Every unit of spend came back nine times over.

Nine-times ROAS isn’t a targeting trick, it’s a creative-testing machine over a strict funnel. The plan: run a wide, cheap testing layer to find the angle the audience actually responds to, then scale the winner through a consolidated (CBO) structure. Cheap-to-test, fast-to-kill is the whole philosophy, we’d rather spend a little to learn an angle is dead than spend a lot slowly discovering it. Watch the funnel end to end, outbound CTR into landing-page views into add-to-cart into purchase, and let the weak links tell us exactly where to cut.

• Ran a near-daily testing cadence, new angles, formats (video, carousel, dynamic, catalog) and audiences shipped constantly, judged on outbound CTR and cost per landing-page view before ROAS ever entered the picture.
• Segmented audiences by buyer, adults, kids, newborn, seasonal (winter), so each creative spoke to one shopper, not everyone at once.
• Found the winner: a “linen” product angle that cleared ~3% outbound CTR (well above category norms) and rode it hard.
• Scaled the winners through best-sellers CBOs and folded in Black Friday and seasonal offer waves; anything that didn’t hold its funnel efficiency got cut within days.

The account settled at about 9× ROAS across roughly 2,000 purchases in six months, at a ~3% outbound CTR that sits well above what the category runs. Cost per purchase stayed low because the funnel was clean at every step, clicks that turned into views, views into carts, carts into orders. And because the winning angle kept clearing every step of that funnel, we could scale it with confidence instead of guessing whether more budget would hold the return.

Reliable ROAS is a byproduct of ruthless creative discipline, not a lucky audience. Find the one angle that works, prove it through the whole funnel, and put your money there. Everything else is a test you should be happy to kill.