The voicemail greeting is the expensive part. Somebody standing in water on a kitchen floor tapped your ad, got a recording instead of a person, and dialed the next plumber on the page before the greeting finished playing. You were charged for that tap the instant it happened, and plumbing gets charged more than almost anything a small contractor can advertise. Nothing about that job was lost to weak ad copy or a slow landing page. It was lost in the gap between the tap and the hello — and most accounts running right now have no field anywhere in them that records it happened at all.
The job is lost before anyone says hello
Put your ad schedule and your call log side by side before you touch anything else. If the campaigns run into the evening and the phone rolls to voicemail after the last truck comes in, you are buying the most expensive hours of the week and handing the work to whoever picks up instead. There are two honest fixes: take those hours off the schedule, or put a human — yours or an answering service that can actually dispatch — behind them. Paying for neither is the option most accounts are quietly running.
Emergency money and planned money
A burst supply line and a water heater quietly rusting through are the same keyword and two different businesses. The emergency caller doesn't compare — they book whoever answers, they barely flinch at the price, and you may never hear from them again. The planned caller is shopping a repipe, a sewer line, a softener, a bathroom that's getting gutted anyway: they read reviews, collect estimates, and sign weeks after the click that found you. Run them through one campaign with one conversion called "phone call" and the bidding will chase whichever is cheaper to produce, which is never the one that carries the bigger ticket. Separate them, and check that your conversion window doesn't close before the estimate gets signed.
Your own off-season is when you go shopping
This is a small topic — 4 related queries merged together, 320 searches a month in total. The shape matters more than the size. It peaked at 480 in September of last year and fell to 10 this past February. Nobody lost interest in marketing in February. February is freeze season, and those owners were under somebody's house. Demand for your service and demand for help with your advertising move in opposite directions, which means most plumbing accounts get built in a slow month by an owner with a free afternoon and then left running unattended through the exact weeks they spend fastest. Whatever you set up in September has to survive January without you watching it.
Grade the account on jobs, not rings
Ad platforms optimize toward whatever you label a win. Label every ring a win and you'll get rings — wrong numbers, price shoppers, a guy asking if you rent augers. The column on the left below is what most plumbing reports are built from. The column on the right is what you'd need to see to know whether the spend paid for itself.
| What the report counts | What you actually need counted |
|---|---|
| Phone calls | Calls long enough to become a booking |
| Cost per lead | Cost per booked job |
| Form fills | Form fills with an address you'd drive to |
| Clicks | Clicks in hours when someone answers |
| Monthly total | Revenue by campaign, once the invoice closes |
What to change before you raise the budget
Trim the schedule to hours a person answers. Draw the service area by drive time, not by a circle, and cut the towns you'd resent driving to at dinnertime. Add negatives for the searches that will never book: how-to, parts, supply house, apprenticeship, salary, DIY. Turn on call tracking with a minimum call length so a misdial stops counting as a lead. Then close the loop — when a call becomes a booked job in your scheduling software, that status needs to travel back to the ad platform, because that feedback is the only thing that teaches bidding the difference between a repipe and a wrong number.
Who should be holding this
An owner can run the schedule, the negatives and the budget alone; that's an hour on a Sunday and it's worth doing yourself. What tends to break is the measurement chain — call tracking that maps to the job, values attached to conversions, the booked-job feed going back out to the platform. That's a setup problem more than a monthly one, which is why plenty of shops bring someone in for a defined stretch and take it back after. If you do, ask what they intend to count and where those numbers come from before the proposal, not after it. And keep the ad account and the analytics property in your own name, whoever is doing the work.