Case study · Paid Performance Media
Breaking every shopping-campaign rule lifted account ROAS 59%.
A UK-based plus-size fashion e-commerce retailer focused on womenswear apparel.
The situation
The client is a UK-based plus-size fashion e-commerce retailer focused on womenswear apparel. Their Meta account ran on the category standard: ROAS-bidding shopping campaigns across narrower, warm-leaning audiences, optimised to protect return on ad spend. The brand wanted to grow new-customer acquisition meaningfully, but had no appetite or budget to fund a separate test alongside existing spend. Any new approach would have to be funded from within the existing account budget, not on top of it, putting pressure on whichever campaigns lost spend to make room.
The assumption
Conventional wisdom, and the account's own history, said ROAS bidding on curated, narrower audiences was the only safe way to grow an e-commerce Meta account without wrecking efficiency. Broadening targeting and bidding to cost-per-acquisition was assumed to dilute return, waste spend on unlikely buyers, and drag the account's overall ROAS down.
What the evidence showed
Rather than requesting new budget, roughly half of total account spend was reallocated into a single campaign built against every one of those assumptions: broad targeting, cost-per-acquisition bidding, and existing customers excluded entirely. Measured as a year-on-year, account-level comparison across the same window (the structure hadn't existed the year before), the results moved the opposite way to what the standard approach predicted: blended ROAS up, cost per acquisition down, new customers up, on essentially flat total spend.
What we did
- i.Identified that the new-customer goal could be funded from within the existing budget rather than through a separate test spend.
- ii.Built a new campaign on cost-per-acquisition bidding instead of ROAS bidding, with targeting opened up to fully broad: no interest layers, no lookalikes, and a customer list uploaded to exclude anyone who had already purchased.
- iii.Shifted roughly half of total Meta spend into the new campaign, taking the corresponding reduction from the existing ROAS-bidding campaigns rather than adding net-new budget, so any change in performance could be attributed to the structural shift rather than more media dollars.
- iv.Ran the approach as a 90-day pilot, judged at the account level across the full comparison window, not on the test campaign's own reported numbers in isolation.
Results
Account ROAS
150 days YoY, 10 Feb–9 Jul, Meta account-level
New customers
Same 150-day window YoY, Meta account-level
Cost per acquisition
Same 150-day window YoY, Meta account-level
The evidence, before and after
Same account. Different rules.
Whole-account metrics, indexed · 2025 window = 100 · 10 Feb–9 Jul both years, Meta account-level
These are whole-account numbers, not the test campaign's own reported figures: the account grew more efficient and acquired more new customers on essentially flat total spend, with roughly half the budget moved into the new structure. Measured as a year-on-year, account-level comparison across the same calendar window in both years. The 90-day pilot has since become the primary focus of the account, and budget continues to scale into it.
“These guys don't waste budget.”
The lesson
Real growth doesn't hide behind best practices. It hides in the exceptions nobody's willing to try.