Split the account by what heats the building, not by repair versus install
A boiler search, a furnace search and a ductless search come from different buildings with different tickets and different crews, so they should not be competing for the same money.
Most accounts we look at arrive with one repair campaign, one installation campaign and a geographic target of the city plus twenty five miles. The heating budget then flows to whichever ad gets clicked most, which is rarely the work you most want.
A better shape: separate campaigns for hydronic and steam, forced air heating, cooling and ductless retrofits, and light commercial. Each gets its own budget, its own landing page and its own negative list. You can then starve one and feed another in a week without touching the rest.
The reason is dispatch, not tidiness. If one technician does your steam work, you cannot buy steam calls at the same rate you buy filter changes. Budget has to reflect who is actually on the truck tomorrow.
Light commercial belongs on its own. Rooftop units, a courtyard building boiler, restaurant makeup air. Different search terms, a longer sales cycle and a buyer who wants an invoice and a certificate of insurance before anything else.
ComEd outage checks and space heater shoppers will happily click your furnace ad
In a deep freeze the search results fill with people who are cold and not buying, and broad match will find every last one of them.
The waste here has a local shape. Outage lookups after a wind event, utility bill and assistance program searches, portable and space heater shopping, thermostat wiring questions, part numbers off a furnace label, and hiring searches from technicians looking for work in the same week you are advertising hardest.
Add the ones that look like customers and are not: builder warranty callers in a newer subdivision, home warranty companies who will dictate your rate, and tenants in a rental two-flat who cannot authorize a repair on the building they live in.
Review the search terms report weekly through the heating season rather than quarterly. Query patterns change overnight when the weather does, and broad match paired with automated bidding drifts fast when the pool of searchers changes.
Negative lists compound in value. Keep a shared library, apply it across campaigns, and treat it as an asset that belongs to the business rather than to whoever built the account.
A Loop high-rise inquiry is a facilities contract, not a service call
Downtown commercial clicks come from people with a purchasing process and a building engineer, and pricing them like a homeowner call loses money twice.
If you do that work, give it its own campaign, its own page and probably its own tracked number. The copy should talk about response windows, insurance, invoicing and after-hours access, not about comfort.
If you do not do that work, exclude it. A campaign built for bungalow furnace replacement should not be paying for clicks from a Loop or River North address, and it usually is because the geography was drawn as a circle.
Between the two sits the market a lot of Chicago shops actually live on: owners of two-flats, three-flats and courtyard buildings, plus the small property managers who look after them. They buy on response time and on whether you will invoice rather than ask for a card at the door.
Write ads for that buyer separately. Multiple units, tenant coordination and monthly invoicing are the words that get the click, and they never appear in a homeowner ad.
Shoulder weeks in April and October buy the maintenance agreements that carry January
Bidding only during the two peaks means paying peak prices for every job you get all year.
Demand here has two hard edges: the first real cold and the first real heat. In those weeks the auction gets expensive because every competitor in Chicagoland has the same idea on the same morning.
October is the cheaper month to buy attention for tune-ups, inspections and agreement signups. Every agreement sold in the fall is a customer you do not have to buy again during the January panic, and it is a service visit you can schedule rather than scramble for.
April works the same way for planned replacement. People who limped through the winter on a boiler that kept losing pressure are willing to talk in spring, and they will schedule around your crew instead of the other way around.
Write the season plan before the season, then revisit it weekly against capacity. If the board is full for six days out, extra spend is buying cancellations and one star reviews rather than revenue.
Bid the 10pm to 6am hours only if somebody actually answers at 10pm
Ad scheduling should match who is genuinely picking up, because a paid emergency click that reaches voicemail is money set on fire.
Winter demand does not respect office hours. The first hard freeze produces calls before six in the morning, from people who woke up cold and are working through a list of numbers.
Before you switch on overnight, call your own line at eleven on a Sunday from a phone nobody at the shop recognizes. If it goes to an answering service, listen to what they promise on your behalf. Most owners have never done this.
Where overnight is genuinely covered, it is worth bidding. Competition often thins in those hours, though that is worth checking in your own auction data rather than taking on faith.
During those hours push the call, not the form. Call assets and call-only ads, with the form as a fallback for somebody who cannot talk because the baby is asleep.
One month of a Chicago winter will not tell you whether the account works
Judge paid search on cost per booked job across a season, because two cold weeks can make a poor account look brilliant.
The chain runs click, call, answered, booked, run, sold, collected. Most reporting stops at the second step and calls it a lead. Everything interesting happens after that.
Feed the booked and sold events back into the platform as conversions so the bidding optimizes toward work you invoiced rather than toward form fills. Automated bidding is obedient and literal. It will get you more of whatever you tell it counts.
Compare like periods. November against November. Month over month movement in a Chicago heating season is mostly a weather report with a chart on it.
Segment the result by campaign and by county. A blended average across the city and the collar counties can hide one campaign paying for itself twice over and another quietly losing money all winter.
Questions we actually get
- What should we budget for Google Ads in the Chicago market?
- There is no honest single figure, and any agency quoting one before seeing your numbers is guessing. The starting point is capacity: how many additional jobs your trucks can absorb in a week, and what an average booked job is worth to you. We would set the initial budget against that, then adjust weekly through the season based on cost per booked job rather than cost per click.
- Are Local Services Ads worth running alongside search ads?
- For many residential HVAC shops they are worth testing, because they are priced per lead and the qualification process rewards the license and insurance paperwork you already keep. They also sit above the standard ads. Eligibility and verification requirements change, so it is worth confirming current terms directly rather than relying on a blog post.
- Should we advertise at all in the summer here?
- Yes, but with a different structure and usually a different budget. Cooling season is real in Chicago and it also carries ductless retrofits in houses with no ductwork, which is high ticket work. The mistake is running the winter campaigns with the ad text changed.
- Do we bid on competitor names?
- Sometimes, and it should be a deliberate decision with its own small budget. Those clicks can be inexpensive and the searcher already wants a contractor. They can also start an unpleasant bidding war with a shop down the street. We would test it in a contained campaign and read the booked jobs, not the click volume.
- How do we stop wasting spend during a freeze when the phone is already overwhelmed?
- Pace to dispatch. Turn budgets down when the board is full more than a few days out, keep the highest intent campaigns running and pause the broader ones, and make sure the phone offers a real later slot rather than a maybe. Spending into a queue you cannot serve buys cancellations.
