Worked example
In the Harvest Readiness examples, a product contributes $12 per order before advertising and clicks cost $1.20. Its break-even conversion rate is 10%. A term with 3 clicks and 1 order has a 33.3% observed rate, but little evidence. A term with 40 clicks and 8 orders has a 20% observed rate and substantially more evidence to review.
Neither observation guarantees that a new exact keyword will perform the same way. The auction context, bid, placement mix and competition can change. A separate target is a test with its own budget and settings, rather than a transfer of historical results into a new campaign.
Why it matters for profit
Harvesting can make bidding and review more deliberate when a query has useful conversion evidence. The important question is whether its future traffic can cover the product’s costs at the proposed CPC. A few orders, a high observed rate or an account-wide ACoS target cannot answer that question alone.
Amazon’s search-term report contains terms with at least one ad click; some rows represent product matches rather than typed queries. Check whether each candidate is a keyword or product target, whether an equivalent exact target already exists, and whether a negative would conflict. Creating an exact target does not itself require blocking the source traffic; assess any routing change as a separate decision.
Check it on your numbers
Use Harvest Readiness.
Read the worked lesson: Choose search terms for an exact-match test.
Related terms
Source: Amazon Ads guidance. Platform rules checked October 30, 2026.