Start with the fact that decides everything else: nobody opens Instagram looking for a furnace.
Google works for HVAC because someone whose AC just died types "AC repair near me" and you appear. The intent existed before the ad did. Meta has no equivalent. Every person who sees your ad is doing something else, and none of them woke up needing you.
This is not a reason to skip Meta. It is the reason most HVAC Meta campaigns fail: they are built as if intent is there, using search logic, on a platform where search logic does not apply.
The short answer
Use Meta to sell tune-ups, maintenance plans and financing-led replacements before the season breaks, and use search to catch the emergencies. Send traffic to your own landing page rather than an instant form unless you have the follow-up speed to survive a colder lead, and judge every campaign on cost per booked job rather than cost per lead.
The numbers, with the close rate attached
| Meta / Instagram | Google search, non-branded | |
|---|---|---|
| Cost per lead | ~$45 ($25–$200 range) | ~$149 |
| Close rate | 10–20% | 15–30% |
| Cost per booked job | ~$300 | ~$662 |
| Intent at the moment of the ad | None | High |
| Can reach people pre-failure | Yes | No |
| Can catch an emergency | Rarely | Yes |
Meta genuinely does produce a cheaper booked job on these blended figures, and that is the honest case for it. What it cannot do is replace search, because the two are not competing for the same moment. One reaches a homeowner in March whose fifteen-year-old system will fail in July. The other reaches them in July, at 2pm, sweating.
Cost by offer type, which matters far more than the blended number:
| Offer | Typical CPL |
|---|---|
| Tune-up form fill | $35–$65 |
| Maintenance plan enquiry | $40–$80 |
| System replacement estimate | $80–$150 |
| High-ticket installation campaign | $115–$117 |
The pattern is consistent with every other channel in this trade: the higher the decision threshold, the more the lead costs and the more of them evaporate. We laid the full cross-channel picture out in what HVAC leads actually cost.
The instant form trap
This is the single most expensive mistake in HVAC Meta advertising, and it is expensive precisely because it looks like a win on the dashboard.
Meta's instant lead forms auto-populate name, email and phone from the user's profile. Friction drops by an estimated 70% to 80%. Cost per lead falls 30% to 60%. Every metric you are likely to be reporting on improves at once.
Then look at what happens next. Lead-to-appointment runs 15% to 30% on instant forms against 30% to 50% on landing page leads. Reports across US advertisers put website conversion leads closing at two to three times the rate of instant form submissions.
Watch out
Halve the cost of a lead and halve the rate at which it becomes an appointment, and you have not saved anything. You have doubled your lead volume, doubled your follow-up workload, and landed in the same place on cost per booked job, or slightly behind it.
The reason is mechanical rather than mysterious. The friction you removed was doing a job: it was the smallest possible test of whether someone actually wanted to talk to a contractor. Two taps on a pre-filled form is not a decision. Typing a phone number into a landing page is.
When instant forms are still right: you have genuinely instant follow-up, you are selling something low-threshold like a tune-up where the lead only has to show up, and you have added a qualifying question or two to the form to restore a little friction on purpose.
When they are wrong: anything approaching a system replacement, and any business whose median first-contact time is measured in hours. Which, in this trade, is most of them, since only 11% of HVAC businesses respond to a new lead within an hour. That constraint decides more campaigns than any bidding setting, and it is the subject of why your HVAC leads are not converting.
What Meta is genuinely good at in HVAC
Four things, all of which exploit the fact that you are reaching people before they need you.
Pre-season maintenance and tune-ups. This is Meta's structural advantage and search's structural blind spot. Nobody searches for a tune-up in March. They can absolutely be sold one in March, and every tune-up sold in March is a replacement conversation you are having instead of a competitor in July. Front-load spend before the season rather than into it, which is the timing argument in the best time of year to advertise HVAC.
Maintenance plans. Recurring revenue, low decision threshold, and the single best defence against a bad quarter. A homeowner on a plan is not a lead you have to buy again.
Indoor air quality. The highest-margin service line most contractors under-sell, and one with almost no organic search demand because homeowners do not know to look for it. That combination is precisely what a demand-creation channel is for.
Financing-led replacement offers. Not the replacement itself, the financing. "Monthly payment" reframes a five-figure decision into a considerable smaller one, and it converts from a cold feed in a way that a bare replacement offer does not.
What Meta is bad at: emergency repair. By the time the furnace dies, they are on Google, and they will not scroll back to find your ad from Tuesday.
We build the landing pages and the follow-up that make paid social pay, because the ad is the cheap part. If your leads are arriving and going cold, more budget makes the problem larger rather than smaller, and we will say so before taking any.
Creative decides the price, not targeting
The $25 to $200 spread in HVAC cost per lead is not mostly a targeting story. It is creative age and season.
At the good end sits a tune-up form fill, in peak season, with a creator-led video still inside its effective window. At the bad end sits a tap-to-call estimate request, in shoulder season, on a creative that has been running too long. Same account, same audience settings, eight times the cost.
Practical consequences:
- Refresh creative every four to six weeks in-season. Rising CPL with flat targeting is fatigue, and no bid adjustment fixes it.
- Run video from real technicians. A tech explaining what a failing capacitor sounds like outperforms a stock photo of a smiling family every time, because it is the only thing in the feed that looks like it came from a real company.
- Show the work. Before and after, a rooftop unit, a duct cleaning. Home services is one of the few categories where the actual job is inherently watchable.
- Lead with the problem, not the company. "Your AC will fail in the first heatwave, here is how to tell" beats any version of "we have served the region since 1994."
- Keep the offer price-free. A free or fixed-scope check gives someone a reason to act without starting a price comparison you did not want.
A campaign structure that survives the season
Most HVAC Meta accounts are one campaign, one audience, one offer, running all year. That structure guarantees the account looks broken twice a year, because a single tune-up offer cannot make sense in both April and January.
A structure that holds up:
Campaign 1: pre-season maintenance. Runs the six weeks before each season turns. Tune-up or plan offer, broad local targeting, video creative. This is the campaign that earns its keep, and it is off during peak season because you do not need to buy demand you already have.
Campaign 2: maintenance plan. Runs continuously at a modest budget, retargeting anyone who has interacted with you or visited the site. Low volume, excellent economics, and it compounds because every sign-up is a customer you never repurchase.
Campaign 3: replacement, financing-led. Runs during peak and shoulder season, targeting older housing stock. Higher cost per lead, much higher job value, and the only one of the three where slow follow-up genuinely ruins the unit economics.
Campaign 4: retargeting. Anyone who visited a service page or started a quote and did not finish. Cheapest leads in the account by a distance, and the closest thing Meta has to search intent, because these people already declared themselves.
That fourth campaign is the one most contractors are missing entirely, and it is usually the first thing worth turning on. If your site gets traffic from search, from your Google Business Profile, or from the other three campaigns, there is already an audience of people who considered you and left. Reaching them again costs a fraction of reaching a stranger.
The offer matters more than the audience
One reframe worth sitting with, because it inverts how most HVAC ad budgets get debugged.
On search, targeting is most of the work. Someone typed "furnace repair," and the job is to be there and be credible. The offer barely has to exist.
On Meta, the audience is largely a solved problem: local homeowners is not a hard target, and Meta's delivery system finds the responsive ones without much help. What varies by an order of magnitude is whether the thing you are offering is worth interrupting someone for.
So when a Meta campaign underperforms, the debugging order is almost the opposite of search:
- Is the offer worth stopping for? A tune-up at a fixed scope, yes. "Contact us for a quote," no.
- Is the creative fatigued? Check how long it has been running before touching anything else.
- Is the landing experience consistent with the ad? An ad promising a tune-up that lands on a generic homepage loses most of the traffic it paid for.
- Is follow-up fast enough to survive a colder lead than search produces?
- Only then, targeting.
Contractors habitually start at step five, because targeting is the part of the interface that looks like a lever. It is the least sensitive input in the account.
The measurement that keeps you honest
Meta's reporting will show you cost per lead in large numbers and will not show you what happened afterwards. Track four things yourself, weekly:
- Cost per lead, by campaign and by offer.
- Lead-to-appointment rate. This is where instant forms fail, and it is invisible inside Meta.
- Appointment-to-sale rate.
- Cost per booked job, which is the only figure that decides budget.
If you cannot connect a Meta lead to a booked job, you are optimising toward whatever produces the cheapest form fill, which on this platform is reliably the worst-performing thing you can buy. That connection is what a CRM is for, and it is the difference between a channel you can scale and one you are guessing at.
One last piece of context worth holding onto. Facebook is used by 32% of home service businesses, making it the most common channel after referrals and repeat customers, and top performers run three to five sources rather than betting on any one. Meta earns a place in that mix as the pre-season and plan-selling channel. It does not earn the whole budget, and the contractors who hand it the whole budget are usually the ones who conclude a year later that social does not work in this trade.
It works. It just does not do the job they hired it for. The full comparison, with cost and close rate for every channel, is in the best lead sources for HVAC companies, and the targeting mechanics get their own treatment in HVAC Facebook ads targeting homeowners.