Use automatic targeting for discovery when it is useful, and manual targeting when a query or product opportunity merits dedicated control. Moving a term into manual targeting does not require keeping every automatic campaign active forever or immediately excluding every promoted query.

Give the two roles distinct objectives

Role	What to learn or control	Review measure
Automatic discovery	Relevant queries and product opportunities	Useful evidence gained within the discovery allowance
Manual keyword targeting	Bid and match-type control for supported intent	Mature spend, sales and contribution
Manual product targeting	Specific product or category opportunities	Relevant product traffic and economics

Amazon supports different keyword match types and applicable variations. Inspect the actual search terms rather than assuming a manual exact target receives only a literal spelling. Amazon match-type guidance.

Step 1: export a usable period

Collect search terms or product-target performance with campaign, ad group, clicks, spend, orders and sales. Record the report definition and export time. Refresh the same period after its conversion window and reporting delay; recent zero-order traffic is not automatically mature evidence. Amazon conversion metrics.

Step 2: classify before creating targets

Separate product mismatch, relevant but sparse data, supported opportunity and mature uneconomic traffic. Use product facts for relevance and contribution for economics. There is no universal click count or CVR threshold that makes promotion correct.

Hypothetical decision: a relevant query has £60 spend and £300 attributed sales on mature data, so ACoS is 20%. If the matching product contribution margin is 35%, it may deserve a dedicated test. Check order count, margin scope and current retail conditions before extrapolating; one profitable period does not guarantee another.

Step 3: create a bounded destination

Check for an existing manual target first. Select the product, match type, starting bid and campaign budget deliberately. Record the source term and old settings. Confirm that the destination is eligible and can serve before changing discovery routing.

The following is pseudocode, not executable SQL or an Amazon import schema:

for each reviewed discovery opportunity:
    verify product relevance and mature report
    check existing targets and negatives
    calculate affordable bid from current economics
    prepare a scoped manual target proposal
    after approval, submit and verify provider state
    review routing separately

Step 4: decide whether routing needs a negative

An exact negative in the source can help route a selected query to its destination. It can also remove traffic if the destination lacks budget or eligibility. Use a negative only when that routing is intended and its scope is correct. Phrase exclusions need wider relevance checks.

Keep automatic discovery active when it continues finding useful traffic at an acceptable cost. Reduce or pause it when it does not, or when stock and budget constraints justify prioritizing established traffic. “Never pause Auto” is not a sound universal rule.

Step 5: review the combined result

Compare source and destination together so a transfer of attributed sales is not mistaken for growth. Track contribution after ads, sales volume and discovery quality using comparable mature periods. Retain a comparison group where practical; before-and-after changes can reflect seasonality or retail changes.

How often should the loop run?

Choose a cadence supported by traffic and reporting maturity. Operational checks can be daily while performance decisions wait for useful evidence. See negative design and bid calculation before applying changes.

Related Arctavia resources

Use these supporting pages to compare Amazon PPC operating models and implementation choices.

• Amazon Ad Rejections: Check Claims, Images and Destinations
• Amazon Ads ACoS Suddenly Increased: What to Check First
• Amazon Ads ML Bid Review: Guardrails Before Aggression
• Amazon PPC Automation: Data, Decisions and Operating Controls
• Amazon PPC Automation Migration Checklist
• Amazon PPC Automation vs Agency: Which Operating Model Fits Better?