HVAC Marketing in the Off Season: Where the Money Actually Is

The shoulder is when most HVAC owners cut ad spend and wait for weather. That is the expensive version of the decision. A lead you do not generate in the slow months is not a lost lead, it is a technician you cannot keep busy, and payroll does not take the shoulder off.

The trade almost never markets on the next part. ENERGY STAR's maintenance guidance tells homeowners to have the cooling system checked in spring and the heating system checked in fall, and it says why out loud: contractors get busy once summer and winter come. The federal guidance your customer is already pointed at tells them to call you in April.

What the Shoulder Actually Changes

The slow months are not a smaller version of peak. They are a different buyer.

At peak you sell to whoever is broken today. No shopping, no deliberation, and you take the jobs in the order they fail. In the shoulder you sell to whoever is going to break. Same houses, different conversation. Not the same customer. Not the same job.

That distinction is the whole argument. Every offer below works because the person on the other end still has time to think, and the person calling you in July does not.

The Installed Base Is Older Than the Phone Suggests

Air conditioning sits in nearly nine in ten American homes, so the reachable base in your territory is close to every house with ductwork in it. That part is not surprising. The age of it is.

Add up the age buckets in the last federal count and it gets specific. Of the 109.51 million households using air conditioning, roughly 42 million had a main unit ten years old or older, and about 21.5 million were at fifteen or more. Nearly four in ten past ten years, one in five past fifteen. The federal table publishes counts per bucket, not percentages. The arithmetic is ours.

Say the year plainly, because it changes how you read the number. That survey was fielded in 2020 and it is still the most recent federal count of equipment age by end use. Absent replacement, every one of those systems is six years older now. A floor, not a snapshot.

Now hold that against the federal replacement thresholds, which put the consideration point at more than ten years for an air conditioner or heat pump and more than fifteen for a furnace or boiler. A large share of the equipment in your service area is already past the line the government draws in its own consumer guidance.

Almost none of those homeowners are thinking about it in April. In July they will not be thinking about it either. They will be reacting to a failure, and you will be quoting a system to somebody standing in a house that is already too warm.

Three Things Worth Selling Before the Weather Turns

A shoulder campaign is not the peak campaign at a lower budget. It is a different offer set, aimed at people who are not in trouble yet.

Maintenance, Sold as a Scheduling Problem

The pitch that works is the one the federal guidance already wrote for you: book the cooling check in spring, because your contractor will be buried in July. That is a scheduling argument, not a price argument. Sell the calendar, not the coupon.

The counter-argument deserves airtime, because plenty of good operators run it. A discounted tune-up fills a dead board, gets a technician into the house, and creates the inspection that finds the weak capacitor or the cracked heat exchanger. Volume now, pull-through later. That is a real mechanism, and the size of the discount is not what makes it work.

We still land the other way, and this is a position, not a coin flip. A discount-led tune-up selects for the customer shopping on price, who is the customer least likely to sign a full system replacement eighteen months later. A scheduling-led tune-up selects for the customer who plans. Two ads, two address books.

Maintenance agreements belong here as a mechanism, not a promotion, and without the return figures circulating in the trade press, none of which survive a look at their sourcing. What a plan does is structural: the holder has agreed to let you in twice a year, so you see the equipment age before the homeowner feels it. Run the economics on your own book rather than on anyone else's: a plan holder is the only customer whose replacement you can schedule instead of wait for, which is the whole of the return and the reason it does not need a number attached to be worth selling.

Planned Replacement, Sold to the Age Bucket

The second offer targets the system rather than the emergency. The same federal guidance lists the symptoms a homeowner can identify without a technician standing there, and those symptoms are your targeting criteria.

• •Age first: an air conditioner or heat pump past ten years, or a furnace or boiler past fifteen, which is where the federal replacement thresholds put the consideration point
• •Energy bills climbing while the repair calls get more frequent
• •Rooms that will not hold temperature no matter where the thermostat sits
• •Humidity the system cannot pull down in summer
• •Dust that comes back a week after the house was cleaned
• •A unit that has gotten noticeably louder

Every one of those is something a person notices about their own house on a quiet Saturday. None register during a failure, because a failure has one symptom and it is that the house is hot. The structure underneath is ordinary: a cold prospecting layer, a retargeting layer, a lead form, the same architecture that works for local service businesses on Meta year round.

The third offer is not really an offer, it is a list. Your past customers are the cheapest revenue in a slow month, and when we audit an HVAC account in the shoulder they are almost always the thing left untouched. Automated follow-up and reactivation is the least glamorous row below and usually the highest return, because you already paid for that audience once.

Nobody Can Tell You What a Shoulder Lead Costs

Start with what is knowable. In the home services benchmark data, over a window ending March 2025, air conditioning installation and repair ran $127.74 per lead and heating and furnaces ran $129.02, against a $90.92 average across home services. HVAC leads cost roughly forty percent more than the category they sit inside.

WordStream's 2026 search benchmarks give a fresher read on a wider bucket: $8.33 per click across home and home improvement, from 13,474 US search campaigns running between April 2025 and March 2026. Their cost per lead lands at $90.92 as well, the same figure to the cent as the older category average above, which is worth noticing before you treat either number as precise. Neither dataset breaks HVAC out by month, and that gap matters.

It matters because the claim every article on this subject makes is that clicks get cheaper when demand falls. It is repeated nearly everywhere and no public dataset verifies it at the trade level. The people repeating it may well be right, they just cannot show you. If your own account shows a seasonal curve, believe it over any blog, this one included.

Here is the version that holds either way. Cost per lead is one half of the equation and the half you control least. A shoulder lead becomes a scheduled install with a chosen date, a crew planned in advance, and equipment ordered instead of grabbed off the shelf. A peak lead becomes an emergency: whatever is on the truck, whoever is free, overtime attached.

A lead price means nothing without the job underneath it, and this is where most HVAC owners flinch at the wrong number. A contractor software vendor's HVAC pricing guide, published April 2026 and scoped to typical residential work across the US, puts a furnace install at $3,000 to $7,000, central air at $3,500 to $10,000, and a heat pump at $4,500 to $12,000 for a complete single system install, with service call fees running $70 to $200. That is a vendor guide rather than an association survey, and equipment pricing moves with tariffs, refrigerant rules and labour, so check the date before you lean on it.

Run it against your own average ticket rather than that range. Say one booked job in four, purely as an illustration and not a rate anybody measured for you, and a $128 lead puts about $512 of ad spend into a job you closed. Against a $6,000 install, taken from the middle of the range above, that is well under a tenth of the ticket. Against a $150 service call from the same source it is a losing trade unless the call becomes something. Same $128 either way, which is the entire shoulder argument in one line: in April it buys a replacement conversation, in July it buys a diagnostic.

Same $128 at the top of the funnel, very different job at the bottom. And a $128 lead that sits in an inbox for four hours costs a great deal more than $128, in the shoulder and at peak alike.

The Calendar, Counted Back From Peak

One caveat. This is written for a market with two real shoulders, where heating and cooling both carry weight. If you sell somewhere with one long season, the sequence still applies, you just run it once a year instead of twice.

Count backwards from peak rather than forwards from today, because that is how the schedule gets built. The ramp point is two weeks out, the same two weeks on either side of the year.

1. 1.Early shoulder, both at once: fix what you cannot fix in July, and run the maintenance offer on scheduling instead of price. The first is the landing page, call tracking, review responses, and the Google Business Profile nobody has opened since the last busy season. The second is the cheapest way to get a technician standing in front of aging equipment before it quits.
2. 2.Mid shoulder, again both at once: turn on the planned replacement campaign against the age and symptom criteria above, and work the reactivation list. Replacement is the longest sales cycle in this post and it needs the runway. Reactivation costs the least and is the first thing skipped once everybody gets busy.
3. 3.Two weeks before peak: raise budgets and change the message from planning to availability. Two weeks is the interval on both sides of the year, cooling and heating alike.

Into peak, let the emergency campaign carry the spend and leave the replacement campaign running against the people who told you no in the shoulder. Some of them are about to find out you were right.

The shoulder is not the part of the year when marketing stops working. It is the only part when you get to decide which jobs you are running in July: the ones you scheduled, or the ones that found you. The equipment in your territory will fail on its own timetable regardless. You get one window on each side of the year to have that conversation first.