Who Owns Training When Nobody Owns Training
Why this matters
Ask a nine-person shop who owns training and you get a pause, then "all of us, really." That answer is the diagnosis. Training in most small shops does not fail because nobody cares or because the plan was bad. It fails because the work is real, recurring, and unassigned, so it gets done by whoever happens to feel worst about it that week - and that person changes, so nothing accumulates.
Every other recurring obligation in your shop has a name attached. Somebody owns the truck inventory. Somebody owns invoicing. Nobody would say "we all own payroll." Training is the one function shops leave collectively owned, and collectively owned means owned at the priority of whatever else that person is holding on a Tuesday in July.
The default owner is the person who feels worst about it
When a role is unassigned, it does not go undone. It gets absorbed by the most conscientious available person, silently, without authority or time. In a trades shop that is almost always the best senior tech, because they are the one who sees the apprentice struggling and cannot watch it.
That arrangement has three specific defects, and they are structural rather than personal:
- It is unfunded. The hours come out of that tech's own billable time and nobody accounted for them, so the training work shows up on their numbers as a shortfall.
- It is unauthorized. They can teach, but they cannot change the dispatch pairing, cannot block a job, and cannot say no to a schedule that makes teaching impossible. So they teach when the schedule permits, which is exactly when teaching matters least.
- It is unrecorded. Nothing they do lands in a system, so when they take a week off or move on, the shop discovers it had a training program only in the sense that one person was carrying it.
Split the job in two, because it is two jobs
The single most useful move here is recognizing that "training" bundles two roles that need different people and different amounts of authority.
Owner of the system. One person, shop-wide. Keeps the skills matrix current, keeps the training record honest, runs the credential expiry report, decides what the shop is trying to build this quarter, and reconciles what got dispatched against who was signed off for it. Administrative, low-volume, high-consequence. This role must be singular.
Mentor of the person. Many people, one per learner per task. Teaches, observes, corrects, signs off on instances. Field work, high-volume, distributed. This role must be plural, because a shop whose teaching all runs through one senior has replaced a training gap with a bottleneck.
Shops conflate these and then argue about the wrong thing. "We do not have anyone who can run training" is usually true of the mentor role, meaning nobody with time to teach every skill, and false of the system role, which is a handful of hours a month of bookkeeping and decisions.
What the system owner actually does
The duty list is short and it is finite. Write it down as a real role description, because a role nobody can enumerate is a role nobody can hold accountable.
| Duty | Cadence | Rough time |
|---|---|---|
| Reconcile dispatch against sign-offs | Monthly | 1 hour |
| Run the credential expiry report and start renewals | Monthly | 0.5 hour |
| Set each learner's named target for the month | Monthly | 1 hour |
| Facilitate one shop or bench session | Monthly | 1.5 hours |
| Enter and audit records from field sign-offs | Weekly, in small pieces | 1 hour per month |
| Observe or arrange competency assessments | As they come up | 1 hour per month |
That totals roughly 6 hours a month for a shop with two or three active learners. It is not a job. It is a named responsibility with a real but small footprint, which is precisely why it survives being unassigned for years - it never gets big enough to force the issue, and it never gets done.
What the shop owner cannot hand off
Three things stay with whoever runs the business, no matter who holds the system role. Delegating them is the most common way a training owner is set up to fail:
- The hours budget. Somebody has to decide, out loud, what fraction of paid hours goes to training and what that means for the schedule. A training owner without a budget is negotiating for every hour, every week, against dispatch. They will lose, because dispatch has revenue on its side today and training has revenue on its side next year.
- Sign-off authority scope. Who is permitted to assess and sign off on what. If a lead can be talked into a sign-off by a busy dispatcher, the record stops meaning anything.
- Consequence. If a person misses their target three months running, something happens. If nothing happens, the target was a suggestion and the crew works that out fast.
Picking the owner, and how each candidate fails
The best senior tech. Highest credibility, worst availability. They are your most billable person, so every hour of the role costs the shop its most valuable hour. Works if you actually reduce their dispatch load and protect the hours. Fails, reliably, if you do not, and you will have burned your best tech's goodwill in the process.
The office manager or dispatcher. Best availability, no technical credibility to assess anything. Excellent fit for the record, credential, and scheduling half of the role, useless for the assessment half. This is often the right answer with a modification: they own the system, and assessment authority is explicitly delegated to named seniors per task.
The owner. Highest authority, worst consistency. It survives the slow months and vanishes in the busy ones, which teaches the crew that training is what you do when there is nothing else to do. Workable in a shop under about five people where the owner is still turning wrenches. Above that it is the least stable option.
An outside consultant or a part-time trainer. Real capability, no context. They can build a matrix and a curriculum, but they cannot see that your board runs 40 percent on one job type, and they will not be there when the sign-off decision gets contested. Use for building the initial structure, never for holding the role.
Authority, or it is only a title
A named owner with no levers is worse than no owner, because now the shop believes the problem is handled. Give the role exactly three powers and write them down:
- The pairing call. The training owner can specify which learner rides with which senior for a given job, and dispatch honors it unless the owner overrides. Without this, learners get paired by who is free, and who is free is systematically the person with the fewest calls, who is often not your best teacher.
- The block. The training owner can say a person is not sent unsupervised on a task they are not signed off on. This is a safety and liability control, and it needs teeth.
- The hours draw. The training owner can spend the budgeted training hours without asking each time.
A worked example: assigning the role in a nine-person shop
Six field techs, two office, one owner. Two apprentices in their first year, one mid-level tech cross-training onto a second job type.
The owner assigns the system role to the office manager, with assessment authority delegated to two named seniors, one per job-type family. The office manager already holds the credential file and the scheduling board, which is over half the duty list, so the incremental load is the reconciliation, the monthly target-setting, and the record auditing.
Measure the cost honestly. The duty list runs about 6 hours a month. For the office manager, roughly 5 of those 6 hours displace other office work, not billable work, and the office month absorbs it by dropping a low-value reporting task. The two seniors carry the assessment half: about 1 hour a month each, and because assessments happen on live jobs the incremental billable displacement is closer to half of that.
So the shop-wide bill is roughly 6 office hours plus about 1 hour of displaced billable field time per month. Against a shop with six field techs, a rough 160-hour month and a 70 percent utilization target gives about 112 billable hours per tech, so 1 displaced hour is under 1 percent of a single tech's billable month and well under a quarter of a percent of the field crew's total.
That is the number to put in front of an owner who says the shop is too busy to run training. The system half of the job costs less than one percent of one tech's month. What is expensive is the mentoring, and mentoring was already happening, unfunded and unrecorded, which is why nobody could see what it cost.
What changes the answer
Under five people. Do not create the role. The owner holds it, the duty list shrinks to about 2 hours a month because there are one or two learners at most, and the record can be a single spreadsheet. Formalizing at this size adds process without adding capability.
Ten to twenty-five people. The split described above is the right shape and the pairing authority becomes load-bearing, because at this size dispatch stops knowing every learner's status personally.
Above about twenty-five, or multiple locations. The system role starts to justify dedicated time rather than being an add-on, and the mentor role needs its own qualification: not every senior should be signing off, and you now need a documented standard for who can assess.
A shop running a formal registered apprenticeship. The program sponsor requirements will specify supervision ratios and record obligations, and those override your internal design. Build around the program's requirements rather than layering your own on top.
The failure modes
The role is announced and never resourced. Person is named in a meeting, nothing is taken off their plate, no budget of hours is set. Dies within two months and the next attempt is harder because the crew has now seen it fail once.
Assessment authority creeps. It starts with two named seniors and within a year anybody senior signs anything, usually because a sign-off was needed on a Friday afternoon. Catch it by auditing who assessed the last twenty competency entries. If the assessor list is longer than your named list, the control is gone.
The owner takes it back during a busy stretch. Usually framed as helping. What it actually does is show the crew that the assignment was provisional, and the named owner stops making decisions because they might be overridden.
Nobody notices it stopped. The clearest tell is the record: if no competency entry has been made in six weeks, the role is vacant regardless of whose name is on it.
How to verify you got this right
Ask three people in the shop, separately, who owns training. If you get one name from all three, the assignment is real. If you get three answers or a pause, you have a title, not an owner.
Then ask that named person two questions: how many hours a month is this, and what are you allowed to decide without asking. If they cannot answer either one, the role has no budget and no authority, and it will be vacant again by the end of the quarter.
References
- U.S. Department of Labor registered apprenticeship program standards, for sponsor and supervision requirements where a formal program applies
- See related: The Annual Skills Review SOP, for the reconciliation this role runs
- See related: Why Most Small-Shop Training Fails in the Second Month, for what happens when the role goes vacant mid-program
- See related: The Mentor Pairing SOP, for the plural half of the split described here
- Trade-standard practice for role definition and delegation in small contracting businesses