Train Apprentices or Hire Journeymen: How an Electrical Contractor Should Run the Numbers

The problem: The choice between training apprentices and hiring experienced journeymen gets made on the hourly wage, because nobody tracks how much work each level actually produces or how much of it comes back.

The solution: Cost labor by finished work rather than by hours purchased, so the comparison includes productivity and rework instead of only the rate on the pay stub.

Every electrical contractor at this size faces the same fork, usually in the middle of a good year. The backlog is strong, the crews are stretched, and you need more hands. You can bring in journeymen at market rate, which is fast and expensive, or you can build apprentices, which is cheap now and expensive in a different way, in supervision and in time.

The debate in the truck is almost always about the wage. If a journeyman runs $40 to $45 an hour loaded in your market and an apprentice about half that, hiring experienced looks like paying double. That comparison is the wrong one, and almost every contractor knows it is the wrong one, and makes it anyway, because the numbers that would make the right comparison possible are not written down anywhere.

What the wage rate leaves out

An hour of labor is not the thing you sell. Finished, correct, inspected work is. The wage rate tells you what an hour costs to buy and nothing about what an hour produces.

Three things sit between the two, and none of them appear on a pay stub.

Productivity. How much of the estimated work actually gets done per hour, by experience level. A first-year apprentice on rough-in might run at a fraction of plan. A strong journeyman might beat it. Nobody in the business knows the ratio, though the foremen could all guess.

Rework. How much of what gets installed has to be corrected, redone, or fails inspection. Rework costs twice: the labor to do it again, and often the material. It also tends to be silent, because a crew fixing its own work on Thursday just looks like a crew working on Thursday.

Supervision drag. Apprentices do not work unattended. Every hour of apprentice time consumes some fraction of a journeyman's hour in direction, checking, and correcting. That fraction is real cost, and it is charged to nobody.

Put those three together and the honest unit is cost per unit of finished work, not cost per hour. Contractors who reach that number sometimes find the apprentice is still clearly cheaper, particularly on repetitive scope where supervision is light. Sometimes they find the gap is much narrower than the wage suggested, and on complex or inspection-heavy work it can invert. The point is not that one answer is right. It is that the answer varies by scope, and you cannot know yours without the data.

Why nobody has these numbers

Not because contractors are careless. Because of how the information gets captured.

Time is recorded for payroll. A hand or a phone logs eight hours, and those hours flow into the payroll run and then into a weekly labor total. Sometimes they carry a job number. They very rarely carry the phase, and almost never carry the person's experience level in a form anything can sort by.

Rework is worse. There is generally no code for it. A redo gets logged against the same job and phase as the original work, because from the field's point of view it is the same work. So the labor hours for doing it right and doing it twice are indistinguishable in the record, which means the business has structurally no way to see its own rework rate.

The estimate, meanwhile, sits in the estimating system with the hours by phase that the job was priced on. That is the benchmark that would make productivity measurable, and it is on the other side of a wall from the actual hours.

Nothing here is missing. It is just never brought together, so the one comparison that would settle the training question, estimated hours by phase against actual hours by phase by experience level, cannot be run.

How the answer becomes available

Making it available takes two changes to what gets captured, both small enough for a field crew to live with.

First, hours land against a job and a phase, not just a job, and they carry who worked them, which the system already knows from payroll. Second, rework gets its own code, so a redo is distinguishable from original work at the moment it is logged. Foremen resist this at first, reasonably, because it sounds like blame. It goes down easier when it is presented as what it is, a way to find out which scopes need more supervision rather than which hands to punish.

Then those actuals sit next to the estimated hours the job was priced on, in one connected picture, so productivity against plan by phase and by experience level is something the business simply knows.

The upkeep should not land on the office. Instead of someone building a labor analysis each quarter, automation watches the hours as they come in and flags the exceptions: the phase running consistently over plan, the crew mix where rework clusters, the apprentice whose productivity curve has flattened and who probably needs different work rather than more of the same.

A look at an electrical contractor

Consider an electrical contractor doing about $8 million a year, roughly 45 in the field, a mix of commercial tenant improvement and light industrial. Two journeymen have given notice within a month of each other. The owner can replace them at about $42 an hour loaded, or take on three apprentices at about $24 and lean on the foremen.

Suppose the company codes hours by phase and by person and separates rework for two quarters before deciding. What you would expect to surface is a productivity curve rather than a single number: apprentices closing on plan quickly on repetitive scope like device trim and branch runs, and staying well off plan on panel work, terminations, and anything an inspector looks at closely.

Run the math on one phase. Say a journeyman completes rough-in at about plan with a rework rate near 1 percent, at $42 an hour. Say a second-year apprentice on the same phase runs about 40 percent slower with rework near 6 percent, at $24. The apprentice needs roughly 1.4 hours for the journeyman's one, plus about 0.08 of an hour of redo, which is roughly $35.50 of direct labor against the journeyman's $42.40. Then add supervision: if each apprentice hour consumes even 15 percent of a journeyman's hour, that is another $8.80, taking the apprentice to about $44 for the same finished work.

On that phase, at that experience level, the cheap labor is not cheaper. On device trim, where the productivity gap is smaller and supervision is light, the same math would likely favor the apprentice decisively. So the answer the contractor would arrive at is not "hire" or "train". It is a scope-by-scope allocation: apprentices where the curve closes fast, journeymen where it does not, and a training plan that moves people between the two deliberately instead of by whoever is available on Monday.

That is a better decision than either of the two the owner started with, and it is unavailable to a contractor who only knows the wage rates.

How to start

You can have a usable version of this within two quarters, without new field software.

1. Code hours to a phase, not just a job. The phase is where the productivity comparison lives, because it is what the estimate was built from.
2. Give rework its own code. Until a redo is distinguishable from original work in the record, your rework rate is unknowable and it is almost certainly not zero.
3. Set actuals against estimated hours by experience level. That single comparison, run by phase, is most of the answer.
4. Let the flagging run itself. Set automation to surface phases drifting off plan and rework clustering by crew mix, so the training decision is a standing read rather than an annual argument.

The takeaway

Deciding between apprentices and journeymen on the hourly wage is deciding on the one number that does not measure what you are buying. What you are buying is finished work, and the cost of that depends on productivity against plan, on how much comes back, and on the supervision it consumes. Those are all measurable from records you already keep, once hours carry a phase and rework carries a code. Get two quarters of that and the question stops being train or hire. It becomes which scopes each is right for, which is the answer that actually protects your margin during a growth year.