Amazon Sponsored Products Structure: Auto, Manual, Match Types

Last updated October 23, 2026 · First published November 5, 2026 · By Chad Norris

Disclosure: the author runs an Amazon advertising agency (Tide Flow) and builds Amazon software (AdSpark). The structures below come from Amazon’s documentation and general practice and do not depend on any product. See the disclosure.

A Sponsored Products account is a hierarchy: campaigns hold budgets, ad groups hold products and targeting, and targeting is either automatic (Amazon picks the queries and product pages) or manual (you pick keywords or product targets, and for keywords a match type). The short answer on structure: run one automatic campaign to discover queries, run manual campaigns where you control keywords and bids, move proven search terms into exact match, and separate products whose economics differ. Everything else is tuning how much separation you can afford to maintain.

The building blocks

Amazon’s Sponsored Products best-practices guide (read October 2026) says keywords, product targeting and bids apply to every product in an ad group, so it advises grouping closely related products, keeping only same-category ASINs in one campaign, and grouping within a category by price range and conversion performance so products with very different ROAS are not mixed. It also says a single ad group can take up to 1,000 products at a time and suggests running manual campaigns alongside automatic ones.

Automatic targeting: the discovery tool

Automatic targeting lets Amazon match your ad to queries and product pages based on your listing. Amazon’s targeting guide lists four match groups:

You can set one default bid or a bid per group. The practical point is that each group can behave very differently. A common move is to read the four groups separately in the targeting report and lower the bid or pause the groups that spend without selling, rather than judging the automatic campaign as a single number.

Automatic campaigns are also where negatives matter most: you cannot choose keywords there, so negatives are your main way to steer the traffic.

Manual keyword targeting and match types

With manual targeting you choose keywords from recommendations, your own list, or both. Amazon’s guide states a limit of 10 words and 80 characters per keyword, and says keywords are not case-sensitive.

Amazon’s separate keyword targeting guide suggests balancing specificity against reach, grouping keywords by theme or product type, including longtail phrases of three or more words, and starting with 25 to 50 relevant keywords per ad group. It also recommends letting a campaign run for a couple of weeks before reviewing the search term report.

Two cautions on match types. First, exact does not mean “only the typed string”: Amazon says plurals match, and its keyword guide says close variants can match as well. Second, match type names describe what is allowed to match, not what will convert, so a broad keyword is a hypothesis and the search term report is the evidence.

Product targeting

Product targeting is a type of manual targeting that targets specific products, categories, brands or features. Amazon says you can refine category targets by price range, ratings and Prime eligibility, and that ads can appear on the target’s detail page and in search results where that product ranks high. Sellers use it for competitor conversion pages, for their own product pages (cross-sell), and for category-level reach on new items. Keep product-target ad groups separate from keyword ad groups so the report rows are not mixed.

Three structures you will actually see

We use the match-type split when a product has enough budget that each layer gets meaningful clicks, and the lean layout when it does not. Splitting a $10 per day product into five campaigns gives each one too few clicks to learn from. Amazon’s budget guide suggests a minimum of about $10 per campaign, so five campaigns implies roughly $50 a day before the split pays off.

The harvest loop

The funnel that most structures implement is simple:

1. Discover with automatic and broad or phrase campaigns.
2. Harvest terms from the search term report that produced orders at acceptable ACoS. Amazon’s guide says to add high-converting search terms as exact match.
3. Isolate by adding the harvested term as a negative in the source campaign, so the two campaigns do not compete for the same query. This is where our method is stricter than Amazon’s documentation: the isolation step is practice, not an Amazon requirement. Only do it where the source campaign is the weaker performer. Negating a term out of a campaign that converts well, to hand it to a campaign with no history, can lose more than it gains.

Thresholds for “proven” are a judgment call. A rule of thumb in our own skill notes is two or more orders before promoting a term; with few clicks per term in small accounts that is a reasonable floor, but it is a heuristic, not an Amazon figure.

Bids by structure

Amazon’s guidance is to bid highest on exact, lower on phrase and lowest on broad. The reason is intent and control: exact queries are the ones you already know convert. For the ceiling, use the same arithmetic as in the break-even guide: maximum bid is price times target ACoS times conversion rate of that term.

Illustrative item at $40 with a 30% target ACoS:

The conversion rates are assumptions to show the mechanism. Some tools apply fixed multipliers by match type (our own tool uses exact at 1.0, phrase at 0.85 and broad at 0.7 of the base bid); a multiplier is a shortcut for the table above, and measured conversion by term beats it once you have data.

Two other controls sit on top of bids. Amazon offers three bidding strategies: dynamic bids up and down, dynamic bids down only, and fixed bids. Placement adjustments for top of search, rest of search and product pages go up to 900% per Amazon’s guide, and the January 2024 launch announcement added rest of search. Change one lever at a time, otherwise the next report cannot tell you which one worked.

Budgets

Per Amazon’s budget guide, a daily budget is averaged over a calendar month, and a campaign can spend up to 25% above the daily figure on a given day, with the monthly total still averaging out. A campaign that runs out of budget stops showing until midnight. If a campaign with conversions runs out early, raise its budget or move money from campaigns that are not spending. If it spends without converting, fix bidding, targeting and the detail page first, and cut the budget if that does not help.

Mistakes that wreck a structure

• Mixing products with different economics in one ad group. The bid is shared, so the thin-margin item loses money while the strong item looks fine.
• Judging an automatic campaign as one number. Read the four groups separately.
• Harvesting without isolating. Two campaigns bidding on the same query inflate your own CPC.
• Over-splitting a small budget. Five campaigns at $5 a day each learn nothing.
• Reading new-campaign ACoS at day three. Attribution lags and early data is thin; wait for mature date ranges before cutting.
• Setting one target ACoS for every segment. Branded, harvest and discovery campaigns deserve different targets; the ACoS and TACoS guide shows how.

When structure is not the problem

If a well-structured campaign still loses money, the cause is usually upstream: price, reviews, images or stock. Amazon’s own readiness list says to advertise products with the featured offer, competitive price, stock, and at least five reviews and a 3.5-star rating. Fix that before restructuring. For help choosing who manages it, see how to choose an Amazon PPC agency, and for the numbers that tell you whether a structure is working, read the PPC metrics that matter.