Tile and foam bring different money, and one campaign averages them away
Splitting the account by covering type is the first structural decision, because the jobs hiding behind those searches are not worth the same.
A tile underlayment replacement, a foam or coating job on a low-slope section, a shingle repair and a light commercial flat roof are four businesses sharing a keyword theme. Blended into one campaign they produce a cost per lead that describes none of them.
Give each its own campaign, its own budget and its own landing page. The ad that mentions tile should land on tile content, and the form on that page should already assume tile. Match all the way down or the click is half wasted.
Separate repair from replacement while you are at it. Repair searches convert quickly, spend less, and often mature into replacement work later, which is a good reason to keep buying them and a bad reason to judge them on first-job revenue.
Once the split exists you can see which lines deserve more money. Before it exists, every reallocation decision is a guess dressed up as optimization.
Target the named communities your dispatcher already says out loud
Geography here is a grid with long drives, so build targeting from the places your team names rather than a circle on a map.
Your crews talk about Summerlin, Spring Valley, Enterprise, Paradise, Silverado Ranch, Centennial Hills, Anthem and Green Valley. They also talk about Henderson and North Las Vegas as different runs entirely, and Boulder City as the trip nobody volunteers for. Build the account the way the schedule board is built.
The 215 sets the real cost of a job as much as the roof does. Two identical jobs on opposite sides of the loop have different margins once you count crew hours, fuel and a second trip for a punch item. Bid adjustments should reflect that, and some areas deserve to be excluded outright rather than bid down.
Jurisdiction matters too. Permitting differs between unincorporated Clark County and the separate cities, and if that changes your lead time or your cost, it belongs in the geographic structure, not buried as a surprise at contract time.
Watch the location reports monthly. Radius targeting drifts, and the money leaks toward the edges where nobody is looking.
Storm and claim vocabulary buys the wrong call in a market with no storms
Copying ad copy written for hail country invites conversations your team cannot help with and puts risky language on a public page.
Storm-market roofing ads run on claims, adjusters and free inspections tied to weather events. Here that language attracts confusion at best. Keep insurance framing out of the copy unless your own counsel and your insurer have looked at it, and describe what you actually do: inspect, quote, replace, warranty.
The negative list is where the real savings sit. Start with hiring and salary searches, DIY coating products, big box retail, RV and trailer roofs, roof racks, roof cleaning, and anything that reads like a homeowner looking for a free product sample. Review search terms weekly for the first month and then monthly forever.
Commercial and residential bleed into each other constantly. If you do not want light commercial flat work, exclude it explicitly, because the broad match engine will happily sell it to you at residential prices.
Every negative you add is permanent savings. Every bid change you make expires the next time the auction moves.
A shift-work city bids at hours your office does not keep
A round-the-clock hospitality economy puts serious buyers on their phones at times a standard ad schedule treats as dead.
Someone finishing a swing shift is researching a roof at one in the morning. Someone starting a graveyard shift is making calls in the middle of the afternoon. If ads run and nobody answers, you paid for a competitor's callback.
Measure before you schedule. Pull missed calls by hour for a month, then decide which hours to buy. Buying an hour you cannot staff is worse than not buying it, because the click still costs the same.
If you can answer outside business hours, say so in the ad and on the landing page, and make it true. Availability is a genuine differentiator here rather than a throwaway line, and it is one of the few claims a roofer can make that competitors cannot copy without hiring somebody.
A text-back option and a form that books a real inspection slot cover the hours nobody picks up. Both are cheaper than losing the call.
In July your constraint is crew hours before the heat, not budget
Spend should follow the work you can actually install, and installed capacity moves with the season here in a way that flat monthly pacing ignores.
Summer heat compresses the productive part of a roofing day. Crews start early, finish earlier, and a tear-off takes the hours it takes. Buying more demand into a backlog you cannot schedule creates canceled appointments and reviews that describe a company that never showed up.
Pace against the schedule board. When backlog runs long, pull spend to the highest value covering lines and let the rest sit. When the board opens up in the milder months, push harder, because the clicks are the same and the crews are free.
Set the account up so pacing is a lever you can pull weekly rather than a monthly budget line you notice on the fifth. Shared budgets and automated bidding will happily spend everything by the middle of the month if nobody is watching.
The number that should scare you is not underspend. It is a full calendar with three angry homeowners waiting on a reschedule.
An approval step sits inside your attribution window, so widen it
Cost per lead flatters an account that buys the wrong jobs, and standard conversion windows cut off before a valley roofing job is actually sold.
Feed signed jobs back into the platform rather than form fills. A tile underlayment contract and a repair invoice cannot share a column, and the algorithm optimizes toward whatever you feed it, which is usually whoever fills forms fastest.
Because association approval frequently sits between the quote and the job, the gap between click and signature is longer here than the default reporting window assumes. Widen the window, and read monthly cohorts rather than weekly totals, or you will kill campaigns that were working.
Report cost per signed roof by covering type and by area. Two campaigns with the same cost per lead can differ by a factor you would not believe once drive time and job size are in the same table.
Keep one honest column for jobs you turned down. Buying demand you decline is a targeting problem, and it shows up nowhere else in the reporting.
Questions we actually get
- How should we split budget between tile, foam and shingle work?
- Start by splitting them at all, then let signed job data move the money. Give each covering its own campaign and landing page so the reporting is honest, and hold the split long enough to read a real cohort. Reallocating weekly on cost per lead is how accounts end up buying whichever job is cheapest to inquire about rather than best to sell.
- Is it worth running ads overnight?
- Often yes here, but only after you check missed calls by hour. A shift-working population produces genuine inquiries at hours other markets do not, so the demand is real. The question is whether somebody answers. If no one does, buy a text-back path and a booking form first, then turn the hours on.
- Do we need separate campaigns for Henderson and North Las Vegas?
- If your drive time, permitting or crew assignment differs, then separate campaigns give you the reporting to prove it. They are separate cities, not neighborhoods, and lumping them into one valley-wide campaign hides both the cost and the win rate. Start with separate ad groups if budget is tight, and split when the data justifies it.
- What is a reasonable cost per lead for roofing here?
- We will not quote you one, because a number without your covering mix, your close rate and your average job size is a made up figure. What we will do is build the account so you can see cost per signed roof by covering and by area within your own data, which is the number that actually decides whether to spend more.
- Can paid ads and organic run at the same time without wasting money?
- They usually help each other. Paid buys reach today and produces search term data that tells you which content to write next; organic lowers your dependence on the auction over time. The waste comes from pointing both at the same generic page instead of covering-specific ones.
