A radius drawn around downtown buys clicks from three Indiana counties
Set geography by what you are set up to serve, not by a circle, because a few miles north of the Ohio River you are in a different state.
Clark and Floyd County are inside almost any sensible radius from central Louisville. If your business is set up to work only on the Kentucky side, that radius is buying inquiries you have to turn down, and you are paying full price for the privilege.
Use explicit locations rather than rings: Jefferson County, then the outlying Kentucky counties you serve, then Indiana as a deliberate decision with its own campaign. Set targeting to presence rather than interest so you stop paying for people researching Louisville from out of state.
Requirements differ between Kentucky and Indiana and they change, so confirm what applies to your business with the relevant state and local authorities and your own counsel before you advertise across the line. The ad account is not the place to find out.
If you do serve both sides, say so in the ad copy for the Indiana campaign. A New Albany homeowner who has been burned by a Kentucky company that would not cross the bridge will click the ad that removes the doubt.
Split the account by which side of the river before you split it by product
Two states means two service realities, two sets of proof and often two different costs to serve, which is a bigger structural break than window versus door.
Start with a Kentucky campaign and an Indiana campaign. Inside each, split by job type: whole-house replacement, entry and patio doors, and single-opening or small jobs if you take them at all.
Whole-house replacement deserves most of the budget and the highest bids. The two-window inquiry deserves a cheaper path, either a lower bid or a page that sets expectations before it consumes an estimator's morning.
Negatives are where the money is saved. Sash cord repair, glass replacement, foggy or fogged glass, window film and tint, screen repair, window cleaning, curtains and blinds, DIY sizing questions, new construction and builder bids. Build the list before launch, then read the search terms report weekly for the first two months and keep pruning.
Keep the door campaign separate even though volume is smaller. Door searches are cheaper, closer to a decision, and they get smothered when they share a budget with whole-house terms.
Bid the pre-1930 blocks and the postwar subdivisions as separate campaigns
Housing age changes the job, the ticket and the close rate, and in this metro it maps neatly onto geography you can target.
Old Louisville, Germantown, Butchertown, parts of the Highlands and Crescent Hill are narrow lots and original openings. Custom sizes, more labor, more careful work, and a buyer who wants to talk about matching and trim before price.
St. Matthews, Jeffersontown, Middletown and the newer development out past the Watterson bring more standard openings and a more direct efficiency conversation. Prospect and parts of Oldham County bring larger homes and larger tickets.
Write ad copy to match. An ad that mentions original openings and matching a hundred year old profile will underperform in a 1994 subdivision and outperform everywhere inside the old street grid.
Bid accordingly. If your average signed value in the older neighborhoods is meaningfully higher, that difference belongs in the bid, not in a note somebody keeps in their head.
Overnight shift work means the ad schedule should not stop at nine
A large logistics and air-freight base puts a lot of this market on hours that do not match a standard ad schedule or a standard phone answer window.
People finishing a shift at two in the morning search then. Turning ads off overnight to save money looks efficient in a spreadsheet and quietly removes a slice of demand nobody else in the market is competing hard for.
The real constraint is not the ad schedule, it is what happens after the click. Overnight clicks need a page that can take a full inquiry and a follow-up process that responds early in the morning, not two days later.
If you cannot cover those hours at all, restrict the schedule deliberately and put the money into the hours you can answer. Both choices are defensible. Drifting into a default schedule nobody chose is not.
The same population turns over. Relocating workers arrive with no contractor, no neighbor to ask and no brand preference, which is exactly the buyer paid search is best at reaching.
Winter weather cancels measures, so the budget should pause when the crews do
Pacing in this market follows the install and measure calendar, and real winters interrupt both.
Cold weather drives demand and also takes days off the calendar. Spending at full pace during a stretch when nobody can hold a measure appointment produces leads that go stale before anyone reaches them.
Tie daily budgets to measure capacity for the coming week rather than to a monthly number divided by thirty. When the schedule fills or the weather closes it, pull back. When it opens, push.
Ordered unit lead times matter to pacing too. If you are quoting long delivery, some of the demand you buy in the winter converts to a signed job that installs much later, which is fine as long as the follow-up is built for it and the cash flow expects it.
Avoid the reflex of spending the remaining budget in the last week of the month. It buys the worst inquiries of the cycle and teaches you nothing.
Your ad account cannot tell a Floyd County contract from a Fern Creek one unless you feed it back
Optimizing to form fills teaches the platform to find you form fills, which is not the thing you sell.
Get signed jobs back into the ad platform. Import the offline conversion when a measure is held and again when a contract is signed, with the value attached. Once the platform is optimizing toward contract value, geography and keyword decisions start correcting themselves.
Track cost per booked job, not cost per lead. A lead number that looks good while the measure hold rate collapses is a number that is lying to you.
Read the results by county. Jefferson, Oldham, Bullitt, Clark and Floyd should each have their own line. A blended metro average is where a losing Indiana campaign hides inside a winning Kentucky one.
Give it enough time to mean something. One metro does not generate the volume for weekly conclusions, so set a review cadence measured in months and resist rebuilding the account every time a bad week shows up.
Questions we actually get
- What should we spend to test paid search in Louisville?
- There is no honest published number for this trade and metro, and anyone quoting one is guessing. Work backward instead: take your average signed job value and your historical close rate from a booked measure, decide what you can pay for a booked job and still make money, and fund a test long enough to produce a meaningful number of booked measures rather than a meaningful number of clicks.
- Should we run Local Services Ads as well as search ads?
- They compete for the same phone and are worth testing separately, with their own tracking, so you can see which one produced the job. Keep the eligibility and badge requirements current, because they change. Do not let a strong week in one starve the other before you have enough booked jobs to tell them apart.
- Is paid social worth it for window and door replacement here?
- Search catches people who have already decided. Paid social is how you reach the household that has put the job off for three winters and the recently relocated worker who does not know anyone yet. Expect a longer path to a signed contract and measure it to the contract, not to the lead form, or it will always look worse than search.
- How do we keep from paying for repair calls?
- Build the negative list before launch and prune it weekly at first. Sash cord and sash repair, glass replacement, foggy glass, screen and storm door repair, tint and film, blinds, cleaning, and new construction terms are the usual drains in an older housing market. Then check that your landing page is not inviting repair inquiries with soft language about fixing windows.
- Should the Indiana campaign send traffic to a different page?
- Yes, if you serve Indiana. A Clark or Floyd County searcher wants confirmation that you cross the river before anything else, and a page that says it plainly, with work you did over there, will hold that click. Confirm what your business is required to hold in each state with the appropriate authorities and your own counsel, and keep anything you publish about it accurate and current.
