Retail replacement, patio doors and storm damage need three separate campaigns
Three buyers with three budgets and three sales cycles do not belong in one auction.
Whole-house replacement is a considered purchase with a long deliberation and a high ticket. Patio and entry doors are often a single-item decision made in an afternoon. Storm damage inquiries run on a different clock entirely, with paperwork and a third party involved.
Blending them means the algorithm optimizes toward whichever converts fastest, which is almost never the whole-house work you actually want. Cheap conversions crowd out expensive ones.
Split the campaigns, split the budgets, and let each carry its own landing page and its own definition of a conversion. A whole-house campaign should be measured on booked in-home measures. A single-door campaign can be measured on a completed quote request, because sending an estimator across Denton County for one slider rarely pays.
Keep the storm campaign narrow and careful. It should route to a page that explains what you do, tells the homeowner to speak with their own insurer, and makes no representation about coverage. Texas regulates what a contractor may say about an insurance claim, and that campaign is worth showing to your own counsel before it runs.
Half of the "my door will not close" clicks are a foundation problem, not a door sale
North Texas generates a whole category of search intent that looks like your customer and is not.
Expansive clay moves slabs, and moving slabs rack openings. So people search for doors that stick, windows that will not slide, and gaps at the corners. Some of those are your job. A good share are foundation work, and you pay the same click price either way.
Handle it in two places. In the ad copy, be specific about replacement units and installation rather than writing to the symptom. On the landing page, say plainly what movement does to an opening and when the answer is a licensed engineer, so the wrong caller self-selects out before they occupy your phone line.
Negatives do the rest of the work. Repair, glass only, screen, tint, auto, board up, and the long tail of manufacturer part numbers all drain a budget quietly. Pull the search terms report weekly for the first two months, then monthly.
Where the intent is genuinely mixed, a cheaper path is better than an exclusion. A short form that captures photographs and the opening count costs you nothing and occasionally turns a foundation question into a real job six months later.
One drive from Frisco to Arlington can eat the margin on a three window job
Geography is the largest lever in this account and most dealers never touch it.
Dallas and Fort Worth are separate markets with an hour of freeway between them. A radius drawn to include both buys inquiries that a crew cannot serve profitably on a small job, and the sales team burns days proving it.
Draw the target from the install calendar backward. Where do crews already go without complaint? Where does a job need a minimum ticket to justify the drive? Build tiered geography that reflects those answers instead of a circle around the office.
Use location bid adjustments to enforce it. Core zip codes where you win and the drive is short get the aggressive bid. The outer ring gets a reduced bid and, ideally, a separate landing page with a minimum project size stated on it.
If you serve Tarrant County properly, with crews and stocked product and real local work, run it as its own campaign with its own copy. If you do not, exclude it. A half-served county is worse than an unserved one.
Bid Southlake and Highland Park at one number and Mesquite at another
The same click is worth different money depending on the housing stock behind it.
Ticket size varies enormously across this metro. Older Highland Park and Preston Hollow houses need custom sizing and matching, which is high-value and slow. Newer Southlake and Frisco homes often need a whole-house swap of failing builder units, which is high-value and fast. Other submarkets skew toward one or two openings.
Let the bid follow the expected job, not the population. A modest CPC in an area that produces single-opening inquiries can be more expensive per signed job than a high CPC where whole-house work is normal.
Copy should follow the same logic. Ads pointed at older neighborhoods can talk about matching existing openings and non-standard sizes. Ads pointed at the northern corridor can talk about failed seals and fogged glass in homes built during the growth years.
Review the split quarterly against actual sold work. Bid adjustments set once in January and never revisited are how paid accounts drift away from the business.
Paid social sells the upgrade a homeowner has put off for three summers
Search catches people who already decided; social creates the decision in a market where summers make the problem obvious every year.
Nobody wakes up wanting to buy windows. They notice a hot room in July, a door that sticks, a bill that annoyed them, and they file it. Social is where you make that itch specific enough to act on.
What works is unglamorous. Real installation footage, a walkthrough of an old opening being brought back to square, a plain explanation of what a retrofit does not fix. Stock renderings of houses that could not exist in North Texas do not.
Target by geography and homeowner status, and keep it inside the same radius your search campaigns respect. A cheap impression outside the drive is still waste.
Expect the reporting to look worse than search and value it anyway. Social conversions are usually earlier and softer, so a form fill from social often needs a longer follow-up sequence rather than an immediate estimator visit.
Cost per booked measure, and a budget that survives the week after a hail night
Two numbers govern this channel: what a booked appointment costs and whether your pacing survives the weeks the market moves without warning.
Cost per lead is misleading in this trade, because a lead can be a whole house or a broken latch. Track through to the booked in-home measure, and where your CRM allows it, to opening count on signed work. Feed the appointment back into the ad platform as the conversion so the bidding optimizes toward it.
Storm activity distorts spend fast. Auction pressure jumps, unrelated intent floods in, and a budget set for a normal month empties in days. Set caps that do not blow through the month, and decide in advance whether you raise or step back when it happens.
Seasonality matters the other way too. Demand rises ahead of the hot months as people remember last summer. Pace toward that, and pace toward the install calendar rather than the accounting month, because selling work you cannot install for ten weeks damages the follow-through.
Nobody can promise you a cost per job or a volume of calls before the account has run. What is knowable is the structure: separate the buyers, respect the drive, measure the appointment. The rest is arithmetic you do with your own numbers.
Questions we actually get
- How should I set my service area radius in Google Ads?
- Work backward from the install calendar rather than drawing a circle. Identify where crews already travel without eroding margin, tier the geography around that, and use location bid adjustments to reflect it. If a job in a far county needs a minimum project size to be worth the drive, either say so on a separate landing page or exclude the area.
- What are the most important negative keywords for a window and door dealer here?
- Repair, glass only, screen, tint, auto glass, board up and manufacturer part numbers are the usual budget drains. In North Texas add the foundation-adjacent phrasing, because searches about doors that will not close often lead to structural work rather than a replacement unit. Read the search terms report weekly at first, then monthly.
- Should I bid during and after hail events?
- It depends on whether you are set up for that work and comfortable with the compliance side. Auction prices rise, competition includes operators who arrived with the storm, and the intent is mixed. If you do run it, keep it as a separate campaign with its own budget cap and a landing page that directs the homeowner to their own insurer and makes no claim representations. Have counsel review it.
- Is paid social worth it alongside search?
- Usually as a complement, not a substitute. Search captures people who have already decided to replace something. Social is where the decision gets made, which means softer conversions and a longer follow-up. Judge it on assisted booked measures over a quarter rather than on same-week form fills, and keep the targeting inside the geography your crews actually serve.
- What should I report on instead of cost per lead?
- Cost per booked in-home measure is the minimum useful number, because it filters out inquiries that never become an appointment. If your CRM can carry it, track opening count and signed value by campaign and submarket. A campaign producing cheap single-opening inquiries can easily cost more per signed job than one producing expensive whole-house appointments.
