How to Govern Amazon Ads Product Targeting

Write the advertised ASIN, marketplace, campaign objective, eligible product family, and decision window before opening the target picker. Decide whether the operator is trying to appear around close substitutes, complementary products, a broader category, or the brand's own catalog. Those are different hypotheses and should not be blended into one target list. A written purpose makes later bid changes and exclusions explainable instead of turning the campaign into a collection of untraceable additions.

Amazon describes product targeting as a manual option for selecting products, categories, brands, or other relevant product features. It is different from keyword targeting, which begins with words and phrases. Preserve that distinction in naming, reporting, and reviews. If one ad group combines several hypotheses, split the next canary rather than trying to assign one result to several possible causes. The operating record should say exactly what relationship each target is meant to test.

Start with a small list that a reviewer can inspect. For individual products, record the ASIN, relationship to the advertised product, observed source, and reason for inclusion. For categories, record the category and every refinement applied. Amazon states that categories can be refined by attributes such as brand, price range, review star rating, and Prime shipping eligibility in supported campaign contexts. Capture the interface values because a label like relevant category is not enough evidence for another operator to reproduce the setup.

Use account evidence as a discovery input, not an automatic instruction. Amazon's targeting guide points operators to auto-campaign metrics, suggested targets, and category targeting as starting paths. A suggested item can enter the review queue, but it should not bypass product fit, availability, margin, policy, or objective checks. Mark whether the seed came from an automatic campaign, a report, a suggested list, a catalog relationship, or a manual hypothesis so later comparisons do not mix discovery methods.

Group advertised products only when they share a useful decision context. A target and bid applied across products with different price, availability, margin, or catalog roles can hide which product actually justified the change. Use campaign and ad-group names that expose marketplace, product family, target type, relationship, and launch date. Keep category targets, exact ASIN targets, and expanded targets distinguishable so an operator can export the account and understand the logic without relying on memory.

Do not copy a winning target into every campaign on sight. Confirm the advertised product, placement context, landing detail page, inventory state, and intended role first. When reuse is appropriate, create a controlled copy with its own baseline and owner. The purpose of structure is not cosmetic tidiness. It is to preserve a stable unit of analysis, make rollback possible, and prevent a target discovered for one product from becoming an unsupported account-wide rule.

Amazon explains that product targeting can make an ad eligible on relevant product detail pages and on search-result pages where the targeted product appears. That means a target name alone does not describe every context in which impressions may occur. Review the supported reports and account views for the chosen date window. Keep impressions, clicks, attributed orders, spend, bid, placement, advertised product, and target identity together, and label missing fields instead of filling them from another period.

Use the search term report to inspect ASIN-like customer search terms discovered by automatic or product-targeted activity, then compare those discoveries with the targeting report. Promotion to a manual target should be a documented decision with a reason and a date, not a silent copy action. Likewise, low activity is not proof that a relationship is wrong. Check eligibility, inventory, bid, time in market, and reporting completeness before deciding whether the evidence is sufficient to keep, change, or remove a target.

Amazon's current support describes expanded product targeting as a way to reach additional products related to a selected individual target. Keep that option labeled separately from exact product targeting. Use a bounded budget, fixed review window, and named rollback condition. If expanded and exact eligibility run together without a separate record, the operator cannot tell whether the original ASIN relationship or the broader related set produced the observed activity.

Apply negative product or brand targets only after checking the same context used to qualify positive targets. Record the excluded object, evidence window, reason, affected campaign, and reviewer. Do not infer poor fit from one click, from a different marketplace, or from a product that was out of stock. Exclusions can prevent future eligibility, so treat them as production configuration. Review whether the exclusion still matches catalog and objective changes before carrying it into a new season or product launch.

Use one review sheet with the target expression, type, refinements, advertised product, bid, status, source, start date, evidence window, decision, owner, and next review. Reconcile the sheet with the advertising console before analysis so deleted, archived, paused, or newly added targets are not treated as comparable active rows. Separate discovery, qualification, activation, observation, refinement, exclusion, and retirement. Each state needs an explicit entry and a reason that another operator can audit.

At the close of each review, keep, adjust, expand, isolate, or retire each target individually. Compare the next complete period with the recorded baseline and state any reporting limitation. Do not turn correlation into a promise of causal impact. The durable output is a target map that connects every eligibility decision to product fit, official platform behavior, account evidence, and a named owner. That makes product targeting manageable even as the catalog and campaign portfolio grow.