Tech Onboarding Is a Revenue System. I Learned That on a Saturday in My Garage.

Six weeks after my first hire at Reeves Electric, I had a callback problem I couldn't explain.

The tech was capable. Good instincts, clean work, customers liked him. But his callback rate was running high — I pulled four hours of job records on a Saturday trying to figure out if it was a skill gap, a scope problem, or something he was saying to customers that set wrong expectations. I couldn't tell. Not because the answer wasn't in the data, but because I hadn't built any checkpoints in week one that would have caught which habit was broken.

Turned out to be one thing. He wasn't noting what he left uninspected. Not incompetence. Not a bad hire. An untaught habit — because I never taught it. I handed him a truck at week six and assumed six weeks of riding around with me had transmitted everything relevant. It hadn't.

That Saturday cost me more than four hours. It cost me eight weeks of money I was losing and couldn't find, and a customer conversation I didn't want to have. I built the onboarding system the following week.

The First 30 Days Sets the Output Ceiling

Onboarding isn't orientation. It's habit formation. Whatever a tech learns to do in month one — how they write a ticket, how they communicate scope to a homeowner, how they handle an unexpected panel condition — that's what they'll do for years. Habits form fast in field work. There's no one watching.

Most small shops treat the first week as supervised babysitting. New guy rides with someone good, asks questions, meets the team. By week two he's got his own truck. By week six he's flying solo. What the shop is actually doing is inheriting whatever the senior tech's unexamined habits are, including the ones that generate callbacks.

I rebuilt my intake process in the first ninety days of running Reeves Electric. My NPS in month two was a 4.

By month nine it was an 81. But that paragraph usually gets the tidy lesson attached — "I listened to calls and found the broken patterns." Here's the messier version: I didn't know what was broken for weeks. I just knew the number was a 4. I started recording every call and listening back on Sunday mornings with a notepad, and I found six distinct problems across fourteen calls before I understood what I was actually fixing. The same forensic approach is what you need for field tickets.

You don't get good outcomes by hoping good habits transfer by proximity.

Most Shops Don't Have an Onboarding System. They Have a Tradition.

At Lonestar Electrical Services, where I ran service jobs from 2014 to 2021, undocumented handoffs got people hurt or fired. Forty trucks, commercial scope, actual project schedules. There were written procedures for how work got transferred between crews, how scope changes got documented, what a completed job package looked like. Not because anyone loved paperwork — at that scale, "just figure it out" costs real money and real liability.

Here's what I find strange: small residential shops, where the margin per truck is tighter, are the ones flying without documentation. One callback on a $900 job costs you more than a large outfit — there's no volume to absorb it. One tech developing a bad habit in month one runs that habit on every ticket for the next two years.

"Our guys are pretty solid" is not a measurement. If you haven't defined what good looks like in week one, you have no idea what you're producing by week twelve.

I cannot stand "most of our work comes from Google" as a business answer. That's a guess. Same thing here. "Our guys are pretty solid" tells you nothing about your actual output.

Ask yourself what a new hire's first day looks like. If you can answer it in a sentence — "rides with Mike, asks questions, meets the team" — you don't have a system. You have a tradition. Traditions feel like process because they're repeated. They're not process, because they're not written and nobody owns them and you can't improve what you can't examine.

"Ask me if you have questions" is an open door with no threshold. It's not a system.

What a Documented 30-Day System Actually Contains

A checklist. With owners and checkpoints. That's it.

Week one is observation and intake, not truck riding. At Reeves Electric, a new tech spends their first morning sitting with my dispatcher, not in a truck. They listen to two live intake calls. We pull two CallRail recordings from the previous week — one clean booking, one that didn't convert — and they hear both before they've touched a single tool.

If a tech doesn't understand what the customer was told before the truck rolled, they're working with half the picture. The customer asking about an EV charger install has a number in their head from somewhere. The intake call either managed that expectation or it didn't. The tech on-site is downstream of that conversation. They need to know what was said.

Days 8 through 15 introduce technical checklists tied to our actual ticket types. Not generic NEC review. The jobs we actually run: AFCI/GFCI installs, panel diagnostics, EV charger pre-assessments. Each one has a written checklist that maps to how we document scope, what we photograph, and what we flag for follow-up. I'm not testing code knowledge. I'm building the documentation habit — the one that breaks down under time pressure when they're on their own.

By day 30, every new tech does what I call the shadow invoice audit. They write the ticket on a completed job. My dispatcher writes it separately, from the same job notes. They compare. Where the two versions diverge is exactly where the habit isn't formed yet. We've caught missed scope items, inconsistent parts notation, vague customer communication notes. Small things on day 30. Callbacks on day 90.

The total calendar time is not dramatic. It's specific, it's written, and one person owns each piece. That's what makes it a system instead of a tradition.

Your best tech is your best tech because of thousands of service calls. Pattern recognition. Field judgment. That's real and genuinely valuable. It is not something a new hire absorbs by watching.

When your senior tech says "you'll just get a feel for it," they're describing something they can't explain because they've never had to write it down. They've internalized it. It's invisible to them now. So the new hire watches, tries to absorb the rhythm, picks up some of it — and also picks up the shortcuts. The panel inspection steps that get skipped when the job's running long. The customer conversation that gets abbreviated because your senior tech can read the room. The notation habits that work fine until someone else has to read the ticket.

Your gut after 4,000 service calls is data. Your gut after 80 is hope. That applies to your senior tech's "feel" too. It's his data, accumulated over years. Watching someone use data doesn't transfer it to you.

I post our full benefits package on the careers page. Openly. Secrecy about compensation is a tax on hiring quality — candidates fill the information vacuum with pessimism. Same logic applies to process. If it's not written down, it doesn't exist as a system. It exists as a personality. Personalities leave. Personalities vary by how busy the day is, by how much the senior tech clicked with the new hire. A written system doesn't.

Your best tech is one input into the onboarding system. He's not the system.

The Real Cost Is in the Callbacks You Never Attributed Correctly

Back to that Saturday in 2022.

When I pulled the job records, I was looking for a clean answer. Technical? Scope? Communication? What I found was ambiguity — because I had no week-one checkpoint data to compare against. I couldn't tell if this was a habit the tech came in with or one that formed while he was riding with me. I had no baseline because I'd never built one.

The habit generating callbacks was one documentation step: he wasn't noting what he left uninspected. On jobs with partial scope — where we touch one thing and flag another for a follow-up quote — he was closing the ticket without writing down what he'd looked at and what he hadn't. Customer calls back two weeks later about the thing we didn't touch. Tech doesn't remember the specifics. Nothing in the ticket. The customer's experience is that something got missed. The review reflects that.

One untaught step. Across dozens of tickets before I caught it.

I sample calls weekly — two that booked, two that didn't. I now do the same with tickets for any new hire's first sixty days. Two clean jobs and two with a flag (a callback, a follow-up dispatch, a customer note that reads lukewarm or worse) sampled every week. I'm not trying to catch failure. I'm trying to see the pattern before it's locked in. Four weeks of ticket sampling tells me more about a new hire's documentation habits than six months of assuming things are fine.

The measurement practice for intake calls and field tickets is the same. Record, sample, find the pattern, fix it before it runs for two years.

What to Build by Next Monday Morning

Not a full onboarding program. Three things.

A day-one checklist. Twelve to fifteen items, specific, owned by a named person, signed off by end of day one. At Reeves Electric, that person is my dispatcher — not me, not the senior tech. If you can't name who runs it right now, it won't run. The checklist lives in one Google Doc. Items include: tool inventory confirmed, truck stock reviewed, first week schedule briefed, CallRail extension assigned, first intake observation scheduled. Specific. Owned. Done.

The three-recording session. Pull three CallRail recordings from the last thirty days before the new hire's first solo dispatch. One clean booking. One callback job. One call that didn't convert. Have the new tech listen to all three. Don't narrate. Let them hear what the path from call to invoice sounds like from the customer's side before they've driven the truck alone. This takes forty-five minutes. It reframes the technical work in a way no safety briefing touches. The customer's own voice does more than your explanation will.

A 30-day ticket review on the schedule. Not a whenever-you-feel-like-it conversation. A standing thirty-minute slot, same time each week, for the first thirty days. You or your dispatcher review two of the new hire's tickets. Flag what's missing, what's inconsistent, what would confuse a customer reading it three weeks later. Show them what a clean version looks like. Four sessions. By day 30 the habit is either forming or you know exactly where it isn't — and you can fix it before it becomes two years of callbacks you can't explain.

A Google Doc. A CallRail account you probably already have. Thirty minutes a week for a month.

Start with the checklist. Name the person who owns it. Do it before Friday.

FAQ

How long should the onboarding process actually run?

Thirty days of structured activity, then ticket sampling for thirty more. After sixty days you have a real picture — not a gut feeling, an actual record. After that the formal structure backs off, but the weekly ticket sampling stays in my shop indefinitely. Two tickets, thirty minutes. The return is worth it.

What if I only have one or two trucks and can't take a tech off billable work?

You're not taking them off billable work. The morning intake session is one morning, not a week. The recording session is forty-five minutes. The day-one checklist is an hour. Most of this fits inside a normal first week — it just happens in a different sequence than you're used to. The truck still rolls.

My best tech has been with me for six years. Is it insulting to tell him he can't run onboarding?

Frame it as protecting his time. His job is to produce quality field work. Asking him to translate six years of internalized judgment into repeatable steps for a new hire is a completely different job — one he hasn't signed up for and shouldn't be expected to do well without a written system to work from. You're not cutting him out of onboarding. You're giving him structure so he's not reinventing it every time you hire someone.

What's the most common mistake at the five-truck level?

Treating the ninety-day probation window as the onboarding window. They're not the same thing. Probation is about fit and retention. Habit formation happens in the first thirty days. If you wait until the ninety-day check-in to assess what someone learned, you've already paid for whatever they taught themselves in the gap. Probation can run ninety days. Onboarding needs to run hard in the first thirty.

Do I need special software?

No. Google Doc for the checklist, your existing field management software (Jobber, Service Fusion, whatever you're running) for ticket review, CallRail for recordings. Build the process in Google Docs, run it twice, find where it breaks. Then decide if software would help. Buying software for a process you haven't defined yet is just expensive distraction.

What if a new hire pushes back and says they already know how to do the job?

That's fine. The onboarding isn't about their technical skill. I'm not retraining journeymen on how to pull wire. The first thirty days is about how we do the job: how we document, how we write a ticket a dispatcher can read without calling the tech for clarification, how we communicate partial scope to a customer so they don't call back confused. That's specific to Reeves Electric. It's not a comment on their experience. Lead with that framing and the pushback usually stops.

Your Tech Onboarding Is a Revenue System — Not a Formality