Build the account around who pulls the permit, not around product lines
The clearest way to split a window and door account in this market is by the job's administrative shape, because that is what actually drives cost.
A whole house replacement in a brick home, a single patio door swap, and a landlord replacing units across a small portfolio are three different sales. They have different ticket sizes, different close rates, and different amounts of paperwork before anyone installs anything.
Group campaigns so that budget cannot slide between them. Whole house replacement in one, doors in another, and anything that involves a property manager or multiple addresses in a third. Each gets its own budget, its own landing page and its own definition of a good outcome.
The permit angle matters because it predicts the work. A job that needs a permit pulled in a small county municipality carries scheduling cost that a screen replacement does not, and blending them into one campaign means the cheap clicks quietly starve the profitable ones.
Once the split exists you can finally answer the question owners actually ask: which of these three is worth more money next month.
A radius from your showroom buys clicks in two states before lunch
Default radius targeting in this metro is close to a guarantee that you will pay for jobs you cannot profitably reach.
Draw a thirty mile circle around most St. Louis dealers and you have bought the City of St. Louis, a large stretch of St. Louis County, part of St. Charles County and the Metro East on the Illinois side. Licensing generally differs by state, so half that circle may be territory you are not set up to sell in at all.
Target by named locations instead. Municipalities, counties, and the Illinois side handled as its own thing. If you do not sell in Illinois, exclude it outright rather than trusting a bid adjustment to hold the line.
Location settings also need attention. Presence based targeting keeps you from paying for someone in another state researching a move, which is a real and expensive category of click in a metro that people relocate into.
Then set bids by what the drive actually costs. Chesterfield and Florissant sit in the same county and are nowhere near each other, and a crew that crosses the metro twice in a day has lost a half day of install capacity.
Tuckpointing, glass repair and screen jobs belong on the negative list before launch
An older brick housing stock generates enormous repair intent, and repair intent will eat a replacement budget by Thursday.
The failure mode is predictable. Freeze-thaw damage produces searches about cracked mortar, water at the head of an opening, fogged sealed units, broken panes and torn screens. Broad match will happily serve your whole house ad to every one of them.
Start the negative list before the campaign goes live rather than after the first invoice. Repair, fix, glass replacement, screen, tuckpointing, pointing, caulking, storm window repair, sash cord, and the whole vocabulary of small jobs you do not want a truck rolling for.
Add the other drains too: DIY and how to phrasing, sizing and measuring questions, new construction and builder bids, parts and hardware, and anything that reads like a rental turnover unless portfolio work is a line you are deliberately buying.
Then review search terms weekly for the first two months. In this market the repair vocabulary is deep enough that you will keep finding new versions of it for a while.
One caution. Do not negate the masonry vocabulary so hard that you lose the homeowner who has a brick problem and a window budget. Some of those become real jobs once the sequencing is explained.
A dozen permit desks in one afternoon's drive is a cost your bids should carry
St. Louis County is made of many small separate municipalities, each with its own permitting and inspection practice, and that administrative variation is real money.
Two identical jobs in two adjacent municipalities can carry different amounts of office time before anyone lifts a window. Drive time is the obvious cost. Paperwork variation is the one that never shows up in a bid strategy.
Handle it by grouping your geographic targets into tiers based on total cost to serve rather than on distance alone. Municipalities where your office already knows the process are cheaper to serve than ones where every job starts with a phone call to find out how they do things.
Bid accordingly, and be willing to bid a place down to nothing. A location you can serve but always lose money in is a location to exclude, not to optimize.
Say nothing specific about any jurisdiction's requirements in ad copy or on the landing page. Requirements differ, they change, and the safe framing is that permitting generally varies by municipality and is worth confirming with the local building department.
Spend should climb ahead of the first hard freeze and hold through the thaw
Demand in this trade is weather driven, and a flat monthly budget systematically underspends the weeks that matter most.
Real winters here mean the first genuinely cold stretch produces a wave of people who have been living with a drafty opening since last year and have finally had enough. The searches spike in a few days, not gradually.
Pace against that. Build headroom into the budget before the cold arrives rather than reacting a week later, and be ready to raise daily caps quickly when the phones move.
The other pressure point is the freeze-thaw stretch at the end of winter, when masonry damage becomes visible and water shows up where it did not before. Demand in that window skews toward urgency, which changes what the ad should say.
Coordinate with the install calendar, not with the calendar month. Spending hard in a week when you are booked six weeks out buys inquiries that go cold before you can reach them.
Slow periods are for the longer sell: the homeowner who has been putting off a whole house replacement for three winters and needs a reason to start now.
Cost per signed contract, read separately on each side of the river
Cost per lead is a number that will let you keep funding the wrong campaign for a year.
Form fills and calls are not commercially equal in this trade. A whole house inquiry from a brick two-family and a broken pane inquiry both cost you a click, and only one of them can pay for the month.
Pass a lead identifier through to the CRM so that campaign, ad group and search term follow the record all the way to a signed contract. Then report cost per booked measure and cost per signed job, with average ticket alongside.
Split the report by geography, at minimum City of St. Louis, St. Louis County, St. Charles County and the Metro East. Blended numbers across a metro this fragmented will hide a losing segment inside a winning average for months.
Do not expect the answer in two weeks. A window job takes time to close, and a report read before the pipeline has matured will point you at exactly the wrong campaign.
Questions we actually get
- What should a St. Louis window and door dealer budget for paid search?
- There is no honest industry number to quote, and any figure presented as a benchmark is usually borrowed from a different market. Work it backward instead: your average signed ticket, your close rate from booked measure to contract, and how many jobs your crews can install in a month. That gives you a ceiling on what a booked measure can be worth, and the budget follows from there.
- Should we run ads on the Illinois side of the river?
- Only if you are actually set up to sell and install there. Licensing generally differs by state, and it is worth confirming your position before you spend a dollar in the Metro East. If you are not selling there, exclude Illinois at the campaign level rather than relying on bid adjustments.
- Do Local Services Ads make sense for this trade?
- Availability by category and market changes, so it is worth checking what is currently offered for your business type. Where it is available it competes for the same phone as your search ads, so treat it as a channel with its own economics rather than an add on, and measure it to booked jobs the same way.
- How do we stop paying for repair clicks in an old brick market?
- Build the negative list before launch, covering repair, glass replacement, screens, tuckpointing, sash and caulking vocabulary, then review actual search terms weekly for the first couple of months. Tighter match types help. Leave a path open for the homeowner with a masonry problem and a replacement budget, because some of those are real jobs.
- How soon will we know whether the account is working?
- Later than you would like. A window replacement decision often runs weeks from first click to contract, so any read taken before the pipeline matures is measuring the wrong thing. Judge early weeks on lead quality and booked measures, and wait for signed contracts before making the big budget calls.
