What HVAC Leads Actually Cost in 2026 (And How to Pay Less)

Here’s the mistake that costs HVAC owners the most money, and it isn’t overpaying for clicks. It’s treating every lead as the same unit. A referral from a past customer, a Google Maps call from someone who already read your reviews, a cold ad click, and a shared-marketplace quote request all get counted as “one lead” — but they don’t carry the same intent, the same trust, or anywhere near the same chance of becoming a paying job.

That’s why a $75 lead can be your most expensive customer. Let me show you the math, the real 2026 benchmarks with sources, and the napkin formula that tells you what your shop — not the industry — should pay for a lead.

Cost per lead is the wrong number (here’s the right one)

Cost per lead (CPL) tells you one thing: what it cost to make the phone ring. It says nothing about missed calls, duplicates, wrong-service-area calls, tenants who can’t authorize work, price shoppers, no-shows, or a technician who doesn’t close. The numbers that actually matter come in sequence:

• Cost per booked appointment = CPL ÷ booking rate
• Cost per sold job = CPL ÷ lead-to-sold rate
• Gross profit after acquisition cost — the only one the bank cares about

Run it once and the illusion breaks: a $150 lead with a 40% booking rate costs $375 per booked appointment. A “cheap” $75 shared lead that books 10% of the time costs $750 per sold job. The cheap lead can be the expensive customer.

The best public data makes the same point at scale. SearchLight Digital’s January 2026 benchmark — 816 HVAC and plumbing contractors, 8,077 campaigns, $14.88 million in tracked Google Ads spend — found non-brand search averaging $149 per lead but $804 per paying customer, while blended Google Ads ran $104 per lead and $472 per customer. The gap between those two columns is the part almost nobody measures:

One honest note about every benchmark in this article: numbers like these come from observed ad accounts and platform-reported conversions, each platform defines a “lead” differently, and your market is not the average market. Use them for calibration — then trust your own customer records over any of it.

What leads cost in 2026, source by source

The reference points, each cited: Coast333’s benchmarks put blended Google Ads around $104 per lead, non-brand at $149, branded at $34, Performance Max at $72, Local Services Ads (LSAs) at $51, and allocated search engine optimization (SEO)/organic leads at $69 — with urgent-service campaigns in competitive metros exceeding $200 per lead. SearchLight’s February 2026 LSA benchmark showed HVAC LSAs at $51 per lead with a 44% book rate and a $2,110 average ticket. And the reason owners tolerate any of these prices: the jobs are worth it. Angi’s 2026 cost guide puts the average HVAC replacement around $7,500, with a $5,000–$22,000 range.

The lead-quality ladder

Price is only half the equation. Here’s how lead quality ranks, in my experience — and why:

1. Repeat customers and referrals. Someone who already trusts you, or heard from someone who does. They ask “can you solve this?” — not “are you the cheapest?” Always the best leads, and the reason review and referral systems are worth more than any ad budget.
2. Organic search and Google Maps. The homeowner chose you — after seeing proximity, reviews, photos, your brand. They arrive partly sold.
3. Branded search. It shows up in your ads report, but it behaves like owned demand — the person searched your name. That’s why it books at 55.3% versus 37.6% for non-brand in the SearchLight data.
4. Local Services Ads. High intent, often efficient — but the homeowner is choosing from Google’s comparison set, not seeking you out.
5. Non-brand Google Ads. Real, profitable urgent demand — but you’re buying your way into consideration against everyone else doing the same.
6. Shared marketplaces. More competition, weaker brand attachment, more price sensitivity.

The pattern that tells you everything: a Maps caller says “I saw your reviews — when can you come?” A marketplace lead says “I’m getting three quotes — what’s your diagnostic fee?” Both count as “one lead” in a report. They are not the same commercial object.

Shared marketplaces: the honest version

Shared lead platforms aren’t automatically worthless. They make sense in a narrow set of situations: a brand-new shop with no local visibility yet, idle capacity that needs filling this week, testing a new service area, or a shop with a very fast response process and disciplined spending caps. Treat them as spot-market demand — never as the foundation of the business.

The trap is structural: you’re often paying for a conversation, not a customer. When three to five contractors chase the same homeowner, the platform has built a reverse auction — margins compress, estimate-chasing rises, and the customer remembers the platform, not you. Worth knowing as context: the Federal Trade Commission (FTC) ordered HomeAdvisor (affiliated with Angi) to pay up to $7.2 million over misleading claims to service providers about lead quality. That doesn’t make every marketplace lead bad — but it’s a strong argument for demanding transparent lead definitions, refund rules, and your own conversion tracking before you spend. (Thumbtack’s own docs note lead prices vary by job type, market, and how many pros are competing — which is exactly why national averages can’t price your ZIP code.)

Receipts vs. assets

Here’s how I frame the rent-versus-own decision without moralizing, because paid isn’t bad and rented isn’t bad — the question is what the spend leaves behind:

Some marketing spend creates a receipt. Some creates an asset. Paid leads are useful when you need demand now. But when you stop funding the auction or the marketplace, the flow stops — Google’s own LSA documentation is plain about it: you’re charged per valid lead, prices vary, and when the budget caps, the ad disappears. Visibility work is different when it builds things the company controls: reviews, Google Business Profile strength, service pages, local authority, email and text lists, referral systems, call-handling data, and a brand people search by name.

And to be honest about my own side of the fence — plenty of “visibility work” doesn’t compound either: generic blog posts that never rank, thin city pages, landing pages rented on a vendor’s domain, marketplace reviews that don’t transfer to your brand. Even your ad spend only compounds if you own the account, the conversion data, and the call recordings. (That ownership question has its own article: the contract red-flags checklist.)

When buying leads is the right call

Being fair to paid — it’s the correct move whenever you have a timing or capacity problem and can convert profitably: a new shop that needs call volume before referrals and Maps visibility exist · open install capacity and a known replacement close rate · attacking emergency repair during peak weather · a slow week that needs controllable demand · entering a new service area and wanting data before committing to long-term visibility work · defending your branded searches from competitors bidding on your name. LSAs in particular deserve a fair shake — $51 leads booking at 44% can be very efficient if the office answers fast and the profile is clean.

The napkin math for your shop

Before buying any lead source, you need your own numbers: booking rate, show rate, close rate, average ticket, gross margin, and capacity. Then:

• Gross profit per job = average ticket × gross margin
• Lead-to-sold rate = booking rate × show rate × close rate
• Max CPL you can pay = allowable acquisition cost × lead-to-sold rate

Two worked examples. A repair shop: $900 average ticket at 50% margin is $450 gross profit; if you’ll spend $150 to acquire a job and 30% of leads become sold jobs, your max CPL is $45. A replacement-focused shop: $10,000 ticket at 35% margin is $3,500 gross profit; $700 allowable acquisition at a 20% lead-to-sold rate means a max CPL of $140. Same industry, same year — one shop profitably pays $150 for a replacement lead while another loses money on a $60 repair lead. There is no universal “good” cost per lead. It’s a function of your margin, your close rate, and your capacity.

For calibration: an ACHR News–covered 2026 contractor survey found contractors averaging $422 per service ticket and closing 43% of install jobs (45% residential, 38% commercial). Useful reference points — but your customer relationship management (CRM) records beat every benchmark in this article.

Peak season vs. shoulder season

In peak cooling or heating demand, the goal is not the lowest CPL — it’s the highest-profit use of scarce capacity. Tighten campaigns to high-intent services and profitable ZIP codes, protect your branded search, raise budgets only where crews can actually take the work, pause low-margin tune-up offers that block repair and replacement capacity, watch missed calls like a hawk, match ad schedules to when someone actually answers the phone — and cut the shared marketplaces entirely if your board is already full.

In the shoulder months, move the money toward demand creation and asset building: maintenance agreements, reactivation campaigns, review generation, referral pushes, your email and text lists, Google Business Profile work, service pages, financing offers, replacement planning. Paid can still run — but change the offer. Chasing emergency keywords in months with no emergencies is how shoulder-season budgets evaporate.

The bottom line

Know your gross profit per job. Know your lead-to-sold rate. Multiply your way to your own max CPL, and judge every lead source — Google, LSAs, marketplaces, even us — against it. Rent demand when the math and the moment justify it. But put steady money into the things that are still working the month after you stop paying: your reviews, your profile, your pages, your brand, and a website the search engines and AI assistants can actually read. (Here’s the complete guide to that last part.)