Why This Matters
Turnover almost always gets diagnosed as a pay problem, a market problem, or a culture problem. Sometimes it’s one of those. Often it’s none of them, and the real cause sits somewhere nobody thought to look: not in what people are paid, but in how they’re taught.
Owen, his shop, and the specific details below are a composite, not a real company. The pattern is real.
I. The Symptom
Owen got the text on a Sunday night, which was how these things usually happened. A machinist he’d hired eight months earlier, someone he’d genuinely liked, was giving two weeks’ notice. Family reasons, the message said, though Owen had heard some version of that phrase enough times to wonder.
This made four in eighteen months. Four hires, trained, ramped up, gone before their first anniversary. Meanwhile the core crew, the six machinists who’d been there five years or more, barely turned over at all.
Owen ran a precision machine shop, eighteen employees, just under $2.8M a year. Pay was competitive for the region, benefits were solid, and he made a point of knowing everyone’s name and asking about their kids. He didn’t think of himself as the kind of owner people fled.
He’d started assuming it was the labor market. Good machinists were hard to find and easy to poach, and maybe the newer hires just had more options, or less patience, than the veterans who’d built their whole careers at the shop.
He posted the opening again that Monday, same as the last three times. Same job board, same slightly higher offer than the last one, hoping a better number would finally stick.
It hadn’t worked yet. He wasn’t sure why he expected this time to be different.
II. The Outside Observer
Marcus arrived on a Thursday, a mutual contact having mentioned Owen’s shop the way most of these conversations started.
He asked if he could just walk the floor for an hour, watch how work actually happened. Owen agreed, mostly because a friend had vouched for Marcus and said the visits were harmless, if occasionally uncomfortable.
Marcus watched a newer hire, maybe four months in, hover near an older machinist named Frank, waiting for a break in Frank’s focus to ask a question. He watched that happen twice more over the hour, each time the new hire waiting, then asking, then getting a quick answer before Frank returned to his own work.
“Is that how new people usually learn the specifics here?” Marcus asked. “Watching, then asking when someone’s free?”
“Frank’s been here eighteen years,” Owen said. “He knows this shop better than I do at this point. New guys shadow him for the first few weeks, then they’re mostly on their own.”
“How long does ‘mostly on their own’ usually take to feel comfortable?”
Owen thought about it. “Honestly, it varies a lot. Some people pick it up in a month. Others take three or four and never quite seem settled.”
“Do you know which ones tend to leave? The ones who pick it up fast, or the ones who take longer?”
Owen didn’t have an immediate answer to that.
Marcus asked if he could come back the following week and specifically watch a new hire’s first few weeks.
III. The Wrong Diagnosis
Marcus returned the next week, timed around a new hire’s second week on the floor.
Owen had a theory ready. “I think we need to raise the starting pay again, or maybe bring in a recruiting agency that can find people who are more likely to stick around long term.”
“Maybe,” Marcus said. “What would higher pay actually change here?”
“It’d make the offer more competitive. Give people less reason to leave for a few dollars more somewhere else.”
“Has anyone who’s left told you pay was the reason?”
Owen paused. “Not directly, no. It’s mostly been ‘not the right fit’ or vague stuff like that.”
“That’s what I want to map today,” Marcus said. “Not whether the pay is competitive. Whether we know what actually happens between someone getting hired and someone deciding to stay or go.”
He asked Owen to walk him through the full arc, hiring to fully ramped, for a typical new machinist.
Owen talked, and Marcus wrote it as a chain. Job posted, interviews, hire made, first two weeks shadowing Frank, weeks three through six handling simple jobs with Frank nearby for questions, week seven onward mostly independent, occasional check-ins with Owen about how things were going.
Six stages. All of them, one way or another, ran through Frank’s availability and patience on any given day.
“You don’t have a pay problem,” Marcus said, looking at the chain. “You have a Frank problem. Not because Frank’s doing anything wrong. Because he’s the only place this knowledge lives, and he’s already busy doing his own job.”
Owen looked at the chain, all six stages quietly depending on one person’s bandwidth.
IV. The Real Constraint
“Tell me more about weeks three through six,” Marcus said. “What actually happens when a new hire has a question and Frank’s mid-task?”
“They wait, usually. Or they guess and hope it’s close enough, then Frank corrects it later if it’s wrong.”
“How does that feel, from the new hire’s side, if you had to guess?”
Owen sat with that. “Probably frustrating. Like they’re bothering him, or like they’re supposed to already know things nobody’s actually told them.”
“If Frank were out for two weeks, what happens to a new hire in week four?”
“They’d probably be stuck a lot. Slower. More mistakes, maybe.”
Marcus circled the shadowing and independent-practice stages on the page. “This is the real constraint. Not pay, not the labor market. The entire knowledge transfer for this job lives in one person’s head, and it only comes out in fragments, whenever he happens to be free.”
“So even a recruiting agency would just be bringing me more people to feed into the same bottleneck.”
“Exactly. You’ve been trying to fix who walks in the door. The shop’s been waiting on you to fix what happens to them once they’re inside it.”
V. The Resistance
Owen wasn’t fully convinced, and said so on the next call.
“I hear you, but Frank’s not going to want to sit down and write out eighteen years of knowledge. He’s not a writer, he’s a machinist. He’d probably feel like I’m turning his expertise into a checklist.”
“Fair,” Marcus said.
“And some of what he knows, honestly, might not even be explainable. Feel, instinct, stuff you only get from doing it thousands of times.”
“Also fair.”
“We actually tried a training manual a few years back. Some consultant helped write it. Nobody used it. It sat in a binder in the office.”
Marcus let that sit for a second. “Everything you’re saying is true. It’s also probably why this hasn’t changed.”
Owen didn’t respond right away.
“You’re not wrong that some of it is instinct,” Marcus continued. “You’re not wrong that Frank isn’t a writer. But right now, this entire shop’s ability to onboard anyone runs through one person’s limited hours in a day, and that’s been true through four resignations in a row.”
“So what, I ask Frank to become a full-time trainer?”
“No. Smaller than that. What if we didn’t touch most of what Frank knows yet? What if we just captured how he handles one specific, common job, the kind every new hire runs into in their first month, and tested whether that alone helped?”
Owen exhaled. “One job. Fine. But I’m not putting more on Frank’s plate without asking him first.”
“That’s exactly the right way to start.”
VI. The Small Experiment
They picked one of the most common early jobs new hires ran, a standard part that came through the shop several times a week, simple enough to be a good teaching case, common enough to matter.
Marcus didn’t ask Frank to write a manual. He asked him to talk through it out loud instead, the way he’d walk a new hire through it if he had unlimited patience and time.
Frank spent about ninety minutes doing exactly that, walking through setup, common mistakes, what “off” looked and sounded like before it became a scrapped part. Marcus turned it into a short reference sheet with a few photos and Frank’s own phrasing kept intact.
The next new hire got the sheet during their first week, alongside the usual shadowing.
Two weeks in, that hire needed Frank for questions about that specific job noticeably less than previous hires had at the same point. Not zero. Just less. Frank noticed it too, mentioned that this one “seemed to catch on faster,” without knowing why exactly.
It was one job. One sheet. A few weeks.
But for the first time in longer than Owen could remember, part of Frank’s knowledge existed somewhere other than Frank.
VII. The Shift
A few weeks later, Owen caught himself watching a quality control check on the floor, noticing that it too seemed to live entirely in one senior employee’s judgment, never written down, taught the same fragmented way training always had been.
Nobody had asked him to start looking for this. Once he’d seen it once, in Frank’s corner of the shop, he started noticing the same shape everywhere: knowledge that only existed inside one person, transferred only when that person happened to have a spare moment.
Frank, for his part, hadn’t minded the exercise the way Owen worried he might. If anything, he mentioned it was almost a relief to say some of it out loud instead of re-explaining it from scratch every time someone new started.
When Marcus checked in, Owen got to the point early. “I want to do the same thing with two more common jobs. See if the pattern holds.”
“What do you think you’ll find?”
“Probably that most of what makes someone struggle here isn’t the work itself. It’s that nobody ever wrote down the part that isn’t obvious.”
It wasn’t fixed. Most of Frank’s knowledge still lived only in Frank. But for the first time, Owen wasn’t trying to out-recruit the problem. He was trying to document his way past it, one job at a time.
VIII. The Pattern Named
A year later, the shop looked the same from the parking lot. Same building, same eighteen bays, mostly the same crew, minus the usual small amount of turnover every business has.
But four common jobs now had reference sheets built the same way the first one had, Frank talking, Marcus or Owen writing, Frank’s own phrasing kept intact. New hires still shadowed Frank for their first weeks, but arrived at week six noticeably steadier than hires had a year before.
Of the three people hired since the experiment started, all three had passed their one-year mark, the first time that had happened in years. Owen couldn’t prove the sheets were the whole reason. He was fairly confident they were a real part of it.
Near the end of one of their last conversations, Marcus asked what he’d tell another owner watching good hires walk out the door.
Owen thought about it. “I’d tell them to stop assuming it’s about pay before checking something else first: how someone’s actually supposed to learn the job. If that only lives in one person’s head, on one person’s schedule, you’re not really onboarding anyone. You’re just hoping they get lucky with timing.”
The Pattern
The same shape shows up again, in a fourth system this time: Hiring, tangled up with Knowledge. In each case so far, the constraint hid inside a single person’s undocumented judgment: a quote, a schedule, a forecast, and now, a craft. The industries keep changing. The structure underneath keeps repeating.
Observe. Watch how a new person actually learns the job, not how orientation says they will.
Listen. Ask departing employees what was hard, not just why they’re leaving.
Understand. Map the ramp-up chain: every stage, and who it depends on.
Find the pattern. Look for the one person whose calendar the whole process quietly runs through.
Improve. Capture one small, common piece of their knowledge before trying to capture it all.
Every business has a Frank. Most owners have never asked what happens when Frank is the only door into the job.
Let’s take a look.