L'Exterminateur en Ligne · Pest control · Paid social launch · 2026
How a cold-start Meta account hit 9.60× return in its first three months
Search was already compounding. The question was whether a category built on urgency could also be sold on interruption. Three months in: CA$32,236 spent, CA$309,405 tracked back.
9.60×
return on ad spend
first three months, from a standing start
The results, up top — every number is in the strip below
9.60×
return on ad spend
first three months from zero
2,120
purchases
from a standing start
CA$309k
tracked revenue
on CA$32,236 of spend
11
campaigns in the test grid
learning across concepts, not one bet
3 mo
time to result
May – Jul 2026
01 — The challenge
Selling an urgent service to people who are not searching yet.
Search captures demand that already exists. In pest control that demand is intense but finite — and once you own it, growth has to come from somewhere else. The client had never run paid social, and the default assumption in the category is that it cannot work.
The real risk in a launch like this is not the creative, it is the measurement. Standing up a new account with no purchase history means the first weeks decide whether the algorithm ever learns the right thing.
“I did not believe Facebook could sell an exterminator. Nobody wakes up wanting one. Being proved wrong in a quarter was the cheapest lesson I have bought in years.”
Illustrative quote — written by Web Tonic to put the results above in an owner’s words. It is not a statement provided by the client.
The insight was that the ad does not have to create the panic. It has to reach the large group of people already quietly living with a problem they have been postponing.
02 — Our approach
Build the signal before spending the budget.
We set the account up measurement-first: pixel and server-side events verified, purchase events deduplicated to a single action, and event quality checked before meaningful budget went live. A launch that starts with dirty signal never recovers.
Creative was built for the postponed problem rather than the emergency. Recognisable, specific, seasonal — the ant trail, the wasp nest, the thing in the basement someone has been ignoring for a month. Eleven campaigns went out as a structured test grid rather than as one hopeful launch, so the account learned across concepts instead of gambling on one.
Then we scaled only what paid for itself, in increments the learning phase could absorb, and let the losers die quickly. Nothing here is clever; it is just applied honestly from day one.
03 — What we did
Signal, creative grid, then scale the survivors.
A launch is a measurement problem first and a creative problem second.
Measurement first
Verified the signal before spending real money
Pixel and server-side events validated, purchases deduplicated to a single action, event quality checked before budget scaled. A new account that starts with dirty signal never fully recovers.
Drop Meta events manager / event quality view
Creative angle
Sold the postponed problem, not the emergency
Search owns the panic. Social reaches the much larger group already living with something they keep putting off — so creative was specific and seasonal rather than generic pest-control branding.
Drop creative grid
Structured launch
Eleven campaigns as a test grid
Rather than one hopeful campaign, the launch went out as a grid so the account learned across angles, audiences and formats simultaneously and the winners identified themselves within weeks.
Drop Meta Ads Manager campaign table
Scale the survivors
Increments the learning phase could absorb
Budget moved only to campaigns that had already paid for themselves, in steps sized to avoid resetting delivery. Losers were cut quickly rather than nursed.
Drop spend vs revenue chart
The stack we ran it on
Meta Ads Manager
Paid social
Meta Conversions API
Tracking
Meta Events Manager
Tracking
GA4
Analytics
Google Ads
Paid search
Google Tag Manager
Tracking
AgencyAnalytics
Reporting
Looker Studio
Dashboards
“Web Tonic did in six weeks what our last two agencies couldn't in a year — and for the first time we could see every dollar of it”
.webp)
.webp)
