Amazon keyword harvesting starts with the customer-query data in your Amazon Sponsored Products search term report or equivalent connected search-term data.
Keyword vs search term: the distinction matters
A keyword is a targeting instruction you choose. A search term is what the shopper actually typed. Broad, phrase and automatic targeting can match many customer searches. Harvesting is how you turn useful discoveries into controlled targets.
A practical harvesting workflow
Find converting search terms
Review actual customer queries with orders, sales and enough click data to judge performance.
Check whether the term already exists
Duplicate protection matters. Do not create the same exact keyword repeatedly across overlapping destinations without a deliberate reason.
Choose the destination
Map the term to the correct product, campaign, ad group and match type. Exact match is often used when the term has already proved itself.
Set an appropriate bid
Use the source performance, product economics and campaign objective rather than blindly copying the discovery bid.
Control the source
Where your structure calls for it, add a negative exact to the source campaign so the query is routed to the controlled destination.
Review next week
Harvesting is not finished when the row is created. Check whether the new target receives traffic and converts as expected.
What should not be harvested automatically?
- A term with too little data to justify a conclusion.
- A search that converts once but is clearly irrelevant to the product or strategy.
- A duplicate that already exists in the correct destination.
- A term that should be handled as product targeting rather than a keyword.
- A winner whose destination product is out of stock or commercially unsuitable to scale.
Why duplicate protection matters
Without duplicate checks, automated harvesting can create a messy structure: repeated exact keywords, overlapping ad groups and unclear reporting. A good workflow checks current campaign membership and harvest history before adding anything new.
Amazon's own guidance supports the discovery-to-control idea
Amazon recommends using the search term report to identify customer searches and focusing top performers with phrase or exact match. Harvesting turns that recommendation into a repeatable account process.
Where negatives fit
Negative keywords serve two different purposes: removing genuinely irrelevant traffic and routing relevant traffic into a more controlled structure. Those are not the same decision. A converting term may deserve a negative exact in a discovery campaign only after it has been added to the intended manual destination.
Find missed harvesting opportunities automatically.
BoostPPC reviews search-term opportunities, destination mappings, duplicate protection and negative actions as part of the weekly workflow.
Audit my search termsOfficial reference: Amazon Ads: Getting started with sponsored ads.