When should a contractor hire another technician instead of buying more ads?

It's the second week of a heat wave. Both techs are running back-to-back, the next open slot is Thursday, and the ad rep wants to know if you'll up the budget for the summer push. Your service manager wants a third truck. One of them is right, maybe neither, and the money only stretches to one.

Here's the short answer. Hire when the work is already there and nobody can run it: the next open slot is further out than customers will wait, you're turning jobs away or pushing them, and it's been that way across a season, not one hot week. Buy ads when your techs have open hours and the jobs that fill them aren't coming in.

Before either, check a third leak: calls that already ring and nobody picks up. More ads send more calls into that gap, and a new tech can't close it because they're in an attic or under a sink, not at the phone. An office hire can, and so can the other phone fixes below.

Full disclosure: Willison sells call answering. It doesn't sell ads and it can't run a job, so we have an interest in that third leak. That's why the checks below come in order, why "hire" is the answer in two of the four outcomes, and why there's a section on when answering more calls makes a full board worse.

Three checks, in the order you run them

1. The board. How many open hours did each tech have, and how far out is the next open slot? Compare the same months last year, not just last week.
2. The work you turned away. Jobs declined, jobs pushed past when the customer wanted them, callers who heard "two weeks" and didn't book.
3. The phone. Calls in, calls a person answered, from your phone or carrier log.
4. Read them together. Full board and work turned away means hire. Open hours and a phone that's answered but quiet means buy. Open hours and real callers going unanswered, after you've checked the log by hand, means fix the phone first.

Everything you need is in records you already keep. The rest of this is how to read them without fooling yourself.

Check the board first

Pull the schedule for the last few months and count, per tech, the hours that sat open. Then note how far out the next open slot was each week. That's your capacity picture, and it beats any feeling about how busy things are, because busy and full aren't the same thing. A tech can feel slammed on a day that's two hours of driving and a return trip for parts.

Read it against the calendar. Open hours in a shoulder month aren't a signal to buy ads, and a booked-out July isn't proof you need another truck. A hire is a year-round cost, so the question is whether last year's busy stretch, and the one coming, would keep a new tech in billable work, not whether this week did.

Then split the board by job type. A schedule full of tune-ups and small repairs tells you something different from one full of replacements and installs. You can be booked solid on low-ticket service work and still be short of the replacement leads that pay the bills. In that case the hire might be a helper who frees your best tech up, and the ad spend might belong on the work you actually want more of.

Count the work you turned away or pushed

An open-hours count shows whether you have room. It doesn't show what you lost by not having it. For that, somebody has to write it down: every job you declined, every job pushed past when the customer wanted it, and every caller who heard the next date and said they'd think about it.

A job that never got booked may never become a record, so check whether your software captures it. If it doesn't, give the list to whoever answers the phone, the office manager or the CSR, and keep it for a month. If you're in the truck all day, it's better kept by someone at the desk. It just has to be somebody's job.

This list is the demand a new hire would actually run. If it's long and it's steady, a hire has work waiting on day one. If it's empty, another tech adds payroll and no jobs, however busy the shop feels.

Then check the phone

Your phone system or carrier keeps a call log. Count a few ordinary weeks: every inbound call, and how many a person answered. Leave out your own outbound calls. Watch for two blind spots. A call that rolls to voicemail can show up as connected, and a line forwarded to someone's cell may leave the record on that phone instead. Pull the voicemail count and the cell's call history for the same weeks, or your answer rate will look better than it is.

For a sense of scale, Invoca's 2026 home services benchmark found 34% of calls to HVAC companies were answered by a person, against 74% for plumbing companies. Those are averages across Invoca's own customers, every inbound call, and they're not your number. Your log is.

Don't read every unanswered call as a lost job, either. Invoca's figures include every inbound call, so misdials and quick hangups are in there. Across all home services, not broken out by trade, Invoca found that counting only calls over 15 seconds, 65% were answered. Over 30 seconds, it was 73%. And a length cutoff only clears the shortest calls. Vendors, existing customers checking on a visit, spam, and callers you or your tech rang back ten minutes later are all still in the pile. Go through a week of it by hand before you put a number on it.

What's left after that is the part that matters for this decision: real callers nobody reached. Those callers already found you. Call them back first, which costs nothing but time, though some will have booked someone else by then. Buying ads to make up for the rest means paying for calls, some of which you already got once.

Read the three together

Put the board, the turned-away list and the phone side by side, and you'll be in one of these places.

• Full board, work turned away, phone answered. Hire. The demand's there and you're running out of hands. More ads only lengthen the wait.
• Open hours, phone answered but quiet. Buy, or buy better. Before you add budget, check that the calls you do get are booking: a weak script or a price problem looks like low demand from the board. In a shoulder month, maintenance agreements can fill the schedule too.
• Open hours, real callers unanswered. Fix the phone first, then see whether you still need the ads. This only applies to what's left after the hand count. A handful of missed calls a month isn't a phone problem, and an empty board beside it is a demand problem.
• Full board, work turned away, phone missed. Hire, for the same reason as the first case: answered calls still need someone to run them. Alongside the hire, not instead of it: if your techs or you answer from the field, every call they take is time off a job, so moving the phone off them hands some hours back while the search runs.

Two more cases. Open hours and work turned away at the same time points to a mismatch: the slots are open but not for the work coming in, which is the job-type split above. And a full board with nothing turned away is a shop running at capacity, where neither move is urgent. A full board is a price signal too. If you're turning work away week after week, it's worth checking whether your rates are where the market is before you add a payroll line to keep up with demand you may be underpricing.

Price a hire honestly

"Hire" doesn't have to mean a third full tech, and it doesn't have to mean hiring at all yet. Overtime, subbing out installs, or cross-training someone you already have can cover a stretch. A helper or apprentice can take the setup, the hauling and the second set of hands off your best tech so more of that tech's day is billable. A CSR or office hire fixes a phone leak a tech never could. Match the hire to what the three checks found.

For a full tech, start with the wage and don't stop there. The BLS puts the median hourly wage for HVAC and refrigeration mechanics at $29.33 and for plumbers and pipefitters at $30.67 (OEWS, 2025). That's the middle of the market, so half earn more, and the tech you want, one who can run calls alone, may be one of them. On top of it sit payroll taxes, benefits, on-call pay, a truck, tools, and the weeks before a new tech can work without someone checking their jobs. Those are your numbers to fill in, from your own books.

Then there's the part no spreadsheet shows. Paid search ads can be paused the same day if the work dries up. A hire can't, and letting someone go after a slow fall costs you more than money. Some ad spend isn't quick to undo either: a lead contract, a direct-mail run or a marketing retainer can run for months. But a hire is the hardest of all to walk back, which is why the bar for it is backlog that lasts, not one good month.

And hiring assumes you can find someone. Finding a tech who can run calls alone may not be quick. If that's where you are, the turned-away list is the case for starting the search now, before next season, not after it.

Price the phone fixes the same way

If your checks land on the phone, the fixes have costs and failure modes too, and they get the same treatment as a hire.

• Calling back from the log. Free, apart from time. It fails when nobody owns it, and some callers will have booked elsewhere before you reach them.
• An office hire or CSR. Someone who knows your customers, your schedule and your prices, and can juggle the board as calls come in. It's a payroll line with its own hiring search, and one person takes one call at a time during the hours they work.
• A human answering service. A trained person on the line, including nights and weekends. It's a monthly bill that varies by service, and it's only as good as the script it works from.
• Willison. Answers at any hour and books qualified jobs straight to your calendar. It's a monthly subscription, priced on the pricing page, plus the setup time before it goes live. It doesn't dispatch or run a job, it can mishear an address or misclassify a job, its intake is only as good as the questions it's built around for your shop, and some callers won't want to talk to a virtual receptionist.

Whichever you try, recount a month later the same way as the first time, and compare like weeks so a change of season doesn't fake the result. For callbacks, count the missed callers you reached and how many booked. For the other fixes, count the answer rate and the booked jobs. If answering went up and booked jobs didn't, the phone wasn't the leak, and the board and the turned-away list are where the answer is.

Where this goes wrong

• Reading a spike as a trend. A heat wave can fill an HVAC board. Hire on that and October finds a tech with no work. Compare the same stretch last year before you decide.
• Hiring to fix a schedule problem. Long drive times, return trips for parts and callbacks for redo work eat hours that look like a capacity shortage on the board. Another truck adds more of the same.
• Buying ads to fill a board the phone emptied. If the open hours sit beside real callers nobody reached, some of that work may have been there. Check the log by hand first, because more ads into the same phone add to the same pile.
• Answering more calls into a full board, including with us. Willison answers the call and books it to the calendar you set. If the next open slot is three weeks out, it books three weeks out, and that caller can still go elsewhere. An AI receptionist can also mishear an address or misclassify a job, which is why it reads the address back to the caller before it books. It fills the room you have. It doesn't make more.

Where Willison fits, and where it doesn't

Willison belongs where the phone is the leak: the open-hours case with real callers going unanswered, and, alongside a hire, the full-board case where you or your techs answer from the field. It answers in seconds at any hour, answers every call at once so the caller who rings at 7pm on a quiet October evening gets the same pickup as the one at 10am, qualifies the job, and books qualified jobs into the open slots on your calendar. It won't run a job, and it won't add an hour to a full board.

If your checks point at a hire, hire. If they point at the phone, press play on a real call on willisonhq.com and hear what an answered call sounds like, start to finish. It's our receptionist answering for a sample HVAC shop; a plumbing build runs the same engine around your own services. For the numbers behind the buying side, see whether buying more leads beats closing the ones you get.

Frequently asked questions

It depends on the job. Someone with no heat in January won't wait until next week, while a tune-up customer may be fine with a date a couple of weeks out. The useful test isn't a number of days, it's whether people accept the date you offer. If a growing share of callers hear the next opening and don't book, you're past it for that kind of work.

If your phone check shows a lot of unanswered calls and your techs or you are answering from jobs, an office hire may be the cheaper first move, because it protects the billable hours you already pay for. If the phone's covered and work is still being turned away, a CSR won't run any of it. That's the tech.

No. It's the median hourly wage across the whole occupation, so half of workers earn more, and experienced techs who can run calls alone may sit above it. It also leaves out payroll taxes, benefits, on-call pay, the truck, tools and the training weeks. Treat it as a starting point for a tech who can run calls alone, then add your own figures for the rest.

Cut back rather than switch off. While you're booked out, extra calls turn into longer waits, so trim the weakest source first, the one that books least for what it costs. Keep enough running that a new tech walks into a schedule with work on it, because a hire who starts on an empty board is the expensive version of this decision.