Your Williamson County clicks and your City of Austin clicks are not the same product
An address inside the city carries a review process that an address in Williamson or Hays County generally does not, and the two produce different sales cycles, so they should not share a campaign.
A homeowner in Round Rock or Leander and a homeowner in Hyde Park may type identical searches, but the jobs behind those searches differ in schedule, in paperwork and in how long your money is at risk before revenue arrives. Averaging them into one campaign hides which half is working.
Split by county at minimum, and split the city out on its own. Set separate budgets, separate bids and separate landing pages. Write the ad copy for each so it names what the reader is actually facing, whether that is a suburban new build or a central lot with trees and a review step in front of it.
Target by presence, meaning people located in your areas, not people showing interest in them. Interest targeting quietly buys clicks from someone in another state reading about Austin, and pool ads are expensive enough that the waste shows up fast.
Then exclude the places you will not drive on a Tuesday. A San Marcos job and a Georgetown job cannot both be convenient, and paying to find out is the most expensive form of route planning there is.
The searcher asking whether a pool fits the lot is worth more than the one comparing shapes
Feasibility searches signal a real property, a real constraint and a decision that has not been made yet, which is exactly when a builder can still shape the project.
Head terms attract everybody, including the people three years out and the ones who wanted a price list. Feasibility phrasing is different. Somebody searching about pools on a sloped lot, about building near a protected tree, or about what happens when excavation hits rock is standing in a specific back yard with a specific problem.
Build a campaign for that intent with tightly themed ad groups and landing pages that answer the question before asking for anything. Slope and access. Trees and cover. Small lot feasibility. Renovation of an older pool where access is the constraint.
Keep new construction and renovation apart. Remodel clicks usually cost less and convert on a shorter clock, and if they share a budget with new construction the cheaper clicks will absorb the money while the expensive campaign starves.
Review the search terms report weekly for the first two months, not monthly. In a small metro with limited volume, one bad match type can spend a week of budget before a monthly report notices.
Zilker swimming, aquatic centers and billiard halls: where a Central Texas pool budget leaks
Most wasted spend in a pool account is not competitive, it is irrelevant, and the irrelevant traffic here has a distinctly local flavor.
The word pool does more jobs than any advertiser would like. People search for public swimming around Zilker, for city aquatic center hours, for swim lessons, for pool halls, for apartment complexes with a pool, for hotel amenities, and for jobs on a construction crew. Every one of those can match a loose keyword.
Build the negative list before launch and keep extending it. Free, hours, admission, lessons, swim team, cleaning, service, repair technician, jobs, hiring, salary, apartments, hotel, above ground, kits, DIY, and the names of public facilities and recreation programs as they appear in the reports.
Service and maintenance searches deserve their own decision. If you do not do weekly service, exclude the vocabulary entirely. If you do, give it a separate campaign with its own budget, since a service click and a construction click have nothing in common except a word.
Automated campaign types will find this traffic on their own. If you run them, fence them with tight budgets, strong brand and account level exclusions, and a hard rule that you review the placements and terms they buy.
Never put a utility program claim in ad copy a Pflugerville address cannot use
Austin Energy is a city-owned utility in a state where most customers pick a retail provider, so eligibility for anything utility related depends on which side of a service boundary an address sits.
Equipment ads are tempting, especially around heaters and efficient pumps. The trap is that program and rebate eligibility here follows the utility serving the address, and the boundary does not match the way people describe where they live. Someone with an Austin mailing address may be served by a different provider entirely.
Keep specific program claims, amounts and plan names out of ad copy and off landing pages. Terms change and generic Texas advice about shopping for an electricity provider is simply wrong inside the city. The safe and still useful line is that you help owners confirm what applies with the utility serving their address.
There is a positive version of this. Copy that says you sort out equipment questions correctly for each address reads as competence to a reader who has already been given conflicting advice from a statewide site.
Ad copy in a licensed trade deserves the same care generally. Claims about licensing, insurance and credentials should be verifiable on the landing page, and anything about outcomes should be worth defending later.
A Bee Cave hillside lead and a Pflugerville flat lot need different target costs
One target cost per acquisition across the whole account averages two businesses with different close rates, different job sizes and different amounts of work before a signature.
Hillside and access-constrained jobs take longer to sell. There are more site conditions to explain, more design revisions, and a higher chance the project changes shape. Flat suburban lots move faster. Paying the same acquisition cost for both means you are overpaying on one side and underbidding on the other.
Set targets by segment. Track the segment in the CRM from the first call, then feed signed jobs back into the ad platform as offline conversions so bidding optimizes toward work you actually booked instead of toward form fills.
Pace to capacity as well as to cost. Crews, excavation subcontractors and the front end of the schedule limit how much work you can take. When the calendar fills, throttle the campaigns you can afford to slow and keep the ones that feed the months that are still open.
The number to report on is cost per signed job by segment and by county. A blended cost per lead can look healthy while the profitable half of the account is being starved by the cheap half.
The landing page has to show a hillside if the ad sold one
Message match in this trade means the page proves you have built on the reader's kind of lot, and a shared homepage cannot do that for three different kinds of buyer.
Send hillside traffic to a page about slope, access, structure and what a site evaluation covers. Send central Austin renovation traffic to a page about working an older pool behind a narrow drive. Send suburban new construction traffic to a page about the process from design through fill.
Each page needs proof in the first screen: photographs of that kind of job, a plain statement of where you build, and a way to make contact that fits a phone held in a back yard. Keep the phone number tappable and keep the form short enough to finish standing up.
Ask the two or three questions that decide whether a visit is worth the drive, usually the address, the lot condition and what stage the owner is at. Everything else can wait for the call.
Speed is part of the buy. You are paying for every visit, so a page that stalls loading a gallery is spending your money to display a spinner.
Questions we actually get
- What should we spend before we can tell whether paid search works?
- Enough to collect a meaningful number of real inquiries in each campaign you are testing, run long enough to see them reach a signed job. In pool construction the gap between click and contract is long, so judge the first stretch on inquiry quality and booked site evaluations, and only judge cost per signed job once a cohort has had time to close.
- Should we bid on competitors' names?
- It is legal to bid on them in most cases and it is often expensive and low converting for a considered purchase like a pool. Do not use a competitor's name in the ad text. If you test it, fence the budget, measure it separately, and be ready to shut it off quickly.
- Do we need call tracking?
- Yes. Most serious pool inquiries in this trade arrive by phone, and without tracked numbers you are optimizing on the minority of conversions that happen to be form fills. Record calls where permitted, and use the recordings to fix intake as well as bidding.
- Are Meta and YouTube worth running alongside search?
- They reach owners who are not searching yet, which suits renovation work in established neighborhoods. Treat them as a separate budget with a separate expectation, and do not let their cheaper conversions absorb money meant for high intent search.
- How do we count cost per booked job when the sale takes months?
- Stamp every inquiry with its source and segment in the CRM, then report by cohort: inquiries from a given month, and what those specific inquiries eventually signed. Importing signed jobs back as offline conversions lets the platform bid toward booked work rather than toward whatever generates the most cheap contacts.
