It’s a hot week in July, the phone won’t stop ringing, and you’re still not sure the Google Ads account is working. Half the calls are job seekers, price shoppers, or people who hung up on voicemail at 9 p.m. The invoice says you got leads. The schedule says otherwise.
This guide walks you through a first pass at how to improve Google Ads for HVAC without raising your budget. You’ll fix tracking, split emergency repair from replacement demand, cut wasted spend, and set bids around cost per booked job. You need admin access to your Google Ads account, your call and booking data, and about two hours.
The order matters. Most HVAC waste comes from measurement and structure, not from bids or ad copy, so we start there.
Step 1: Fix call and booking tracking before you touch anything else
Between 40% and 70% of local service leads come by phone. If your account only counts form fills, it’s blind to most of your results, and Smart Bidding is learning from the wrong signal. It will happily find you more people who fill out forms, whether or not they ever book a truck roll.
Work through these in order:
- Add call assets. Use a Google forwarding number on your ads so calls from ads are recorded and measurable.
- Add website call tracking. Install the Google forwarding number swap on your site (or a third-party call tracking tool) so calls from people who click, browse, then dial are attributed to the right campaign.
- Set a minimum call length. Something like 60 seconds is a reasonable starting point, so hang-ups and wrong numbers don’t register as conversions. Adjust it if your intake calls are typically shorter.
- Import booked jobs as offline conversions. Pull them from ServiceTitan, Housecall Pro, Jobber, or even a spreadsheet you update daily. Once the account sees booked jobs, it can optimize toward them instead of toward clicks.
The most common mistake is counting every “Calls from ads” event, every page view, and every button click as a conversion at equal value. Separate them. Booked jobs and qualified calls should be primary conversions that bidding optimizes toward. Page views, scroll depth, and short calls belong as secondary, for observation only.
Then verify. Open your conversion actions list in Google Ads and confirm which ones are marked primary. The interface changes often, so as of 2026 the labels and settings names may differ from what you’ve seen before. Place a test call from your phone and confirm it shows up. If the booked-job import isn’t showing recent data after a few days, check that the click ID is being captured and passed into your CRM.
Step 2: Separate emergency repair campaigns from replacement and install campaigns
These are different buyers. Someone searching “AC not cooling” or “furnace out” wants a truck today and will call the first company that picks up. Someone searching “new AC installation cost” is researching for days, comparing quotes, and probably filling out forms at three companies.
Build at least two campaigns so each gets its own budget, bidding target, and ad copy. A repair call and a system replacement are not worth the same, so they shouldn’t share a CPA goal. If you’re working from ticket-size ranges, use your own numbers from your CRM. Published averages vary widely by region and equipment, so cite a named source with a year (Angi or HomeAdvisor, for example) or skip the figure.
A practical structure:
- Emergency and repair: urgent, high-intent terms, tight geo, aggressive bids during hours you answer.
- Replacement and install: cost, quote, and new-system terms, with financing and free-estimate messaging and a longer evaluation window.
- Maintenance and tune-up: lower-cost, steady demand that helps fill the shoulder seasons.
Seasonality is the reason the third campaign earns its keep. Cooling demand spikes in summer heat waves and heating demand in cold snaps. Between those, repair volume drops, and a maintenance campaign keeps techs busy and builds a customer list you can sell replacements to later.
The mistake to avoid is one campaign called “HVAC” with every keyword in it. The algorithm will spend wherever conversions come cheapest, and that’s usually low-value jobs. You’ll see a decent CPL on paper while the big-ticket replacement leads get starved.
Step 3: Rebuild your keyword list around urgency and cut the junk with negatives
Start with the search terms report. Pull the last 30 to 90 days, sort by spend, and filter for terms with zero conversions. That list is your education. You’ll find searches you never meant to buy.
For match types, keep it practical. Use phrase and exact match for emergency terms like “ac repair near me” or “furnace not heating.” Use broad match only if you have strong conversion data feeding the account and a tight negative list behind it. Broad match plus Smart Bidding without negatives is how budgets disappear into irrelevant searches.
HVAC has predictable traps. People search for HVAC jobs, certification courses, and unit parts, not service. Here’s a starter negative list you can copy by category and adjust:
- Jobs: jobs, job, careers, hiring, salary, apprenticeship
- Education: training, school, course, certification, class, license exam
- DIY: diy, how to, manual, instructions, tutorial
- Parts and resale: parts, wholesale, used, surplus, supply, for sale
- Brands you don’t sell or service: add manufacturer names accordingly
Check each term against your business before adding it. If you do sell parts or training, obviously skip those.
Apply negatives at two levels. Put shared junk (jobs, school, salary) at the account level. Put intent-specific exclusions at the campaign level. For example, exclude “installation” and “replacement” from the repair campaign so those searches land where the right ad and landing page are waiting. Do the reverse in the install campaign for “repair” and “emergency.”
Step 4: Tighten geo-targeting and ad schedule around your real service area and call-answering hours
Open your campaign location settings and look for the option about who sees your ads. You want the setting for people in, or who regularly visit, your targeted locations, not the broader “presence or interest” option that includes people merely interested in your city. Wording changes, so confirm the current labels in the interface. Leaving it broad means paying for clicks from people researching a move or a vacation rental.
Target by zip code or radius based on what’s profitable to drive to, not by county lines. A zip 40 minutes away with traffic may cost you more in windshield time than the job earns. Exclude zips you won’t dispatch to, and review which areas produced booked jobs, not just calls.
Then look at timing. Pull your call logs and sort by hour and day. You’re looking for two things: when calls actually come in, and when they go unanswered. A missed emergency call is the most expensive click in the account, because you paid for it and got nothing.
If nobody answers after hours, you have two options. Route calls to an answering service that can book or dispatch, or reduce bids (or pause ads) during those windows. Don’t keep paying full price for a phone that rings out.
Give these changes room. Wait two to four weeks of data before judging them, and expect full optimization to take 30 to 90 days. Tweaking geography every few days just makes the results unreadable.
Step 5: Write ads and landing pages that match the call a homeowner is about to make
An emergency searcher scans for three things: can you come today, are you legit, and what will it cost. Answer those in the ad. Useful claims include same-day service, licensed and insured, financing availability, upfront pricing, and the names of the towns you serve. Use only what you can back up. If you can’t actually get there same day in July, don’t say it.
Fill out your assets too. Call, location, callout, and structured snippet assets give you more space and more reasons to click. Asset types and availability change, so check what’s offered in your account as of 2026.
The landing page matters as much as the ad. Build one page per campaign type:
- A tap-to-call phone number at the very top on mobile
- A short form for people who don’t want to call
- Trust signals up front: license number, review count and rating, and brand certifications such as NATE if you actually hold them
- Copy that matches the ad, so an “AC not cooling” click lands on a page about AC repair
The classic mistake is sending emergency clicks to the homepage or a long “About Us” style page. Someone with a dead AC at 95 degrees will not hunt for your phone number.
Finally, test mobile load time with Google PageSpeed Insights. Fix the worst offenders, usually oversized images and heavy scripts. A slow page wastes clicks you’ve already paid for.
Step 6: Set bids and budgets using your real cost per booked job
The $18 to $35 benchmark for home services Google Ads is cost per lead, not per booked job. To find your real number, take your cost per lead and divide by your lead-to-booked rate. Suppose, for illustration, that you pay $30 per lead and 1 in 3 leads books. Your cost per booked job is about $90. Use your own booking and close rates, and track them by campaign.
For overall budget, plan to spend roughly 8% to 12% of revenue on marketing, with Google Ads as one part of that. If your account is far below or above that range, ask why before moving money around.
On bidding, build up gradually:
- Start with Maximize Conversions or manual CPC while the account collects data.
- Move to Target CPA only once there are enough recent conversions. Google’s guidance on thresholds changes, so check the current recommendation before switching.
- Set the target based on cost per booked job for that campaign, not a blended number.
Protect the emergency campaign first. During peak weeks, check daily that it isn’t running out of budget by noon. Pull money from maintenance or install if you have to, because the best-converting hours are the ones you can’t afford to miss.
Two mistakes reset your progress. One is judging a campaign after a week. The other is changing bidding strategies every few days, which restarts learning each time. Make one change, then wait.
Step 7: Run a weekly 20-minute review and know when to ask for help
Once the account is rebuilt, maintenance is short. Each week, check:
- The search terms report, adding new negatives
- Call recordings or logs, listening to a handful
- Missed-call count, by hour
- Cost per booked job by campaign
- Impression share lost to budget, especially on emergency terms
Good looks like leads inside the $18 to $35 range with a documented booking rate. A low CPL with a poor booking rate is just cheap noise.
If an agency or previous manager built your account, look for red flags: no call tracking, a single campaign, Performance Max with no exclusions or reporting visibility, a lock-in contract, and reports that only show clicks and impressions. Any one of those means you can’t see what your money is doing.
Also consider Local Services Ads as a parallel channel. Availability, eligibility rules, and pricing vary by market and change over time, so check what’s offered for HVAC in your area as of 2026 before planning around it.
If you’d like a second set of eyes, Clicks Geek runs HVAC accounts month-to-month with no lock-in and can audit an existing account. Our HVAC marketing page shows how we approach the vertical.
Step 8: Recheck after 30 days and keep only what moved booked jobs
After 30 days of changes, go back to the two numbers that matter: tracking accuracy and cost per booked job. Confirm calls and imported jobs are still flowing, then compare each campaign against where it started. Keep what produced booked jobs, drop what didn’t, and resist the urge to keep tinkering with things that are working.
Revisit bidding and budget once there’s real conversion data behind them. That’s when Target CPA becomes realistic and when shifting spend between repair, install, and maintenance is a decision, not a guess.
Tired of spending money on marketing that doesn’t produce real revenue? We build lead systems that turn traffic into qualified leads and measurable sales growth. If you want to see what this would look like for your business, we’ll walk you through how it works and break down what’s realistic in your market.
