The framework
Your account sets the bar.
We test every ad against two bars: what your account already delivers, and the KPI you hired us to hit. Whichever sits higher is the bar.
The four moves
From budget to verdict.
Hypotheses get budgets. Ads get verdicts. The whole system falls out of two numbers: your budget and your cost per acquisition.
Step 1
Set the testing budget
TB = s × B
Testing stays a contained slice of your account, typically 10 to 25 percent of spend, sized to your budget.
Step 2
Fund hypotheses, not ads
Read = 20 × CPA
One ad set per hypothesis: an angle, an audience, a format. Inside it the auction feeds a favorite and filters the rest nearly free, and every ad is judged at a read floor of 2 to 6 times CPA.
Step 3
End every ad in a verdict
Reads = TB ÷ (20 × CPA)
Losers die at the floor. Ads the auction starves cost almost nothing and simply end. Nothing lingers half funded.
Step 4
Scale where they won
Evergreen CPA ≈ 1.4 × test CPA
Winners grow inside testing while they hold the bar. When volume caps, we duplicate them into scale campaigns we create. Your campaigns are never touched.
