The Trades Shortage Isn't a Hiring Problem

Nearly forty thousand HVAC techs short this year. Over twenty thousand plumbers. As many as 2.1 million skilled trade jobs unfilled by 2030 — the technician numbers from industry projections, the 2030 figure from the US Department of Education, by way of JLL. Nobody is building a pipeline fast enough to close that gap — the trade schools can't graduate at that rate, the apprenticeships take four years, and the bodies don't exist.

So the industry bet on wages. Pay more, attract more. The HVAC median cleared $57k last year and the shortage got worse. Wages will keep climbing and the gap will keep widening because this was never a compensation problem. You cannot hire your way out of a structural shortage. The math doesn't work when the supply side is capped.

Higher wages pull a lever that isn't attached to anything

Wages work when the labor exists and is choosing to go elsewhere. That's not what's happening here. Enrollment in trade programs has grown — and it still doesn't cover the retirement rate. Three tradesmen retire for every one coming in. If every kid who enrolled in an HVAC program this year graduated and hit the field tomorrow, you'd still be short by a margin that's embarrassing to calculate.

So wages are pulling a lever that isn't attached to anything. The demand side is real. The supply side won't respond fast enough regardless of what you pay.

The shortage is structural. The fix has to be structural. Not a recruiting problem. Not a compensation problem. An allocation problem.

Where the licensed hours go

A licensed HVAC tech or plumber can run eight to ten billable hours on a good day. Here's what happens instead: he takes the 7 AM missed call himself, spends twelve minutes explaining the estimate to a homeowner who wants to know why the part costs what it costs, waits on hold with the supply house, texts the office the parts list, and fields a callback from a customer whose window ran long.

None of that requires a license. All of it burns licensed hours.

Field service benchmarks put non-billable time at around 30 percent of a technician's day — and while some of that is drive time you can't route anywhere, a large share is dispatch, quote follow-up, customer calls and job notes. Nearly three quarters of technicians say they spend too much time on paperwork. You're paying for diagnostic and installation capacity and spending it on tasks with no licensing requirement.

The work that doesn't need a license

graph TD
  H[Eight hours of licensed time] --> A[Diagnose and install: the work only he can do]
  H --> B[Admin bleed: 25-40% of the day]
  B --> C[Return missed calls and book]
  B --> D[Send quotes and chase them]
  B --> E[Status calls and job notes]
  C --> R[Route to a system that runs around the clock]
  D --> R
  E --> R
  R --> G[Two hours a day back, per tech, with no new hire]

Draw the line hard: anything a licensed tech does that doesn't require his ticket is waste. The list is longer than most owners will admit.

• Return missed calls and book the appointment
• Send the quote and follow up when it goes quiet
• Field status calls from customers whose window is running long
• Capture parts lists and job notes after the visit
• Schedule the next maintenance window before the tech leaves the driveway

That's the portfolio burning licensed labor right now. Route every item on that list to a system that runs around the clock and doesn't need a truck — and the equation changes without a single new hire.

What reallocation recovers, per crew

If your tech loses two hours a day to unlicensed work, that's 25 percent of an eight-hour shift. On a four-tech crew, that's one full tech recovered — no hire, no raise, no four-year wait on an apprentice.

The play is concrete: audit one week of tech time. Log every task. Mark anything that didn't require the license. Build the routing for those items — inbound call handling, quote follow-up, appointment confirmation — and get them off the tech's plate before next Monday.

An AI dispatch and follow-up system that handles inbound at 7 PM costs less than $100 a month. A licensed tech costs $55k a year and takes years to produce — price it against the hire, not the app. The ratio is not subtle.

And it compounds: the same recurring capacity a buyer pays a premium for is the capacity you're burning on paperwork — which is what decides the multiple when you sell.

The shortage is real. The people aren't coming fast enough. But the capacity you need is already there — it's just burning on the wrong work. Stop spending licensed hours on unlicensed tasks and you just found the tech you couldn't hire.