Sponsored Products automatic targeting can organize delivery into groups such as close match, loose match, substitutes, and complements. These groups describe different ways Amazon may connect a product with a shopper search or product context. Their availability and labels can vary, so confirm the current campaign report before changing bids.
The four targeting groups
Close match is intended to reach searches closely related to the advertised product. Loose match can reach broader related searches. Substitutes can show an ad in contexts involving similar products, while complements relate to products that may be used together.
These labels are useful for organizing analysis, but they are not a guarantee that every query will match your own definition of relevance. Search terms and placements should be reviewed against the listing’s actual attributes and the shopper’s likely task.
- Close match: closer query relationship to the product.
- Loose match: broader related-query discovery.
- Substitutes: similar-product contexts.
- Complements: related-product contexts.
Measure groups separately
If the campaign reports group-level results, compare spend, clicks, orders, and sales by group over a period long enough for attribution to mature. A high click count may reflect reach rather than profitable demand; a group with fewer clicks may produce better contribution after fees and returns.
Product availability, price, reviews, and listing content influence whether a click converts. Before changing targeting, check whether performance differences come from the audience match or from the advertised detail page.
- Keep the advertised product set consistent when testing groups.
- Allow for attributed-sales reporting delays.
- Use margin and inventory context alongside ACoS or ROAS.
Control the test without losing discovery
Separate groups into campaigns or ad groups when the platform settings allow independent budgets and bids. This can make a test easier to interpret, but too many fragments can slow learning and reporting.
Use the data to form a specific hypothesis—for example, whether close-match traffic is more efficient than substitutes—then make one controlled adjustment. Preserve some discovery budget if the search space is still uncertain.
These groups are not necessarily mutually exclusive explanations for a shopper’s intent. A query can look close to one product and also relate to a broader product category. Use the platform’s assigned group as a reporting classification, then review the actual shopper terms and product context before deciding that a group is inherently good or bad.
How did this article land?
Choose one reaction. You can change it anytime.