Key takeaways:
- An auto campaign is a Sponsored Products campaign where Amazon picks the targets from your listing content, not from a keyword list you upload.
- Set the four targeting groups separately: close match, loose match, substitutes, complements. One bid for all four hides what is working.
- Start near the low end of Amazon's suggested bid range, use dynamic bids down only, and leave placement adjustments at zero for the first two weeks.
- Read auto campaigns as a research report, not a profit center. Judge them on the number of converting search terms and ASINs they surface each month.
- Keep discovery to roughly 15–25% of ad spend. Above that, you are paying to re-learn things you already know.
An Amazon auto campaign is a Sponsored Products campaign that uses automatic targeting: Amazon matches your ad to shopper searches and to other product pages based on what is written in your listing. You choose the budget, the default bid and up to four targeting group bids. You do not choose keywords. Its job is discovery, and everything else about the setup follows from that.
What automatic targeting actually does
Amazon reads your title, bullets, description, backend search terms and category, then decides which searches and which ASINs your ad is relevant to. That has two consequences worth planning around.
First, a weak listing produces weak auto data. If your title does not contain the words people type, the campaign will match you to adjacent searches and you will conclude the keyword does not work when the listing was the problem. Fix indexing before you judge the data.
Second, automatic targeting is only available on Sponsored Products. Sponsored Brands and Sponsored Display use their own targeting models. If someone tells you to run an auto Sponsored Brands campaign, they mean category targeting, which is a different thing.
Auto campaigns also need the Buy Box. If you lose it, the ads stop serving and your spend graph will show a flat line that has nothing to do with bids.
The four targeting groups and what each one is for
Amazon splits automatic targeting into four groups. In Campaign Manager you can set a separate bid for each, and you can pause the ones you do not want. Do that. A single default bid across all four is the most common reason an auto campaign looks expensive.
| Targeting group | What it matches | Typical job | Usual bid position |
|---|---|---|---|
| Close match | Searches very close to your main keywords | Finds your core converting terms | Highest of the four |
| Loose match | Broader, looser searches around your product | Finds long-tail and unexpected use cases | Middle |
| Substitutes | Product pages of items like yours | Conquesting, and ASIN discovery | Middle to low |
| Complements | Product pages of items bought alongside yours | Cross-sell placements, often cheap clicks | Lowest, or off |
Close match and loose match return search terms. Substitutes and complements return ASINs, which appear in your search term report as a B0 identifier rather than words. Those ASINs are the input for product targeting campaigns later.
Complements is the group to test last. It works well for accessories, consumables and anything with an obvious partner product. For a standalone item in a crowded category it often spends without converting.
Starting bids and bidding strategy
Use Amazon's suggested bid range in Campaign Manager as the anchor, not a number from a blog. Suggested bids move with the category, the season and your own conversion rate.
A workable starting pattern, as an example on a product with an average click cost of about one dollar: close match at the suggested bid, loose match at roughly 70–80% of it, substitutes at 60–70%, complements at 50% or paused. The point is the order, not the exact percentages. Close match should always be the one you are willing to pay most for.
Three more settings matter at launch:
- Dynamic bids, down only. This lets Amazon reduce your bid when a click looks unlikely to convert. It is the safest setting while you have no conversion history. Up and down can raise bids substantially for top of search placements, so leave it until you know which terms convert. Amazon has changed these limits before, so confirm the current behavior in Seller Central.
- Placement adjustments at zero. Give the campaign two weeks of clean data before you push budget toward top of search or product pages.
- Budget set so it does not cap. A discovery campaign that runs out of money by 11am only ever shows you morning shoppers.
How quickly you move those bids afterward is a separate discipline.
How many auto campaigns to run
One auto campaign per product group is usually enough. Splitting by targeting group into four separate campaigns gives you cleaner budget control, but it multiplies maintenance and most accounts do not need it below a few thousand dollars a month in spend.
Run auto campaigns on a tight set of ASINs. Loading twenty products into one auto campaign means the two best sellers absorb the budget and the other eighteen generate no data at all.
Read this flowchart as text
- New ASIN or new category to research
- Check listing is indexed and holds the Buy Box
- Create auto campaign one ASIN group, own budget
- Set four group bids close highest, complements lowest
- Dynamic bids down only placements at zero
- After 14 days: enough clicks per group?
- Raise group bid or widen listing keywords
- Review search term report terms and ASINs
- Term converted two or more times?
- Promote to exact match in a manual campaign
- Spend above target with no orders?
- Add negative exact or negative ASIN
- Leave running recheck next week
How to read an auto campaign
Pull the search term report in Campaign Manager, filter to the auto campaign, and set a date range of at least 30 days. Shorter windows on low-volume products produce noise.
Read it in this order:
- Terms with orders. These are your candidates for a manual exact match campaign. Two or more orders is a reasonable threshold on most products. One order can be luck.
- Terms with clicks and no orders. Compare spend on each term against your break-even cost per acquisition. Anything well past it with zero orders is a negative candidate.
- ASIN targets. B0 rows that converted go into a product targeting campaign. B0 rows that spent and did not convert become negative product targets.
- Group-level performance. In the campaign's targeting tab, compare the four groups on spend, orders and ACoS. If substitutes is eating half the budget at twice your target ACoS, cut that bid before you touch anything else.
Judge the campaign on output, not on its own ACoS. A healthy auto campaign that feeds five new converting keywords into your manual structure each month is doing its job even at an above-target ACoS, because the profit shows up in the manual campaigns.
Keeping discovery from leaking money
Auto campaigns go stale. Once a term has moved into a manual exact match campaign, the auto campaign will keep bidding on it unless you add it as a negative exact. That is how you end up bidding against yourself and paying more for clicks you were already winning.
Three controls we use on every account:
- A negative exact list in the auto campaign, updated whenever a term is promoted.
- Negative ASIN targets for product pages that have spent without converting.
- A ceiling on discovery as a share of total ad spend, usually 15–25%.
Branded searches are worth checking too. If your own brand name is showing up in the auto search term report, that traffic belongs in a dedicated brand campaign where you control the cost per click.
Frequently asked questions
Should I ever turn auto campaigns off?
Rarely all the way off. Most mature accounts keep a small auto campaign running at a low bid as an early warning system for new search behavior, new competitor ASINs and seasonal language. If it is spending real money and producing no new converting terms for two or three months, cut the bids rather than deleting the campaign, so you keep the history.
How long before an auto campaign gives usable data?
It depends on click volume, not on days. You want a few dozen clicks per targeting group before you draw conclusions. On a product with normal traffic that is often two to four weeks. On a low-volume niche product it can take a quarter, and the honest answer is to widen the date range rather than act early.
Can I use negative keywords in an auto campaign?
Yes. Auto campaigns accept negative exact and negative phrase keywords, plus negative product targets for individual ASINs and brands. Use negative exact for terms you have promoted to manual, and negative phrase sparingly, since a broad phrase block can shut off traffic you have not seen yet.
Do auto campaigns help organic rank?
Indirectly. Sales from any source contribute to a product's relevance and sales velocity for the terms that drove them. Because auto campaigns spread spend across many terms, they are a slow way to rank for a specific keyword. For ranking work, promote the term into a manual exact match campaign where you can control bid, placement and budget on it.