The Apprentice-to-Journeyman Path as a System
Why this matters
Most shops have a ladder. Very few have a system. The difference shows up the day a senior tech gives notice and you go looking for the person who was supposed to be ready by now, and find three people who are all about two thirds of the way there.
A ladder is a set of rungs and criteria. A system is a ladder plus a flow rate: how many people can be moving up it at once, how fast, and what limits that. Almost every shop that stalls on developing people has a perfectly reasonable ladder and no idea what its flow rate is, so it takes on more learners than it can carry and quietly stretches everyone's timeline instead of admitting the constraint. This article is about running the pipeline. For how to define the rungs themselves, see the sibling article in References.
The constraint is teaching capacity, not learner capacity
Owners size the program by asking how many people want to advance. That is the wrong end. The pipeline is limited by how many learners your teaching-capable techs can actually carry, and that number is small and stubborn.
A tech who is genuinely carrying a learner is running roughly two to three calls a week at teaching depth: narrating the reasoning, handing over the tool, letting the learner be slow, then debriefing. On the other calls the learner is present but the tech is producing, which is exposure, not instruction. So one teaching-capable tech supports approximately one learner. Two, and both learners get half the depth and twice the elapsed time.
That gives you the only capacity number that matters: seats equals teaching-capable techs, not headcount, not trucks, not enthusiasm. A shop with six field people and two techs who can genuinely teach has two seats, whatever its ladder says.
Note the word "teaching-capable" rather than "senior." They are not the same person often enough to assume it. Being excellent at the work and being able to make someone else excellent are separate skills, and treating them as one is how a shop ends up with four seniors and zero usable seats.
The four things a system needs that a ladder does not
An entry gate. Who is allowed to start, and what is true on day one. Without it, everyone is nominally on the path and nobody is actually on it.
Advancement criteria that are decided by evidence, not by tenure. A rung reached by calendar is a rung that means nothing to the dispatcher, who has to know what the person can actually do.
A declared flow rate. How many seats, how long a seat is occupied, and therefore how many people finish per year. This is the number that tells you whether your pipeline can cover your attrition.
An exit. What happens when someone completes, and what happens when someone stalls. A path with no exit for a stalled apprentice turns into a permanent middle tier of people who have stopped advancing and are still being described as "coming along."
Throughput arithmetic, and why it is worse than it looks
Two numbers produce your throughput: seats, and time in seat.
Time in seat is partly outside your control. Registered apprenticeship programs generally require at least 2,000 hours of on-the-job learning per year, and licensed construction trades commonly run about four years, though whether an hour requirement applies to you at all, and how many hours, is set by your state licensing board and the trade in question. In an unlicensed trade you set the duration yourself, and the honest floor is however long it takes to accumulate the supervised instances your sign-off criteria demand.
The arithmetic is unforgiving. Two seats on a four-year path produce 0.5 completions per year at steady state. If your shop loses roughly one experienced tech a year to retirement, a move, or a competitor, internal supply covers half your loss. You are still a net hiring shop, and you should plan hiring on that basis rather than telling yourself the program will catch up.
To produce one completion per year on a four-year path you need four seats, which means four teaching-capable techs. That is the actual cost of "growing our own," and it is worth knowing before you build the org chart around a pipeline that cannot deliver it.
The overload trap
Here is the failure that looks like ambition. A shop with two seats takes on three apprentices, because a good candidate walked in and turning them away felt wasteful.
Three learners sharing two seats each receive about two thirds of the teaching depth. A four-year path stretched by that factor runs about six years. All three finish later than the two would have, none of them has a sense of progress, and the shop concludes that apprentices are unreliable and stop developing anyone.
This is a work-in-progress limit, exactly like holding six jobs open on one truck. The seat count is a real constraint whether or not you acknowledge it, and the only thing you control is whether you overload it. If the third candidate is genuinely worth taking, the answer is to add a seat by developing a teacher, or to stagger the start by six months, not to squeeze.
Where pipelines stall, and the signature of each stall
| Stall | What it looks like | What it actually is |
|---|---|---|
| The permanent helper | Two years in, still fetching and holding, competent at nothing signed | No advancement criteria, or criteria nobody reads |
| The invisible middle | Can do the work, nobody will say they are ready | No sign-off authority named, so nobody signs |
| The bottleneck teacher | One senior carries every learner and is always behind | Seats miscounted as headcount |
| The rung with no reward | Advances a level, nothing changes in pay, scope, or autonomy | The ladder is descriptive, not operative |
The last one is the one shops underestimate. If reaching a rung does not change what the person is dispatched to, what they are paid, or what they are allowed to decide, then the rung is a label and the apprentice will read it correctly as a label within about a month. Every rung needs at least one of those three to change, and the change should be visible to the rest of the crew, because the crew is watching to see whether the path is real before they invest in it.
Hours and verified skill are both required, and they check each other
Where an hour requirement applies, treat it as a floor for eligibility, never as evidence of competence. Hours are easy to accumulate and prove nothing about what happened during them. An apprentice can log 2,000 hours holding a light.
Verified skill without hours has the opposite problem. Someone can demonstrate a procedure correctly three times and still not have hit the range of conditions that turns a procedure into judgment: the equipment that was installed wrong, the customer who is hostile, the failure that presents as something else.
Use both as gates, in this order: verified skill decides whether the person is competent, hours decide whether they have seen enough variety for that competence to hold. Where the two disagree, the interesting case is the fast learner who is signed off on everything at half the hours. That person is usually genuinely good and usually has a hole, and the hole is almost always in the rare-condition cases the calendar has not handed them yet. Send them at the rare conditions deliberately rather than blocking on the clock.
A worked example: six field people over 24 months
A shop runs six field people: three at journeyman level, one mid-level tech, two apprentices. Of the three journeymen, one is genuinely teaching-capable, one is capable but is the shop's only person on two specialty rows and has no free calls, and one is technically strong and a poor teacher. Effective seats: one.
Two apprentices on one seat. Each is getting about half the teaching depth, so a path the shop believes is 36 months is running at about 72 months for both. Nobody has noticed because the calendar has not caught up yet, and both apprentices are described as "doing fine."
The read: this shop's pipeline produces roughly 0.33 completions per year on paper and effectively zero on the current split. It cannot cover a single retirement.
The fix has three moves, in this order. First, develop the mid-level tech into a teacher, which is a smaller project than it sounds: that person is closer to the apprentice's confusion than any journeyman is, and can carry the routine objectives. That takes seats from one to two, and immediately un-stretches both apprentices back toward 36 months. Second, protect two calls a week on the booked specialty journeyman's board so they can carry only the specialty objectives, which are the ones nobody else can teach. That is not a full seat, call it a partial seat aimed at exactly the rows at risk. Third, stop trying to make the poor teacher teach, and instead have them write down the two procedures only they run, so the knowledge is captured even though the teaching is not.
Twenty-four months later, the arithmetic that matters: two seats plus a partial, both apprentices advancing on schedule, one new teaching-capable tech developed, and the specialty rows two-deep instead of one. The shop did not hire, did not buy a program, and did not add a training budget line. It counted seats honestly and stopped overloading them.
What changes the answer
- A licensed trade with a legal supervision ratio. The state or local ratio of apprentices to licensed personnel caps your seats regardless of your teaching capacity, and it is a legal cap, not a guideline. Check it before planning any expansion of the program.
- A shop under three field people. The pipeline model does not apply at that size. You are not running a program, you are training one person directly, and the right instrument is a written objective list and a completion date, not a ladder.
- High turnover. If the average tenure is under the length of your path, the pipeline mathematically cannot complete anyone. Fix retention first, or shorten the path by narrowing the scope of what "journeyman" means at your shop.
- A seasonal shop. Time in seat should be measured in supervised instances, not calendar months, or your apprentices lose an entire off-season of nominal progress and arrive at their review date with nothing signed since spring.
How to verify the system is running
Ask four questions once a quarter and write the answers down:
- How many seats do we actually have? Count teaching-capable techs with room on their board, not seniors.
- What rung is each learner on and what closed it? If the answer is a date rather than a signed objective, the criteria are not being used.
- What is our completion rate per year, and does it cover our attrition? If it does not, say so out loud and plan to hire, rather than assuming the pipeline will surprise you.
- Who has not advanced in six months, and why? Every stalled learner has one of the four signatures in the table above. Name which one before deciding what to do.
References
- See related: The Apprentice-to-Journeyman Path: Design, Apprentice Pay Progression Tied to Real Skill Milestones
- See related: How to Pick Which Tech Should Be Teaching, The Competency Sign-Off SOP
- U.S. Department of Labor, registered apprenticeship program standards and on-the-job learning requirements
- State licensing board rules on apprentice supervision ratios and required hours, which vary by state and trade