The Challenge
The account had grown the way a lot of accounts grow: by adding more campaigns on top of the ones already running, without ever going back to prune what wasn't working. Spend had doubled inside a year, but revenue hadn't kept pace, and the team could feel margin compressing without being able to point at exactly where it was leaking.
The diagnosis wasn't subtle once we pulled the search terms report. A large share of budget was landing on broad-match terms that had never converted a single sale in the trailing ninety days, sitting alongside a negative keyword list that hadn't been touched in months. Shopping and the newer automated campaign types were competing against the brand's own manual search campaigns for the same queries, driving up internal auction pressure and cost per click for no net gain. On the landing side, most ad groups — regardless of the specific product or search intent behind the click — were sending traffic to the same generic category page, which meant paid clicks were doing the work while the page was actively working against the conversion.
None of this was visible from the top-line ROAS dashboard the team was reviewing weekly; it only surfaced once we went down to the query and landing-page level.
The Strategy
Rebuilding an account that's this tangled is mostly subtraction before it's addition — cutting the spend that was never going to convert, then rebuilding the structure around intent so budget could be allocated deliberately instead of by campaign inertia. In parallel, we treated the landing experience as part of the media plan, not a separate workstream handed off after the fact. Specific workstreams:
- Account restructure by intent — separated branded, non-branded prospecting, and retargeting into distinct campaign groups with their own budgets and bid strategies instead of one shared pool.
- Search term audit and negative harvesting — ran a systematic pass on ninety days of query data to build a negative keyword list that stopped the recurring spend leaks.
- Shopping and PMax feed optimization — cleaned up product titles, attributes, and custom labels so the automated campaigns could actually differentiate high-margin SKUs from clearance stock.
- Landing page rebuild for the top ten ad groups — matched page headline, imagery, and offer to the specific search intent driving each group, rather than one shared category page.
- Attribution and UTM cleanup — fixed tracking gaps that had been quietly under-reporting assisted conversions from Shopping.
- Creative testing cadence for retargeting — rotated offer-led and product-led creative on a fixed schedule instead of letting ad fatigue set in unnoticed.
The Outcome
ROAS moved in two distinct phases: an immediate step up in month one purely from cutting the waste spend identified in the search term audit, followed by a slower, steadier climb as the rebuilt landing pages and restructured campaigns compounded over the following months. That two-phase shape is typical of this kind of engagement — the quick win from cleanup buys the credibility to invest in the slower structural work.
The cost-per-acquisition improvement and the annual revenue figure should be read as the same underlying story told two ways: once spend stopped being wasted on non-converting terms, the remaining budget could be pushed harder against the segments that were actually profitable, which is what drove both numbers at once rather than one at the expense of the other.
These figures are illustrative of the pattern this kind of account rebuild tends to produce, not a specific client's audited results — treat them as directional guidance for scoping a similar programme.
"Same budget, more than double the revenue. The landing page rebuild paid for itself in six weeks."
