E-commerce Turnaround
UK online retailer · Google Ads (Search + Shopping)
Inherited a stagnant ecom account stuck at 4-5x ROAS for 13 straight months. Rebuilt the structure in one intervention cycle and tripled conversions on the same spend — ROAS went from 4.59x to 11.98x and held there.
By the numbers
Where we started
When we took the account over it had been running for nearly two years and performance had rotted. The first half of 2024 had hit 11-12x ROAS, but by autumn 2024 it had collapsed to 4-5x and stayed there. Thirteen consecutive months of flatlining at roughly a fifth of its previous efficiency.
The spend wasn't the problem — roughly £10-14k a month was still flowing through the account consistently. The problem was everything downstream of that spend: campaign structure, bidding logic, negative keyword hygiene, and an uncomfortable over-reliance on Performance Max campaigns hiding the underlying query performance.
The client wasn't looking for a 20% improvement. They needed the account restored to the ROAS they used to be able to hit — and this time, held there.
The approach
Forensic audit of the 13 dead months
We started by pulling apart every month of under-performance and comparing against the 2024 peak months when the account had been hitting 11x+. The problem wasn't one single failure — it was an accumulation of structural debt: Performance Max campaigns cannibalising high-intent Shopping queries, a Shopping feed with zero margin-based segmentation, and a negative keyword list that hadn't been touched in nine months.
Broke the Performance Max monopoly
Performance Max is convenient but opaque, and in this account it was eating the budget without surfacing what was actually converting. We rebuilt the core traffic around tightly-structured standard Shopping and Search campaigns, and kept Performance Max for specific tail inventory only. Within two weeks we could see exactly which product groups and search terms were driving the revenue.
Rebuilt the Shopping feed around margin
Custom labels tied to gross margin, tiered ROAS targets by product group, high-margin SKUs isolated into their own campaigns with aggressive bidding. The feed stopped chasing revenue vanity and started optimising for contribution.
Rebuilt the negative keyword list from scratch
We pulled the full search term report, identified the wasted clicks that had been accumulating for nine months, and built a fresh negative list with more than 1,500 terms. Expensive mixed-intent traffic was cut overnight.
Weekly cadence instead of monthly drift
The previous setup was being reviewed monthly at best. We put it on a weekly cadence — search term reviews, budget pacing checks, bid adjustments, ad copy rotation. Compounding small optimisations week-over-week is what holds the ROAS line once it's been restored.
What happened
- October 2025 baseline (last month before intervention): £10.3k spend, 324 orders, £47.3k revenue, 4.59x ROAS, £31.78 cost per order.
- November 2025 (first month of rebuilt account): £10.5k spend, 999 orders, £128k revenue, 12.23x ROAS, £10.51 cost per order. Roughly the same spend, triple the orders, nearly three times the revenue.
- December 2025 – March 2026 (sustained window): ROAS held above 9.2x every month, with three of the four months above 11x. The November spike was not a fluke — it became the new baseline.
- March 2026 (latest full month): £9.6k spend → £115k revenue → 11.98x ROAS, 949 orders at £10.11 each. On slightly lower spend than the old baseline, the account is now doing 2.4x the revenue.
- Across the full 27-month dataset the account has generated nearly £1.94M in revenue. More than half of that came from the five months after we took over.
Want results like these?
Book a free discovery call and we'll run the same diagnostic on your account.
All figures pulled directly from the client's Google Ads account. Client name anonymised pending permission; real data available on request.