A July condenser call and a January no-heat call cannot share one campaign
The account has two seasons with different vocabulary, different urgency and different job values, and they should never draw from the same budget.
Cooling season terms and heating season terms attract different people at different hours. A furnace that quit at 6am on a Tuesday in January is a different buyer from someone shopping a condenser replacement in July, and the ad copy that works for one is noise to the other.
Run them as separate campaigns with separate budgets, separate landing pages and separate schedules, so you can throttle one without touching the other. When the first hard freeze arrives you want a dial, not a rebuild.
Use the shoulder months deliberately. April and October are when maintenance agreements and planned replacements are sellable, and when clicks are competing against less panic. Build a campaign for that work instead of letting the emergency campaign absorb the budget year round.
One combined campaign hands the bidding system a mixed signal and it will optimize toward whatever converts cheapest, which is almost never the work you want more of.
Franklin and Antioch should not share a bid, and only your own numbers say by how much
Davidson County runs as a consolidated government while Williamson, Rutherford, Sumner and Wilson do not, and the buyer changes as you cross the line.
Draw the geography on counties and honest drive times rather than a radius around your shop. A circle puts money into places you will not send a truck at five o'clock on a weekday and cuts off places you serve happily.
Franklin and Brentwood in Williamson County are a different buyer from East Nashville, and Murfreesboro is different again. Job mix, average system count per house and willingness to wait all vary. So does what a click costs.
Set bid adjustments from your own booked job data, not from a benchmark someone published. Until you have enough of your own results by county, run them level and watch. Guessing at adjustments early is how accounts get quietly starved in the areas that were working.
Exclude what you cannot serve. If Sumner County is a two hour round trip and you only go for replacements, say so in the targeting and in the ad copy, not in an apology on the phone.
Trade school and hiring searches quietly eat a heating season budget
HVAC attracts a large volume of searches from people who will never buy a repair, and paid accounts pay for all of them by default.
Job seekers, apprenticeship and trade school queries, salary questions, parts lookups, model number research, manufacturer warranty checks and step by step reset instructions all sit inside broad HVAC terms. None of them books a truck.
Renters are the other quiet drain. Somebody in an apartment cannot authorize a repair, and in much of Davidson County that traffic is substantial. Ad copy that names homeowners and property managers filters some of it before the click.
Broad match and automated bidding will keep inventing new queries, so the search terms report needs reading weekly during peak season, not quarterly. Exclusion lists are local and they are never finished.
Every dollar removed from a query that cannot convert goes back into the auction for the ones that can. In a short heating season, that reallocation matters more than any change to your ad copy.
The week of an ice storm, turn spend down rather than up
When freezing weather generates more calls than you have technicians, buying additional clicks buys complaints.
A hard freeze or an ice storm produces demand you did not pay for. The phone rings on its own. Paying to add calls you cannot answer means booking dates you will miss, and the reviews that follow outlast the storm by a long way.
Pull back during the surge, then put the money into the two weeks after it. That is when people who limped through on emergency heat start pricing a replacement, and when a booked appointment is actually deliverable.
Pace against dispatch capacity rather than evenly across the month. If the board is full through Thursday, the ads should reflect Thursday, not Monday's optimism.
Daypart honestly. If nobody picks up at two in the morning and the answering service can only take a message, either accept a lower bid in those hours or make sure the message gets returned first thing. Paying peak rates for a ring into an empty office is the most avoidable waste in the account.
The Gulch rental manager is buying a response window, not a price
A meaningful share of housing downtown and nearby is run as a business, and the person making the call is buying turnaround, remote approval and clean invoicing.
Short-term rental operators and management companies around Broadway, The Gulch, Germantown and East Nashville often hold several properties. One relationship can be worth many single visits, and they judge you on how fast you can get in and out with a guest arriving.
Give them their own campaign, their own ad copy and their own landing page covering access arrangements, after-hours response, invoicing terms and who signs off on what. Sending them to the homeowner page wastes the click.
Search alone will not reach all of them, because plenty of this work moves through management companies rather than through Google. Treat paid as one channel into that segment and keep the expectations modest.
Rules covering short-term rental properties are contested and change, so do not state them in ad copy or on the page. Say practice varies and is worth confirming with Metro.
Count the service agreements a paid call produces, not the calls
A call total is the easiest number to move and the least useful one to run a business on.
Connect call tracking to whatever your office uses to dispatch, so an ad click can be followed through to a booked appointment, an invoice and a signed maintenance agreement. Without that link every optimization decision is a guess dressed up in a chart.
Three numbers deserve to be on the report: cost per booked job, cost per installed system, and how often a paid call ends with an agreement. Cost per lead belongs nowhere near the last page.
Agreements are the reason a January panic call is worth more than its invoice. A customer on a plan is a customer who does not search again next winter, which changes what you can afford to pay for the first call.
No account should promise a volume of jobs or a cost per job in advance. What the reporting should show is the direction of those three numbers, split by county, over a period long enough to include a full season.
Questions we actually get
- What should we budget to start?
- Enough to cover one clearly defined segment properly rather than a little across everything. A campaign spread thin over five counties and both seasons produces data too sparse to act on. Pick the county and the job type you most want more of, fund it to the point where you see meaningful click volume, and expand once the booked job numbers are stable enough to trust.
- Are Local Services Ads worth it alongside Google Ads?
- For most HVAC contractors they are worth testing, because they sit above the standard results and are priced on leads rather than clicks. They also demand paperwork and review volume, and they cannot be tuned the way a search campaign can. Run them alongside rather than instead of search, and hold them to the same standard: booked jobs, not lead counts.
- Should we bid on competitor names?
- Often yes in this market, because so many households arrived recently and are choosing between names they have only just heard. It is legal to bid on a competitor's name and generally not permitted to use their name in your ad text. Keep the copy about what you offer, watch the conversion rate closely, and drop it if the calls are people looking for their existing contractor.
- How quickly can we expect results?
- Clicks are immediate; useful conclusions are not. Booked job data accumulates at the rate your phone rings, and this trade adds weather noise on top. Any read on cost per booked job from a single week is unreliable, particularly a week containing a freeze. Plan on judging structural changes across a season.
- Do we need a separate landing page, or can ads point at the website?
- Separate, when the traffic is separate. An emergency no-heat ad and a property manager ad want different pages, different forms and different proof. If the site's existing service page already answers coverage, hours and what happens next, it can carry the emergency traffic. The manager audience nearly always needs a page the site does not have yet.
