Ask a seller why their ACoS is high and they will talk about bids. Bids are rarely the root cause. The root cause is almost always structure: campaigns where match types are mixed, roles are undefined and search term data has nowhere to go. In that environment every optimization is a guess, and every guess costs money.
This is the structure we build on every account we manage. None of it is exotic. The value is in the discipline of running it every week.
Why structure comes before bids
A bid is a decision about what a click is worth. You can only make that decision well if the campaign tells you what the click was for. When an auto campaign, a broad-match keyword and an exact-match keyword all share a budget and an ad group, you cannot see which one produced the order, so you cannot raise what works and cut what does not.
Structure separates the account into units where every search term has one job. Once that exists, bids become simple arithmetic against your target ACoS.
The four campaign roles
Every product family gets four kinds of campaigns. The names matter less than the fact that each one has a single purpose.
| Role | Targeting | Job | What we watch |
|---|---|---|---|
| Research | Auto, broad, category | Discover new converting search terms | Search term conversion rate, spend per new term |
| Ranking | Exact match on priority keywords | Establish and hold organic rank | Rank movement, share of voice, ACoS trend |
| Profit | Exact and phrase on proven converters | Deliver sales at or below target ACoS | ACoS, order volume, bid efficiency |
| Defense | Branded terms, own ASINs | Protect brand searches and detail pages | Competitor impression share, branded CPC |
Research campaigns are allowed to run above target ACoS, because their output is not sales, it is information. Profit campaigns are not. Keeping those two separate is the single biggest structural fix in most accounts.
Match type isolation
Broad, phrase and exact match do different things, and Amazon's own matching has become looser over time. When they share a campaign, broad match will take the impressions and the budget, and the exact keyword you carefully bid on never gets served.
We isolate match types at the campaign level, not just the ad group level, so that budgets, placement multipliers and bidding strategies can be set for each. A phrase-match campaign for a product might run with a lower placement multiplier than the exact-match ranking campaign for the same keyword, because they are doing different work.
The harvesting loop
Structure is only useful if search terms move through it. Every week we pull the search term report and ask three questions of every term with meaningful spend:
- Did it convert at or below target? Move it to an exact-match profit campaign and negate it in research.
- Did it convert but expensively? Move it to phrase or exact with a lower bid, and watch it.
- Did it spend without converting? Negate it, at the level (campaign or ad group) that stops the bleed without killing adjacent terms.
The negation step is where most accounts fail. Without negatives, research campaigns re-buy the same losing terms every week. With them, research spend gets cheaper over time because the pool of unknowns shrinks.
Placement and bidding strategy by role
Top-of-search placement multipliers are powerful and expensive. In a ranking campaign they are the point: you are paying for the position that drives organic rank. In a profit campaign they are usually waste, because you already know the keyword converts and you want the cheapest click that still delivers.
The same logic applies to dynamic bidding. "Down only" is the default for profit and research campaigns. "Up and down" is reserved for ranking campaigns on keywords with proven conversion, where we accept a higher CPC for position.
Where this goes wrong
Three failure modes show up constantly in the accounts we audit:
- Over-segmentation. Hundreds of single-keyword campaigns nobody can maintain. Structure should be as simple as the weekly process can support.
- Ranking campaigns that never end. A campaign that was meant to establish rank keeps running at a loss long after organic sales took over. Ranking campaigns need a review date.
- Research without negatives. The account discovers the same losing terms every month and pays for them again.
What it looks like after 90 days
When the structure is in place and the loop has run for a few months, the account changes character. Research spend falls as a share of total spend. Exact-match profit campaigns carry most of the revenue at a predictable ACoS. Ranking campaigns come and go as launches happen. The search term report gets shorter, because fewer irrelevant terms are being bought.
That is what "optimized" means in practice. Not a lower bid, but an account where every dollar has a job and every result is attributable.
If you want to see how your current structure scores against this, the free audit does exactly that, campaign by campaign. You can also read how we think about ACoS versus TACoS when setting the targets that drive these bids.
Want a structural review of your account?
Get Your Free Amazon Audit
