Block out one hour before your next hire's first day. Write down his first five days: where he goes, who he rides with, what he touches, and what good looks like by Friday. That hour decides whether he stays, or whether you write the job post again.

AGC and NCCER surveyed 1,830 people across the industry in July and August 2026. Eighty-three percent of firms reported at least some turnover among new field employees in their first 90 days. Not year two. The first 90 days.

The reason is not the one you tell yourself

Ask an owner why the new guy left and you get the labour market. Nobody wants to work. The young ones do not last. We have all said it standing beside a truck one man short.

The 2026 workforce survey asked. Respondents named a mismatch between what workers expected and what the work actually is, more often than any other reason. Physical demands came second. Travel and scheduling came third.

Read that order again. The top answer is not that the work is too hard. It is that the work did not match what the worker expected.

The survey stops there. It does not say who set that expectation, so we will. Somebody described the job out loud, and it was not the labour market.

He never really started

The same survey found 42 percent of firms say new hires fail to show up or quit shortly after starting. That is not turnover. That is a seat you paid to fill twice.

Filling it is the hard part. AGC reported that 87 percent of firms have openings for hourly craft positions. Half say the people applying arrive without the skills, certificates or licences the job needs. You fought for this guy, then lost him over something free.

A fence and deck crew hires differently than a plumbing shop, a roofing outfit or a painting crew. The first week breaks the same way in all of them.

The gap you built at the interview

You know how a used truck ad reads. One owner, highway miles, always garaged. Then you crawl under it and the frame tells a different story. Nobody lied; they described the good version.

Interviews work the same way. You need a body by Monday, so the job gets its best coat of paint. Mostly finish work. Home before supper most nights.

You will be running your own crew before long. Every line is true on the right day. None of them is true on his first cold Tuesday in the mud.

John Flick, who owns a repair shop in Delaware County, Pennsylvania, said it plainly to Ratchet and Wrench in September 2026. "You can promise the moon and the stars, but if you don't back that up, then your word is nothing." Your word is the job offer. The pay is only the part you wrote down.

Honest sounds worse than you think, and it works better. Say the drive is 50 minutes each way and the first month is demo, cleanup and carrying. Winter starts early and it is cold.

Then say what changes, and what he has to do to get there. A man who hears all that and still takes the job has already made the decision you were hoping for.

Week one is where the promise gets tested

Picture his first morning. He shows up at the shop and nobody told the crew he was coming. He stands beside the van while you work out who to put him with. By Friday he still does not know what he is supposed to be getting good at.

That is the same failure as a job buried in the group text, aimed at a person instead of a job. He cannot see the week, so he assumes nobody else can either. It looks the same in an HVAC van and on a carpentry job.

Flick starts before that morning. He brings a new hire in for a walkthrough of the shop before the first day. The layout, the tools and the faces are not all new at once.

The survey says firms are moving the same way: structured onboarding, regular check-ins, formal jobsite orientations, and mentor or buddy systems. None of that needs software. It needs a decision.

Here's what this means for you

Take the hour. Take it before he accepts, not after he quits. Here is what goes on the page.

  • The honest version of the job, said out loud: start time, the drive, the weather, and how much of the first month is grunt work.
  • His first five days by name. Which job, which truck, which crew, in order.
  • One person he asks. Not you. A name, so he is never standing there guessing.
  • What good looks like by Friday. One or two things he can do on his own by then.
  • Three dates in your calendar: 30 days, 60 days, 90 days. Ten minutes each.

The check-ins matter more than the page. Ask two questions each time: what surprised you, and what do you still not know. The first answer tells you which line of your interview was optimistic. The second tells you what nobody has taught him yet.

Flick recommends those same three reviews, and his read on the clock is blunt. Within the first 30 days you can tell whether somebody works out, and by 90 days they are either staying or already gone. You are not adding paperwork. You are deciding on purpose instead of by accident.

Say you spend 60 minutes on a Sunday night, 2 weeks before he starts. Now imagine the other version, where he walks on day 9. You eat the wages, the truck seat, and a lead hand who spent his week answering questions. One of those two costs you nothing.

This is the same margin you protect by counting bid hours against burned hours and by pricing small jobs properly. A crew that turns over cannot build the reputation that carries word of mouth either.

Then show him the same picture every morning after that. We built OPS so the crew opens it, sees their jobs, and knows what to do with no training. The written week is the part only you can do.

The people who stay are not the ones who were promised the most. They are the ones who were told the truth, and then watched it hold.

Write down his first week, or write the job post twice.

Sources