PPC restructure: moving a fragmented ad account to a TACoS-driven architecture
Low PPC order volume wasn't the real problem. The real problem was a fragmented ad account where some campaigns at 10% ACoS were starved for impressions and others at 60% ACoS were wasting spend on undifferentiated head terms. We rebuilt the whole architecture around TACoS.
Why “low ACoS” wasn’t the win it looked like
The client’s PPC account had several campaigns reporting 10% ACoS — on paper, elite efficiency. But the same campaigns were running negligible impressions, negligible orders, and negligible contribution to the account’s overall revenue. Those were not good campaigns. They were over-negated campaigns that had efficiency-ed themselves into irrelevance.
At the other end, campaigns targeting high-competition head terms were running 60% ACoS and consuming the majority of the daily budget. Head-term bidding is a legitimate tool — for brand defense, for rank building, for audience research. But it has to be a bounded, purposed line item. It cannot be the ambient state of the account.
The combination was an account that looked busy but wasn’t converting. Optimizing inside any single campaign couldn’t fix it, because the problem wasn’t inside a campaign — it was between them.
What we shipped
Architecture diagnosis, not campaign diagnosis. We mapped spend, impressions, orders, and TACoS contribution by keyword type — brand, branded defensive, head commercial, long-tail commercial, research — rather than by campaign name. That’s when the picture became legible.
Segmentation and budget isolation. High-competition head-term campaigns got isolated into their own campaign group with a bounded budget and an explicit ACoS target reflecting their defensive / research purpose. Long-tail converting campaigns got their own lane, with budget that could no longer be drained by a head-term overrun.
Per-type ACoS targets. Not one ACoS target across the account. Brand-term defense has a different target than long-tail conversion, which has a different target than head-term research. The target reflects the job.
Negative-keyword audit. We removed negatives that had been applied defensively years earlier and were now starving converting campaigns. The over-negation was the silent killer on the “low ACoS, no impressions” campaigns.
TACoS as the operating KPI. The account stopped being optimized for “minimize ACoS per campaign” and started being operated for “maximize total-account revenue given a TACoS budget.” That’s a different job, and it produces different decisions.
The result
The account moved from a fragmented, accreted collection of campaigns to a controllable, scalable architecture. Budget flowed to the keywords driving orders. Head-term spend became a deliberate, bounded line item rather than a black hole. The ground was laid for measured TACoS reduction and revenue growth in the subsequent quarters, without the whack-a-mole behavior that a campaign-by-campaign ACoS chase produces.
Why this generalizes
Most Amazon ad accounts over two years old have this problem. Campaigns accrete. Negatives accrete. ACoS targets get set once and inherited. Nobody owns the architecture because everyone is busy owning individual campaigns. The fix is not more campaign optimization. The fix is a once-per-year architecture review where the question is “what job is each campaign doing, and is that job still worth funding.”