[ Case study ]
Click campaigns ran 24/7 to a landing page with a form. Night clicks — often panicking homeowners — bounced or never submitted, while daytime research clicks cost the same. The economics ignored the single most important variable: when the pipe burst.
CLIENT a 24/7 emergency plumbing service — FOCUS Buy the moment, not the day
Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.
Nights, weekends, and burst pipes define the emergency plumbing service — a licensed operator who answers his own phone, one van, and a small daytime crew for scheduled work. Emergency jobs carry the margin; daytime research queries bring tomorrow's scheduled bookings. Google Ads had run for years on continuous click campaigns pointed at a form-bearing landing page, with a Local Services profile sitting alongside it, both managed by a previous provider who treated the two as interchangeable.
Click campaigns ran 24/7 to a landing page with a form. Night clicks — often panicking homeowners — bounced or never submitted, while daytime research clicks cost the same. The economics ignored the single most important variable: when the pipe burst.
We proposed buying the moment instead of the day. Search campaigns split by daypart and intent temperature: night windows run call-first ads with call-only actions, daytime research queries run the standard flow. The overnight schedule encodes the on-call rota — bids only run when a human can actually answer. Local Services Ads sit on the same dayparted terms to cover the highest-intent searches, and conversions count booked jobs verified against the job book, not clicks.
Just as important is what we ruled out, and why:
Search campaigns split by daypart and intent-temperature — night campaigns use call-first ads with Local Services Ads on the same terms, daytime runs standard flow.
Overnight bids only run when the on-call phone can actually answer — the schedule encodes the plumber's real night, not Google's.
Call-only conversion actions with qualifying prompts verified against the job book — the number being optimized is jobs.
Delivered by the growth pod — paid search specialist over 4 weeks, with working increments reviewed with the client every week.
Obstacle
Call logs from week two showed two overnight windows where nobody answered the on-call phone — the rota the ads were about to depend on had holes.
Handled: We paused bids in those windows and rewrote the schedule against the plumber's real night, converting dead spend into capacity the daytime crew could absorb.
Obstacle
Local Services Ads verification stalled on license and insurance documents — Google's review requested a re-submission, costing ten days on a four-week engagement.
Handled: We ran the search-side restructure first while the review processed, and LSA launched in the final week with the dayparted terms already proven on search.
Obstacle
The first call-only night ads pulled research calls too — homeowners browsing tomorrow's bathroom project at ten at night, dialing an emergency line.
Handled: We sharpened night ad copy to emergency-only language and let the qualifying prompt filter the rest; the job book confirmed the night mix cleaned up within a week.
The headline: night-window booked jobs per month at lower spend (wasted overnight clicks cut 62%) — +73%, read from Job book vs call logs. A second check: cost per booked emergency job at −31%.
The plumber trusts the phone at night again — the calls that come are emergencies, not browsers, and he stopped screening numbers from the ads. His night schedule and the ad schedule are the same document, so the rota conversation happens once, not weekly. Daytime scheduled work keeps flowing from research queries without emergency budget feeding it. He describes the account in job terms now — nights covered, calls answered, work booked — and has stopped describing it in clicks.
The result was read from Job book vs call logs against the pre-engagement baseline over the stated window, with a guardrail check on cost per booked emergency job. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.
What we would do differently
We would have asked about the on-call rota first — two overnight windows had nobody answering, which we discovered from the call logs in week two.
[ Related service ]
[ Related builds ]
70% 34%Revenue share of the single best ad, replaced by a portfolio of five proven concepts
0.8s 2.3sAverage 3-second hold rate on ads (proxy for hook survival), cohort over cohort
[ Next step ]
Next case study