Three campaigns: tonight, this month, and the system they will replace next spring
Three different buyers hide in the same keyword set, and they are not worth the same money.
Somebody with no cooling today. Somebody whose system keeps failing who is finally collecting quotes. Somebody shopping a maintenance agreement or an annual tune up. One campaign averages all three into a bid that is wrong for each.
Split them, because the click value, the ad copy and the landing page all differ. Emergency intent belongs on a page with a phone number and an honest answer about today. Replacement intent belongs on a page explaining how a quote is built and what financing looks like.
The maintenance campaign is the one most shops skip and the one that compounds. Agreements sold in the shoulder months become the customer list you sell changeouts from later, at a fraction of what the July auction charges for the same household.
Weight call assets toward the emergency campaign and let the replacement campaign carry forms. A homeowner comparing systems wants to read before they talk to anybody.
A radius drawn from a South Austin shop pays for the wrong side of MoPac at five o'clock
Distance from the shop is a poor proxy for whether a truck can get there, finish, and get back.
Crossing MoPac or I-35 in the afternoon can cost more time than ten extra miles of open road at nine in the morning. A circle on a map prices none of that.
Build the geography as separate targets that reflect drive time, then bid them independently. Travis County near the shop. The Williamson County corridor through Round Rock, Cedar Park and Leander. The Hays County run down to Kyle and Buda. The Hill Country push west toward Lakeway, Bee Cave and Dripping Springs.
West deserves its own decision. Slope, long private driveways and difficult access add time to a visit that mileage does not reveal. Some shops make good money out there. Others should stop bidding it after mid afternoon.
Check location of presence against location of interest. Someone searching downtown about a property in Georgetown is a completely different job from a downtown address, and default settings will happily blur the two.
Rebate research, filter sizes and job seekers will eat a July budget quietly
The exclusion list is where an HVAC account earns its keep, and it needs maintenance rather than a one time setup.
Start with the usual offenders. Filter sizes, capacitor part numbers, thermostat wiring diagrams, window units and portable units, and anything containing the word manual. All of it clicks. None of it invoices.
Two local ones matter more than people expect. Utility program research pulls in readers who want to know what a program covers and are months from buying, if ever. Hiring searches pull in technicians, and in a labor market this tight they click a great deal.
Add warranty and builder terms if you do not take that work, tenant complaint language if you do not want apartment calls, and competitor brands you cannot honestly out-argue on the landing page.
Read the search terms report weekly in season and monthly outside it. An account that has not had a negative added in three months is not being managed.
Promising a utility program to a Pflugerville address starts an argument at the kitchen table
Eligibility generally depends on which utility serves the specific address, and a wide geographic target reaches plenty of addresses it does not describe.
Austin Energy is owned by the city, which already separates this market from most of Texas, where a household picks a retail provider. Its territory does not line up with everything a county level target will reach.
So an ad promising a city program that shows to somebody out in Williamson or Hays County is setting up a conversation nobody in the office wants to have. The technician on site ends up defending marketing copy.
A safer construction: say your team confirms what applies to a given address before anything is quoted. It is true, it differentiates you from generic statewide advertisers, and it does not need editing every time terms change.
Keep dollar figures out of ad copy for anything you do not control. Program terms change, and it is worth confirming current terms with the utility serving the address every time.
In the worst week of the heat your constraint is trucks, so stop buying calls you cannot run
Demand and capacity peak in the same week, which is exactly when most accounts are told to spend harder.
Buying more emergency clicks while dispatch is booked out produces voicemail, arrival windows you cannot honor, and reviews written by people who waited. Paying for all three at once is a poor trade.
Pace against the board rather than the calendar. When capacity is gone, cut the emergency campaign before the replacement campaign, because a replacement shopper will still be there tomorrow and a no-cool caller will not.
Shoulder months are the bargain. Spring and fall are when replacement research happens, the auction is calmer, and a quote can be a conversation instead of a rescue.
Even twelfths of an annual budget describe nobody's year in Central Texas. Set spend by season and by campaign, and put the money where the trucks are actually free.
Price the account on the truck rolls that produced an invoice, county by county
Cost per lead flatters everyone, because a hang up, an out of area form and a booked changeout all count as one.
Feed booked jobs back into the platform. Importing offline conversions from dispatch or the CRM lets bidding optimize toward work that invoiced rather than toward whatever produced the cheapest clicks.
Then read cost per booked job by county and by campaign. The Travis County number near the shop and the Hays County number will diverge, and the size of that gap is the argument for how next season gets funded.
Two supporting numbers repay the effort. Booking rate on paid calls, which usually says more about the phone than the ads. And maintenance agreements sold from paid traffic, which is the part that quietly pays for next year.
No one can promise a lead volume or a cost per job in advance. What is reasonable to expect is a structure you can audit, a search terms report someone actually reads, and reporting that ends at revenue instead of impressions.
Questions we actually get
- How much should an Austin HVAC company spend on Google Ads?
- There is no honest benchmark to quote, and any number given before someone has seen your booking rate and your average invoice is invented. Work backward instead: what a booked job is worth, what share of calls your phone books, and how many jobs your trucks can absorb in a week. Those three set the ceiling, and the first month of real data replaces the estimate.
- Do Local Services Ads make sense alongside a search campaign?
- For most residential HVAC shops they are worth running in parallel rather than instead. They reward the license and insurance paperwork you already keep, they sit above the standard results, and they price by lead. They also give you far less control over geography and intent, so they rarely replace a properly structured search account across a service area spanning three counties.
- Should we bid on our own company name?
- Usually yes, at a small budget. Competitors bid on it, and the click is generally cheap because your own page is highly relevant. Watch the reporting though: brand clicks convert well and will flatter the whole account average, so report brand separately from everything else or you will overstate what paid search is really producing.
- How do we handle Williamson and Hays counties without wasting money?
- Run them as their own campaigns with their own budgets and their own bids, not as a wider radius. Drive time to Round Rock in the morning and drive time to Kyle in the afternoon are different businesses. Separate reporting also tells you honestly whether an outlying county is profitable or just busy.
- Can we mention utility rebates in our ads?
- Better not to name amounts or programs. Eligibility generally depends on which utility serves the specific address, and terms change without warning. Copy that says your team confirms what applies before quoting is accurate, holds up when the customer is out in Pflugerville rather than inside the city, and never needs rewriting.
