Most Amazon accounts we inherit are not underperforming because the bids are wrong. They are underperforming because money leaves through structural gaps that have no dashboard flags. ACoS looks acceptable, revenue is growing, and a meaningful share of the budget is funding searches that were never going to convert. That gap is architectural, and it compounds quietly.
Why Wasted Ad Spend Almost Never Shows Up in ACoS
Wasted Amazon ad spend hides inside a healthy ACoS because the metric averages profitable and unprofitable spend together. A campaign converting strongly on brand terms can mask an entire ad group burning budget on irrelevant queries. The leak becomes visible only at the search-term and placement level, which is exactly where most accounts are never examined. Structure, not bidding, is where the money goes.
We see this pattern constantly. An account reports a blended ACoS in a range the brand is comfortable with, and underneath it the gap between the best-performing ad group and the worst is not marginal – it is several multiples. Averages are comfortable. They are also the reason a problem can run for eighteen months without anyone noticing.
The instinct at that point is to lower bids across the board. It works for a month. Impressions fall, spend falls, ACoS improves, and the brand loses ranking on the terms that were actually profitable, because the reduction hit those equally. Six weeks later the account is smaller and no healthier. In the accounts we have taken over, bid adjustment alone has almost never been the fix.
Mixed Match Types and the Negative Keyword Vacuum
The most common structural fault is broad, phrase, and exact match living inside one campaign. It feels efficient. It makes the account impossible to read. When a broad match term and its exact match twin sit in the same ad group, they compete in the same auction, drive each other’s CPC upward, and produce data that cannot be attributed to either decision. We cannot optimize what we cannot separate.
Underneath that sits the second leak, the negative keyword vacuum. The consequence here is never dramatic, which is precisely why it survives. A handful of irrelevant search terms absorb spend every day, none of them large enough to appear in a weekly revenue report, all of them compounding across a quarter. By the time the pattern is visible in a profit and loss statement, it has been running for a year.
Negative keyword governance is unglamorous work. It is also the highest-return hour anyone can spend inside an Amazon ad account, and it is the first thing that lapses when a brand gets busy.
Placement Data Is the Most Ignored Report in Seller Central
The third leak sits in placement. Ask a brand where their conversions come from and most can answer by campaign. Very few can answer by placement, and that is a mistake, because top-of-search, rest-of-search, and product pages behave like three different channels with three different economics.
Top-of-search converts at a materially higher rate and costs materially more. Whether that trade is worth making depends on the product’s margin and its stage in the ranking cycle, and it is a decision that has to be made per campaign rather than per account. Applying one placement modifier across an entire account repeats the same error as applying one bid.
Product page placements are the ones most often left running unexamined. They can be excellent for defensive targeting against a competitor’s listing and dismal for generic discovery, and the difference is invisible unless someone reads the report at that level.
Budget Distribution Is a Strategy Decision, Not an Accounting One
The fourth leak is the budget allocated by history rather than by intent. The campaign that spent most last quarter gets most this quarter, and nobody asks whether it is still doing the job the business currently needs done.
A launch campaign and a defensive brand campaign should not be funded on the same logic. One is buying ranking velocity at a deliberate short-term loss. The other is protecting margin on demand that already exists. When both sit in the same budget pool under the same target ACoS, the account quietly starves whichever is currently more expensive – usually the launch – and the brand concludes the product failed when the truth is it was never properly funded.
The fifth leak follows directly from this. A campaign that exhausts its budget by early afternoon is not underperforming; it is being switched off during the hours its buyers are active. It will report a respectable ACoS while doing so, because the spend it never made cannot appear as waste. Budget caps are the only failure mode in advertising that makes the metrics look better as the problem gets worse.
How We Read an Account Before Touching a Single Bid
When a new account comes to Sellers Umbrella, we change nothing for the first two weeks. We map it. Serious Amazon PPC management starts with knowing which campaigns carry the brand’s actual profit and which carry its volume, because these two sets are rarely the same and they require opposite handling. Our Revenue Core framework exists to separate them. Only once that separation is clear does bidding become a sensible conversation.
The Keyword Evolution Path governs what happens next. Terms are not permanently broad or permanently exact. They move – from discovery to validation to defense – and each stage justifies a different match type, a different budget, and a different tolerance for ACoS. Most accounts we audit have terms frozen at the stage they entered years ago. The Performance Amplifier then handles the terms that have earned it: concentrated budget, placement modifiers set per campaign rather than per account, and a defined ceiling beyond which we stop.
None of this is complicated. It is work that requires someone to look at an account at the query level, consistently, which is exactly the thing that stops happening once a brand grows enough to matter.
Wasted spend is not a bidding problem, and it will not be solved by another round of optimization applied on top of a broken structure. More budget poured into a poorly built account does not buy growth. It buys a faster leak. If the architecture has not been examined at the search-term and placement level in the last two quarters, the honest assumption is that it has drifted. We would rather show a brand exactly where their money is going before anyone proposes spending more of it.
Author bio: Nirav Bhatt is the founder of Sellers Umbrella, an Amazon consulting agency that has managed over $200M in client revenue and scaled more than 100 brands across seven international markets. He writes about marketplace profitability, advertising architecture, and the operational side of scaling on Amazon. sellersumbrella.com
