$57K from a restructure that started with deleting half the account

The problem

This account had been running five Smart Shopping campaigns, all targeting the same product feed with no segmentation. Every campaign was competing against itself in the auction. The result was inflated CPCs, inconsistent ROAS, and zero control over which products showed for which queries. Legacy Display and YouTube campaigns sat alongside the Shopping mess, spending budget without meaningful attribution.

What we changed

We paused everything and started over. The new architecture has four campaigns, each with a specific job. A primary PMax covers the full catalog. A sales-tier PMax rotates promotional products using custom labels tied to markdown status. A brand-specific PMax isolates premium athletic products into dedicated asset groups with model-level search themes. And a branded Search campaign defends the store's name at sub-$25 CPA. Each PMax campaign uses separate audience signals so they do not overlap.

The results

The sales-tier PMax hit 4.73x ROAS, which is 84% above the Performance Max platform average of 2.57x. The brand-specific PMax spent $314 and returned 4.54x. The branded Search campaign runs at 26.3% CTR, which is 20x the apparel industry average. Blended across the account, the 30-day window produced $57.4K in revenue on $13.9K in spend at 4.13x ROAS. The industry median ROAS fell 10% year over year in 2025. This account moved in the other direction.

$314 on a brand-specific PMax. 11 conversions. 4.54x return. Budget did not make the campaign. Signal precision did.

Performance vs. 2025 benchmarks

Campaign breakdown

Want results like these?