Inside the client's 20% ACoS target in four weeks, by cutting waste before adding spend
A wood-repair epoxy brand (about $45K a month in 2026) with a strong B2B base. Search-query data separated market contraction from execution, then a conquest campaign and exact-match rebuild landed ACoS inside the 20% target.
Lower is better. Week 4 settled at 10.4%.
Orders per click across all sponsored campaigns. The exact-match rebuild from search-query data landed in March.
Where it started
Sales dipped 6% and the obvious read was 'ads are broken'. Search-query performance showed branded search volume shrinking 16% while the brand's share of it was actually rising.
What we did
- Pulled the cold-weather formula to a seasonal spend floor and paused five zero-sale campaigns.
- Opened a conquest campaign on the category leader (3% ACoS in week one) and an exact-match campaign built from search-query data (12% ACoS).
- Caught an on-time-delivery flag at 89.9%, a tenth of a point from Amazon's threshold, before it hurt the account.
What happened
ACoS of 16.9%, inside the client's 20% target for the first time in months, with TACoS down to 9.7% and organic sales up 47% on 28% less spend. Spend then grew with sales once the waste was gone.
From the ads console
Jan–Aug 2026 vs Jan–Aug 2025 (spend grew with sales after the reset)
| Metric | Before | After | Change |
|---|---|---|---|
| Account sales | $143,928 | $362,242 | 152% Improved |
| Ad-attributed sales | $72,372 | $245,984 | 240% Improved |
| ACoS | 25.6% | 21.1% | 4.5 pts Improved |
| ROAS | 3.90 | 4.73 | 21% Improved |
| Ad conversion rate | 4.0% | 4.6% | 0.6 pts Improved |
Ad-attributed sales use Amazon's standard attribution windows: 7 days for Sponsored Products, 14 days for Sponsored Brands and Sponsored Display.

