Circle your shop on a map and you will pay for open water
Radius targeting is a convenience that ignores the single most important fact about this metro, which is that a straight line between two addresses often crosses a bay.
Draw twenty miles around a location in South Tampa and a meaningful share of that circle is water, with much of the rest sitting behind a crossing. The map looks reasonable. The dispatch board disagrees at four in the afternoon.
Build geography from your own service reality instead. Target by county, by city and by ZIP code, using the list of places you actually send trucks. It takes an afternoon and it is the highest return hour in the whole account.
Set location options to people in your targeted locations rather than people interested in them. The default setting will happily serve your ads to someone in Ohio researching a move, and those clicks bill the same as real ones.
Local Services Ads use their own service area settings, separate from your search campaigns. Set them from the same honest list, then check both quarterly, because service areas drift as crews change.
Run Hillsborough and Pinellas as two campaigns with two budgets
Two sides of the water behave like two markets, and averaging them together hides which one is paying for the other.
Auction pressure, click cost, close rate and drive time all differ across the bay. Blended into one campaign, a strong Hillsborough performance can mask Pinellas clicks that never book, and you will keep funding them for months.
Separate campaigns also give you separate controls. Different daily budgets. Different schedules. Different bid strategies if the conversion volume supports it. You can turn one down on a Tuesday without touching the other.
Within each, keep intent separated the way you already think about it operationally. A no-cool emergency, a planned replacement quote and a maintenance plan sign-up are three different jobs with three different values, and they should not share a budget line.
Watch the conversion data requirement. Smart bidding needs volume to learn, and splitting an account too finely can starve every campaign of signal. If you are small, split by county first and by intent second, once the numbers support it.
Bid up the hours when a truck can still cross the Howard Frankland
Ad schedules should follow your ability to arrive, not the general shape of the workday.
A cross-bay inquiry at nine in the morning is a booking. The same inquiry at four thirty is a promise you will probably break, whether the route runs over the Howard Frankland, the Gandy or the Courtney Campbell.
Use bid adjustments to reflect that. Push harder on cross-bay geography during the hours you can genuinely make the trip, and pull back sharply when you cannot. Keep full bids on your home county all day.
Do the same with days. Weekend coverage is worth paying for only if someone answers and someone can go. Ads running into an unstaffed phone are the most expensive thing in a contractor account.
None of this works without call tracking that records the hour. Pull a month of calls, mark which converted to a completed job, and let the pattern set the schedule instead of your intuition.
Most wasted spend is parts lookups, DIY questions and window units
The negative keyword list is where an HVAC account is won or lost, and in most accounts nobody has opened it in a year.
Three categories account for the bulk of it. People hunting a part number or a capacitor. People trying to fix it themselves and searching how-to language. People shopping a window unit or a portable, who will never buy a service call.
Add the ones specific to a metro with a lot of newcomers and a lot of rentals. Tenant language, apartment maintenance requests, and employment searches, because HVAC jobs and HVAC schools generate volume that looks like demand and is not.
Mine search terms weekly at first, then monthly. Build shared negative lists so a term you kill stays dead across every campaign rather than only the one you were looking at.
Be careful about broad match with automated bidding. It can produce genuinely useful discovery and it can also wander into brand-adjacent nonsense. Run it only with a maintained negative list and a term report someone reads.
Newcomers have no contractor to call, which makes competitor terms cheap
A market with constant arrivals contains a lot of people with no incumbent relationship, and paid search reaches them at the moment they start looking.
Households transferring in around MacDill and elsewhere arrive with no dentist, no plumber and no HVAC company. They are not switching away from anyone. They are choosing for the first time, often under a short timeline.
Brand terms are usually the cheapest clicks in the account, and in this market they matter more than people expect, because a newcomer who hears your name once will search it rather than remember a URL. Cover your own name.
Competitor campaigns are legal on keywords and restricted in ad copy, since you generally cannot use another company's trademark in the text. Keep the copy about you, expect a lower conversion rate, and cap the spend.
Landing pages for this audience should be blunt about credentials and coverage. Someone new to the area has no way to sanity check you socially, so license number, physical address and clear service area do more work than a discount.
Price the account per booked job by county and watch the two diverge
Cost per lead is the number that hides the problem; cost per booked job is the number that finds it.
A lead in this trade can be a wrong number, a parts question, a tenant with no authority or a real replacement inquiry. Averaging them gives you a figure that moves for reasons unrelated to your business.
Pipe the outcome back. Import booked and completed jobs from your field service software as offline conversions, so the bidding optimizes toward work rather than toward form fills. Even a weekly manual upload beats nothing.
Then split the number by county. When Hillsborough books at one cost and Pinellas at three times that, you have a budget decision rather than a mystery, and you can make it in an afternoon.
Pace against capacity, not evenly across the month. A campaign that spends its budget while your board is full is buying calls you will handle badly. Turn it down, service the work, turn it back up.
Set expectations honestly before you start. Nobody can promise a lead volume or a cost per job in advance, and an agency that quotes one is selling you a number rather than a plan.
Questions we actually get
- What should an HVAC company expect to spend on Google Ads in this market?
- There is no honest single figure, and anyone quoting one has not looked at your service area or your close rate. What we can say is structural: budget has to be large enough for each county campaign to gather conversion data, or automated bidding never learns. If the budget only supports one market properly, run one market properly and add the second when the first is paying.
- Are Local Services Ads worth it alongside search campaigns?
- Usually yes, for the license and insurance verification alone, since it puts a badge in front of people who have no other way to check you. They are booked on leads rather than clicks, and the service area is set separately from your search campaigns, so it needs its own honest boundary. Treat it as a complement, not a replacement.
- Should I advertise across the bay if I only have one location?
- Only during the hours you can actually get there, and only with the geography named by city rather than drawn as a radius. Many contractors find the cross-bay campaign works in the morning and loses money after mid-afternoon. Test it as its own campaign so you can see the answer instead of guessing.
- How quickly can paid search produce calls?
- Calls can start the day it turns on, which is the real appeal of the channel. Cost efficiency takes longer, because negative keyword mining and bid learning need a few weeks of data. Nobody can promise a volume or a timeline, and the first month is usually more expensive per job than the third.
- Who should own call tracking, us or the agency?
- You should own the account and the numbers, with the agency given access. If the tracking lives in an agency account, you lose your call history the day the relationship ends, and that history is the only record of what your marketing actually produced.
