The brand
whose ads stopped
working.

The challenge

Six months prior, the Meta account was printing money. Then performance fell off a cliff. Every new ad looked like every old ad, and the algorithm had nothing fresh to optimize against. ROAS halved, and the growth team was out of ideas.

The account structure was a mess — campaigns overlapping, audiences competing against each other, and no clear framework for testing creative. They were throwing spaghetti at the wall and wondering why nothing stuck.

What we did

• Implemented a creative testing framework: 20 concepts per month, systematic iteration on winners, controlled holdout groups for measurement.
• Rebuilt the account structure around creative clusters — grouping by concept angle rather than audience, letting Meta's algorithm find the buyers.
• Introduced a UGC + studio content mix that tripled the creative variety without tripling production cost.
• Built a creative scoring system that killed losers at day three and scaled winners within 48 hours of statistical significance.

The outcome

ROAS climbed from 1.8× to 3.1× in 90 days. The creative win rate nearly tripled — from 8% to 22% — because we were testing more ideas faster and killing bad ones earlier. Ad spend efficiency improved 72% on the same budget.

The brand went from dreading Meta's algorithm updates to outperforming them. When iOS changes hit the rest of the category, their account barely flinched.