HVAC marketing: a season-proof system for booked calendars

Most HVAC work goes to a company the household can already name. FIELDBOSS surveyed 1,000 US homeowners who had paid an HVAC professional in the previous twelve months. 42.6% called a company they had used before, and another 26.3% called one that a friend or a relative named. Online reviews and marketplace apps such as Yelp and HomeAdvisor accounted for 7.7% of those hires.

Two thirds of the work in that survey ran on memory, and memory gets built long before a compressor quits on a Saturday in July. So the budget has two jobs rather than one. It has to reach the household with nobody to call, and it has to hold every household that has already paid you once. Most advice in this trade covers the first job and skips the second, which is how a shop ends up buying leads all year and still watching March go quiet.

The market itself is close to saturated. About 88% of US households run air conditioning, and two thirds use central AC or a central heat pump as their main cooling equipment, according to the Energy Information Administration. IBISWorld puts the trade at $159.4 billion across roughly 120,000 US businesses in 2026. Almost every household becomes somebody's customer, and the shop that gets the call is the one a homeowner can name without looking it up, in HVAC as across construction and trades.

Two peaks and two troughs, in the same order every year. Cooling work builds from April and runs hard through June and July. Heating work starts at the first hard freeze and holds into February. The gaps fall in early spring and again in early fall, roughly March into April and September into October, and those are the weeks when trucks sit or owners overpay for clicks.

How deep the swing runs depends on where those trucks are. 93% of Southern households and 92% of Midwestern households use air conditioning, against 73% in the West, and half of Northeastern households cool with individual units rather than a central system. A Phoenix shop and a Monterey shop do not share a demand curve, and a budget calendar copied from a Sun Belt blog will misprice both. Roofers live with a sharper version of the same problem, where one hailstorm decides the quarter.

Three triggers are worth building the calendar around. A first heat wave or a hard freeze arrives on its own schedule and decides only who gets the call. Equipment age is the trigger a shop can see coming, because the install dates already sit in its own records. The pre-season tune-up window is the one a shop makes for itself, out of a list rather than a budget.

A diagnostic call and a full system change-out both start with the same phone ring, and they differ by a factor of a hundred. That spread is what makes an HVAC budget hard to set, because the average job is an average of two very different businesses.

One replacement is worth fifteen to twenty-five repairs, and the household that buys a system almost always bought a repair from somebody first. So the cheapest line in the table does the most work. A maintenance visit at $70 to $200 puts a technician in front of the equipment a year or two before it fails, with the install date written down and the homeowner watching a professional work.

The two kinds of job are also won in different ways. An emergency is won in the ten minutes after the air stops blowing cold: 29.8% of the homeowners FIELDBOSS surveyed expected service within a few hours, and another 44.0% expected it by the next day. A replacement is won slowly, on a quote that sits in an inbox for three weeks while a couple argues about the money. The first job needs answering, and the second needs following up.

The published benchmarks are close enough to plan against. SearchLight Digital measured 816 HVAC and plumbing contractors running 8,077 campaigns and $14.88 million of Google Ads spend in January 2026, producing 143,008 leads at a blended $104 each. PPC Chief, working from a separate dataset, puts the average HVAC click at $9.12 with a usual band of $6.84 to $12.31, against $5.26 across all industries. San Francisco runs near $14.59 a click, and rural and suburban markets sit about 20% below the national figure.

What a booked job costs

Book rates finish the sentence that cost per lead starts.

On cost per lead, branded search looks four times cheaper than non-branded. Measured per booked job it runs six and a half times cheaper, and that gap is the argument for spending on brand, reviews and the map pack before spending in the auction. A branded click is cheap because the person typing your name has already chosen you, and the work that earned the decision happened months earlier.

A workable budget follows the calendar rather than the month. Branded search stays on all year as cheap defense. Non-branded and Performance Max scale up into the two peaks and throttle down through the shoulder weeks, when a message to your own list costs less than any auction. For scale, WordStream's home improvement benchmark, built on 18,228 campaigns between December 2020 and February 2022, put clicks at $6.11 and leads at $54.05. HVAC costs in 2026 sit well above that older cross-category floor.

49.2% of the homeowners in the FIELDBOSS survey said their last HVAC visit was routine maintenance, against 31.8% who called about a breakdown. Most contact in this trade is scheduled, and scheduled contact is the cheapest kind to start, because it goes to a list you own instead of an auction you have to win.

Sold as an operations product, an agreement is $150 to $500 a year for two visits, and no owner gets excited about that line. Sold as a marketing product, it turns a $396 acquisition into the repeat business 42.6% of homeowners default to, and it books the visit where a technician stands in front of ten-year-old equipment and writes down its age. One agency benchmark puts sign-up after a repair between 12% and 25%. The report behind it names no sample and no period, so treat the range as a target to beat rather than a number to plan on.

The email program underneath the agreement is plain, and most shops already hold the parts. Four fields on the customer record decide who hears from you and when: the equipment install date, the agreement renewal date, the date of the last visit, and whether a replacement quote is open. Export those from your field-service software and keep them current, and the segments write themselves. Sending takes a verified domain with SPF, DKIM and DMARC records in place, one from-address a homeowner will recognize, and plain templates that survive being read on a phone in a driveway. That is the whole setup an email program in this trade needs before anyone talks about design.

1. 1

   Enroll the agreement while the technician is still in the house

   The invoice is in the customer's hand and the problem was fixed ten minutes ago. Every platform below can attach a recurring plan to a job, so the software records a conversation the technician already had.
2. 2

   Book pre-season tune-ups off the member list in March and September

   One email and one text to households who already pay you. This is the flow that fills the shoulder months, and its cost per booked job is the cost of sending the message.
3. 3

   Answer inside five minutes, and text back every missed call

   Almost 30% of homeowners expect service within a few hours. Google states that if you regularly fail to answer calls or respond to messages on Local Services Ads, your ad ranking may be affected. Speed raises the yield on every channel at once, which is the case for automated follow-up.
4. 4

   Ask for a review on every completed job, week in and week out

   74% of consumers look for reviews from the last three months, so a quarterly push leaves the profile looking abandoned nine weeks out of twelve.
5. 5

   Start the replacement conversation at year ten of equipment age

   The install dates sit in your records already. A light note at ten years and a firmer one at fifteen works, and the exact year matters less than having one.
6. 6

   Follow an unsold replacement quote for thirty days

   A $9,000 to $16,000 decision does not close on the first visit, and five to seven touches over a month cost nothing while working the most valuable job in the business. The mechanics sit in our guide to email automation flows.

We watched two of these ideas carry a shop in a nearby trade. Our work with 101 Exterminators took a pest control company from $120K to $6.8M in two years, on local pages that answered the searches people typed and a response system that let no lead sit.

Fewer than most owners assume, collected far more often than most shops manage. BrightLocal's 2026 survey of 1,002 US adults found that 97% read reviews for local businesses and 41% always read them. 74% look for reviews written in the last three months, and 32% want them from the last two weeks. 85% say positive reviews make them more likely to use a business. A shop with 60 reviews arriving every week reads better, to a homeowner and to Google, than a shop with 300 that stopped in 2024.

The same survey found consumers using AI tools to find a local business rose from 6% in 2025 to 45% in 2026, which makes those tools the third most popular recommendation source. They read structured, factual pages: services named in plain language, a service area a machine can parse, hours, license numbers, and the equipment brands you carry. That is ordinary local search work, and it now feeds two ranking systems instead of one. If you want to see what a machine finds on your site today, our free SEO report runs the check in about thirty seconds and orders the fixes by impact.

Local Services Ads sit at the top of local search results and charge per lead rather than per click. Google gates them behind license and insurance checks, plus a background screen on the owner and the field techs. The catch is operational. Google states that if you regularly fail to answer calls or respond to messages, your ad ranking may be affected, so the channel pays the shop that picks up on the second ring.

The platform is not the strategy. Every product in the table can run a renewal sequence and a review request, and few shops turn either one on. Pick the cheapest tier that supports recurring plans and automated reminders, then spend the difference on the list.

A three-truck shop with no maintenance base does not have a software problem. The order that works runs opposite to the sales pitch: build the list, run one renewal sequence by hand for a season, then buy the tier that automates what you proved works.

Four numbers repeat across the highest-ranking pages in this category, and none of them survives a check. Those pages rank anyway, which is how the numbers keep spreading into slide decks and budget plans.

• Retention costs five to seven times less than acquisition. Repeated everywhere, traced to nothing. One vendor's version, $200 to $300 to win a customer against about $40 to keep one, runs with no method and no sample behind it.
• A 5% lift in retention raises customer spend 25%. This is a garbled version of Reichheld and Sasser's 1990 finding about profit, not spend, and the range has been contested for decades.
• Email returns $40 for every $1 spent. The lineage runs back to self-reported marketer surveys, never audited, and never specific to HVAC.
• 82% of customers read online reviews. The measured 2026 figure is 97%, from a named survey of 1,002 US adults.

Renewal-rate claims deserve the same treatment. Agency posts quote 70% to 85% agreement renewal, sometimes 90%, with no dataset, sample or period behind them. Set a renewal target and measure your own against it. A plan built on borrowed benchmarks is a plan built on somebody else's imagination.

The first heat wave will arrive when it arrives, and no ad account has ever moved that date. What a shop controls is how many households already have its number in a drawer by then, and that gets decided in March and September, on the days the phone is quiet.