-page written procedure for anything that changes shop practice, filed where procedures live.
• Their materials go into the shop's reference, not their truck.

Without this, one person's training stays one person's training and you have converted a shop capability purchase into an individual benefit. The teach-back is also the only reliable way to find out whether the class was any good, because a person who has to explain it discovers within ten minutes which parts they did not absorb. Set the expectation that a poor class gets reported as a poor class; shops keep sending people to a session for years because nobody wants to say the trip was a waste.

On repayment agreements

Shops sending someone to expensive certification often want protection against that person leaving immediately after. Agreements that require an employee to repay training cost if they leave within a set period do exist and are used in the trades.

Understand the shape before you write one. Enforceability varies substantially by state, and these provisions have drawn regulatory and legal scrutiny in recent years, particularly where the repayment amount is not tied to actual cost, where the covered training is generic rather than specialized, or where the repayment operates to deduct from final wages, which is separately restricted in many states. If you want one, have counsel in your state draft it, tie it to documented actual cost, prorate it down over the covered period, and never treat it as a wage deduction without checking your state's rules. Do not copy one from another state's shop.

The more reliable retention move costs nothing: a tech who comes back and immediately gets to do the new work is far less likely to shop the credential elsewhere than one who returns to the same calls they were running before.

A worked example: one gap, two paths

Illustrative, and the method is the point.

A five-tech shop has a task family with recurring demand of about 21 instances a year and exactly one qualified person: the owner, who is also the estimator and the permit filer. The audit says build, and the question is how.

Path A, in-house. The owner teaches two techs. Preparation and a bench session, about 6 hours of owner time. Then five supervised field instances per tech at roughly 2 hours of overlapping owner attention each, so 10 hours per tech, 20 hours total. Owner time comes to 26 hours. Tech time is roughly 16 hours each, 32 hours total, and much of that rides on calls that were happening anyway.

Path B, outside. Both techs attend a three-day session. Four days off the board each including travel, at 8 hours a day, so 32 hours per tech and 64 hours total. Owner time is zero, except a two-hour teach-back review on return.

Compare them honestly. Path A consumes about 58 hours of shop time in total, Path B about 66. Path B is roughly 14 percent more total hours, which on its own would favor Path A.

But the hours are not interchangeable. Path A spends 26 hours of the single most constrained person in the building, and those 26 hours come directly out of estimating and permit filing, which is a second gap the same audit already flagged. Path B spends zero of them.

So the decision turns on one question: is the owner's time the binding constraint right now? In the busy season it plainly is, and Path B is correct despite costing more total hours. In the slow season it plainly is not, the owner has the hours, and Path A is better because the teaching happens on the shop's own equipment and customers, and the sign-off is real field work rather than a classroom certificate.

Which means the actual answer is a scheduling answer, not a training answer: run Path A in the off season, and if the gap will not wait for the off season, run Path B. A shop that never asks the question defaults to whichever path the owner is more comfortable with, and pays for that comfort in the constraint they cannot see.

What changes the answer

A jurisdiction with mandatory continuing education. The hours are compulsory, so the question is not whether to send but how to make required hours useful. Pick the required course that most closely overlaps a real gap.

A shop under about four people. The crew hole from a multi-day absence is proportionally much larger, and losing one of three field techs for four days is a genuine revenue event. Weight online and evening formats much more heavily, and prefer sessions that run one day at a time.

A learner with a poor track record of applying training. Do not send them. That is not a punishment, it is an accurate reading: the return depends entirely on application, and this is a person for whom in-house supervised instances work better.

Training you can get delivered on site. Some distributors and associations will run a session at your shop for a group. This changes the math completely by removing travel and letting you train several people in one hole rather than several.

How to verify you got this right

Sixty days after the training, ask one question: name a job we did differently because of it. A specific job. If nobody can name one, the training did not land, and the fault is usually not the class - it is that no work of that type was dispatched to that person in the first month.

Second check: did the teach-back happen, on the date set, with a written page produced. A skipped teach-back is the clearest possible signal that the shop treated the trip as an individual benefit.

Third check: run the same task family through your next gap audit. If the qualified-and-available count did not go up, you bought attendance, not capability.

References

• EPA Section 608 and equivalent credentialing requirements, as examples of capability that can only be certified externally
• OSHA training requirements applicable to your trade, for the mandatory categories where no in-house option exists
• See related: The Skills Gap Audit SOP, for identifying which gap justifies the seat
• See related: The Training Record SOP, for recording an external credential and its expiry
• Consult counsel in your state before using any training repayment or wage-deduction provision