The Tech Who Has Plateaued and What to Do About It
Why this matters
The plateaued tech is the hardest development case in a small shop because nothing is wrong. They show up, the work is clean, callbacks are low, customers like them. There is no incident to react to. So year three looks like year two, year four looks like year three, and one day you notice that a tech with six years in still hands off anything unfamiliar and still cannot be dispatched to half your job types.
That costs you twice. You are paying a rate that assumes growth you are not getting, and you still cannot take a week off because the coverage map never widened. It also costs them, because a tech who stops climbing eventually gets bored, and bored competent techs leave for a shop that looks like a bigger ladder.
Plateau is not the same as underperformance
Keep these separate, because the fixes point in opposite directions.
| Underperformer | Plateaued tech | |
|---|---|---|
| Current output | Below standard | At or above standard |
| Trend | Falling or never arrived | Flat |
| What it costs you | Rework, callbacks, customer risk | Coverage gaps, key-person risk, their eventual exit |
| The conversation | Corrective, with a standard to return to | Developmental, with a destination to move toward |
| Getting it wrong | Coaching a will problem as a skill problem | Treating a flat line as a performance problem and insulting a good tech |
The most common mistake is running a plateau conversation in the shape of a correction. A tech who has done nothing wrong, called in and told they are "not where we need them to be," hears that their good work has been reclassified as bad. That conversation costs you the tech more often than it fixes the plateau.
The four causes, and the different move each one takes
Plateaus have a small number of causes and they are distinguishable if you look at the right evidence.
Cause 1: exposure ceiling. They have stopped growing because they have stopped seeing anything new. Dispatch keeps sending them the same job types because they are reliable at those job types, which is rational for this week and corrosive over a year. Evidence: pull their last 60 to 90 tickets and count distinct job types. If a large majority of tickets fall into a handful of types while the shop runs many more, this is your cause. Move: change the dispatch mix deliberately, not the person.
Cause 2: no reason to climb. Nothing at the top of the ladder is visible or worth reaching. If pay, dispatch priority, and standing are identical for someone with 6 competencies and someone with 14, the shop has priced growth at zero and gotten what it paid for. Evidence: ask what the next step up is. If they cannot name it, or they name it and it does not exist, this is your cause. Move: build and publish the rungs before you push anyone up them.
Cause 3: the competence trap. They are genuinely excellent at a narrow band and have organised their working life around never leaving it. Being the best in the shop at three things is more comfortable than being a beginner at a fourth, especially in front of people who look up to them. Evidence: they decline or deflect unfamiliar work with plausible reasons, and the reasons vary. Move: make the new skill low-exposure before you make it public.
Cause 4: fear of the next tier. The next step is not more field skill, it is something else: reading a customer, quoting, running a two-person crew, signing off someone else's work. That is a category change, not an increment, and plenty of strong techs stall there quietly rather than admit they do not want it or do not know how. Evidence: they are fine with any new technical skill and consistently avoid anything with people or money attached. Move: name the actual skill, teach it as a skill, and separately let them opt out of the management track without penalty. Not every good tech should become a lead, and pretending otherwise creates a bad lead and a lost tech.
Read the evidence before you read the person
Do not diagnose a plateau from a feeling. Three pulls, all of which you already have data for:
- Job-type distribution over the last quarter. Distinct types run, and the share taken by the top three. This separates cause 1 from everything else in about ten minutes.
- The skills matrix column. How many competencies have moved in twelve months. A column that has not moved at all points at cause 2 or 3; a column that moved on technical rows and not on the customer-facing or leadership rows points at cause 4.
- Who they get asked for. If other techs call them for help on the same three subjects and never on anything else, you have located the exact edge of their band, and that edge is where the plateau lives.
The conversation that works
Open with the record, not with the concern. "You have run 61 tickets since April. Fifty-one of them were three job types. That is not a complaint, those are the three I trust you with most. I want to talk about the other ten." That framing is accurate, it is not a criticism, and it puts the dispatch pattern on the table as a shop decision rather than a personal shortfall.
Then ask, and stop talking. "Where do you want to be in two years" is a fine question if you have actually built somewhere for them to be. If you have not, it is a trap, because they will answer honestly and you will have nothing to offer.
Close with one commitment on each side. Yours is usually a change in what work they get sent. Theirs is usually a named competency with a date. Two commitments, written down, reviewed in ninety days.
A worked example: the six-year tech who looks fine
A tech with six years in, three of them at your shop. Callback rate is the lowest on the crew. Customers request them by name. The skills matrix shows 9 of 22 competencies at the sign-off level, and the same 9 it showed at last year's review.
The pull. Last quarter: 58 tickets. Job-type breakdown comes back 44 tickets across 3 types, 14 across 5 more, 0 across the remaining 14 types the shop runs. So 76% of their work is in 3 of 22 job types.
What that rules in and out. Exposure ceiling is clearly present. But it is not the whole story, because 5 other types did get worked, and none of those 5 turned into a matrix move over twelve months. Exposure alone would have produced at least some drift.
The second pull. The 9 signed-off competencies are all technical. Every row on the matrix that involves quoting, explaining a recommendation to a hesitant customer, or signing off a helper's work is blank, and has always been blank. Cause 4 is in play alongside cause 1.
The conversation. They confirm both, which is common: "I like the ones I am good at, and I do not want to be the guy who tells someone their unit is done." That is a specific, teachable fear, not a character trait.
The plan. Two commitments. Dispatch commits to routing them 4 tickets a month from outside their top three types, chosen from the teaching-job bucket. Against 58 tickets a quarter, about 19 a month, that is roughly 21% of their volume - a real change to their week, not a rounding error, and worth saying out loud to dispatch rather than slipping it in. They commit to one competency: run the customer conversation on their own diagnoses, with the lead listening in on the first three.
Ninety days later. New job types worked: 3. Matrix rows moved: 2, one technical, one customer-facing. Callback rate up slightly, from the lowest on the crew to second-lowest, which is the expected cost of working unfamiliar job types and is worth it. The tech's own read: the customer conversation was harder than the new job types, which is exactly the information the blank matrix rows were already telling you.
What the shop got. Coverage on 3 more job types, one fewer single point of failure, and a tech who has a reason to still be there in year eight. Note what it did not require: no pay change, no title, no training budget. It required a dispatch decision repeated 12 times and one honest conversation.
What changes the answer
A tech in their final working years. Someone five years from stepping back is a different case entirely. Pushing new competencies at them has a short payback window and is often unwelcome. The high-value move there is conversion: their remaining years are worth more spent transferring what they already know than acquiring what they do not.
A plateau that started recently in a tech who was climbing. A flat line that appeared suddenly in someone with a rising trend is usually not a training problem at all. Something changed: at home, in their body, in how they are being treated. Ask before you plan.
A shop with genuinely narrow work. If you run three job types and that is the whole business, a tech mastering all three is not plateaued, they are finished with the technical ladder. Their growth has to come from a different axis: teaching, quoting, running a crew, or owning a part of the operation. Manufacturing new technical competencies you do not need is busywork and they will see through it.
Pay compression. If a plateaued tech is already at the top of your band, no development plan survives contact with the fact that the climb pays nothing. Fix the ladder first or expect the plan to stall in month two.
How to verify you got this right
Ninety days is the right review window: long enough for exposure to accumulate, short enough that a stalled plan is caught before it becomes a shrug.
- Did the mix actually change? Count new-to-them job types worked. If the answer is zero, dispatch quietly reverted to the comfortable assignment and the plan never started. This is the single most common failure and it is a shop failure, not a tech failure.
- Did a matrix row move? One row in ninety days is a healthy pace for a working tech with no dedicated training time. Zero rows with plenty of new exposure means they are being sent to unfamiliar work with nobody teaching, which produces frustration and callbacks rather than growth.
- Is the callback rate acceptably worse? A small rise on unfamiliar job types is the correct signal that real learning is happening. No change at all usually means they are quietly handing the unfamiliar work to someone else.
- Would they describe the plan the same way you would? Ask them to say it back. If your version is "we are broadening their coverage" and theirs is "the boss thinks I have gotten lazy," the conversation landed wrong and you have started a retention problem while trying to solve a development one.
References
- See related: Separating a Skills Gap From a Will Gap, Coaching a Struggling Employee Back to Standard
- See related: Build a Skills Matrix: Who Can Do What, Apprentice Pay Progression Tied to Real Skill Milestones
- U.S. Department of Labor, registered apprenticeship competency-based progression models
- Trade-standard practice for technician development reviews