Set St. Tammany up as its own budget, or leave the Northshore alone
Lake Pontchartrain is not a suburb boundary, and an account that ignores it spends Orleans money in Covington.
A radius from a shop in Mid-City or Metairie will happily include Mandeville and Slidell. The clicks are real, the calls are real, and the dispatch is a different business. Unless you have crews stationed on the Northshore, every one of those calls ends in an apology or an arrival window you should not have offered.
Run the Northshore as its own campaign with its own budget, its own ad copy and its own landing page, or exclude St. Tammany Parish entirely. The middle option, taking the clicks and hoping, is the expensive one.
The West Bank deserves the same decision. Algiers is close on a map and separated by a bridge in practice. Decide whether you cross, then set the targeting to match the decision rather than the ambition.
Use location targeting set to presence rather than interest, and check the location reports monthly. Metro level targeting has a way of drifting into places nobody agreed to serve.
The damp house call is worth more than the no cool call, so bid it separately
Two people with a comfort problem type very different things, and only one of them is shopping for a repair.
Split the account by what the searcher is trying to buy. Emergency repair. System replacement. Humidity, duct and sizing problems. Maintenance agreements. Four campaigns, four sets of ad copy, four landing pages, four bids.
Emergency terms convert fast, cost more and reward availability. Somebody typing about a house that feels wet is earlier, slower and often worth more, because the honest answer is frequently a sizing or duct problem rather than a part. Judging both on the same conversion window makes the second look like a failure.
Adjust by hour with the same logic. There is no point winning the top position for an emergency term at four in the afternoon if the truck cannot get there, and no point bidding replacement research at midnight if nobody follows up until Tuesday.
Give the humidity and duct campaign a longer measurement window before you cut it. Deciding within a fortnight on a purchase that takes a month to make is how good campaigns get killed.
Mini split retrofits are a second business inside your ad account
A city full of houses with no ducts creates a search market your repair campaign was never built to catch.
Shotgun houses and Creole cottages built without duct chases are a standing demand for ductless work. The searches are distinct, the buyer is not in a hurry, and the ticket is not a service call.
Give it a dedicated campaign and a landing page that shows the work: a head mounted in a real room, a line set run handled cleanly, an outdoor unit placed sensibly beside a raised house. Photography of your own installs does more here than any headline.
Exterior placement in parts of the city is subject to historic district review, so write the ad copy to promise a conversation rather than an outcome. Never imply an approval, and say that requirements for a specific address are worth confirming with the relevant commission or the city.
Measure this campaign on booked estimates rather than clicks. If the field software records which estimates came from paid, you will find the ticket sizes justify a bid the repair campaign could never support.
Renters, window unit shoppers and Entergy bill questions all look like demand
The exclusions decide the account, and this market adds a few entries most templates miss.
Start with the obvious waste: parts numbers, wiring diagrams, portable and window units, DIY instructions, training and job searches, home warranty terms. Add utility billing and outage queries, which pull in Entergy customers looking for anything other than a contractor.
Then handle the local one. Doubles and rented halves of doubles mean the caller is often not the person who owns the equipment. You cannot exclude that with a keyword, so handle it in the ad copy and the first form field: ask whether they own the property. A tenant call that reaches a dispatcher is thirty minutes nobody gets back.
Read the search terms report on a schedule, not when results dip. Broad match will keep finding new ways to spend, and the negative list is a living document rather than a setup task.
Structure helps as much as exclusion. Tighter ad groups with matching landing pages generally reduce the number of loose matches Google feels entitled to serve you.
A street that will not drain stops the trucks, not the spend
Pacing should follow your ability to actually show up.
Heavy rain and street drainage can take half a dispatch day out of the schedule. The board fills with rescheduled work, the crews are behind, and the ad account keeps buying emergency calls at full price into a business that cannot take them.
Build the ability to react. Automated rules that pull budget back when the board is full, and push it up when the schedule thins, are worth more than a clever bid strategy. Somebody should have the authority to make that call the same morning.
Pace against capacity rather than evenly across the month. Spending in equal daily amounts assumes every day has the same number of available truck hours, which is not true in any week here.
Keep a note of the days you throttled and why. Over a season it becomes the clearest picture you will get of where dispatch capacity, not the account, is the ceiling.
Nobody has matched an invoice back to a click, so your parish numbers are guesses
Cost per lead is a comfortable number that hides which half of your map loses money.
Put dynamic call tracking on the site so calls carry a source. Make source a required field in the field software. Then reconcile invoiced jobs back to campaigns at least monthly. Without that chain, every conclusion about paid search in this market is an opinion.
Once it is in place, cut the numbers by parish. Cost per booked job in Orleans, in Jefferson, and on the Northshore if you serve it. They will not match, and the gap is usually where the budget decision lives.
Push the outcomes back into the platform as offline conversions. Bidding optimizes toward whatever you report, so reporting form fills produces form fills. Reporting released changeouts changes what the auction chases on your behalf.
End every report on cost per booked job and, where the data supports it, cost per released replacement. Impressions, clicks and cost per lead belong in an appendix for diagnosis.
Questions we actually get
- How much should we budget to start?
- The honest answer is that budget should follow dispatch capacity rather than a benchmark. Work out how many additional jobs per week you can actually take, in which parishes, and build the spend up toward that ceiling. Starting larger than the schedule can absorb produces missed calls, which cost more than the clicks did.
- Should we bid on competitors' names?
- Sometimes, carefully. Competitor terms tend to be cheap and can work when somebody is checking a second option, but the intent is mixed and the volume is thin. Test it in its own ad group with its own budget so it never quietly consumes the emergency campaign, and judge it on booked work rather than on clicks.
- How do we know paid ads are not just capturing calls we would have received anyway?
- You partly cannot, and anyone claiming certainty is overselling. What helps is separating branded from non branded spend, watching what happens to total booked jobs rather than to paid booked jobs, and running geography or schedule holdouts when volume allows. Treat branded search as cheap insurance, not as evidence the program is working.
- Do we really need separate landing pages for each campaign?
- Yes, and it is usually the cheapest improvement available. An emergency repair visitor and someone researching a ductless retrofit need different first screens, different proof and different calls to action. Sending both to the home page wastes the click you paid a premium for, and it inflates cost per booked job across the whole account.
- Can you promise a cost per lead?
- No. Auction prices move, competitors change budgets, and demand swings with weather nobody controls. What we will do is instrument the chain from click to invoice, report cost per booked job by parish, and tell you plainly when a campaign is not earning its place rather than defending it with impression share.
