Question

Should You Turn Off Loose Match but Keep Close Match for a Book?

Loose and close match serve different discovery roles in automatic book ads. Evaluate each group by search terms, economics, and sample size.

By Publisher Ads · Published · 5 min read

You can turn off loose match while keeping close match, but do it because the loose-match evidence is weak—not because loose match is inherently bad. Amazon's automatic targeting groups explore different kinds of demand. Evaluate each group's traffic, cost, orders, and discovery value separately before shutting off part of the campaign.

Close match and loose match have different jobs

In Sponsored Products automatic targeting, close match aims at shopping terms closely related to the advertised product, while loose match can reach more broadly related terms. Amazon documents automatic and manual targeting in its Sponsored Products targeting guide.

For books, the distinction can be useful. A tightly titled nonfiction book may receive predictable close-match traffic, while loose match uncovers adjacent problems or reader language. A genre title may find unexpected thematic or trope-based demand. Broader discovery also creates more room for mismatch.

Do not treat the group name as a quality score. “Close” can still find unhelpful traffic, and “loose” can find valuable searches the author did not anticipate.

Pull evidence for each group

Use the same account, marketplace, advertised edition, and completed date range. For each automatic targeting group, record:

  • Impressions
  • Clicks
  • Spend
  • Attributed orders and sales
  • Average CPC
  • Search-term or matched-product detail
  • Current bid
  • Any relevant changes during the period

Keep the group totals separate. A profitable close-match group can conceal a weak loose-match group when only the campaign total is reviewed. The reverse can also happen.

Search-term reports need careful reading. Some rows are shopper text queries, while automatic or category traffic can also surface matched ASINs. Amazon's book advertising reporting guide describes the reports used to examine this activity. Text rows can suggest keyword decisions; ASIN rows can suggest product-target decisions.

Decide whether the issue is relevance or price

Read the highest-spend loose-match rows and classify them:

  1. Relevant and converting: possible candidates for controlled manual targeting.
  2. Relevant but uneconomic: consider the bid, sample size, and book-page conversion.
  3. Clearly irrelevant: consider a carefully scoped negative where supported.
  4. Ambiguous: gather more evidence within an affordable limit.

Suppose, hypothetically, loose match finds “decluttering workbook” for a general home-organization guide. That may represent useful adjacent intent. If it finds repeated traffic for industrial warehouse storage, the mismatch is clearer. The label “loose” did not make either decision; the reader intent did.

If traffic is relevant but costly, lowering the group's bid may preserve discovery at a more affordable price. Turning it off removes both poor and promising opportunities. If traffic is broadly irrelevant across meaningful evidence, pausing the group is cleaner.

Use book economics to set the test boundary

There is no universal click threshold or test duration. Use actual royalty contribution for the advertised format and marketplace to decide how much uncertainty you can afford.

A simple hypothetical framework is to set a maximum learning budget for the group, then review when it reaches either that spend or a chosen evidence interval. If a book contributes $3.50 per order, allowing $70 of unexamined discovery spend carries a very different risk from allowing $7. Neither number is a rule; the appropriate amount depends on cash flow, conversion evidence, and the strategic value of discovery.

Do not judge using attributed retail sales alone. ACoS compares spend with attributed sales value, while the publisher retains a royalty. The low-ACoS profit guide explains the difference.

Also allow recent reporting to settle rather than assuming one fixed attribution window. Record the extraction date and rerun the same date range if the newest activity could still change.

Keep close match only when it earns its place

The question often assumes close match should remain enabled. Test that assumption too. Close-match traffic may be more semantically obvious yet still expensive, poorly converting, or redundant with another campaign.

Compare both groups on:

Question Why it matters
Is the traffic relevant? A close label does not guarantee reader fit.
Is CPC affordable? Relevant clicks can still cost too much.
Does the book page convert? Targeting cannot repair a weak or misleading offer.
Does the group find new terms? Discovery can justify a bounded learning cost.
Is another campaign serving the same role? Overlap can obscure where results come from.

Keeping close match and pausing loose match is sensible only when this comparison supports it.

Choose among four measured actions

You have more options than “leave everything on” or “kill the campaign”:

  • Keep both unchanged when samples are small and the test remains affordable.
  • Lower the loose-match bid when traffic has value but auction cost is the main concern.
  • Add precise negatives when specific irrelevant queries recur while the group still finds useful demand.
  • Pause loose match when broad evidence shows the group itself no longer merits spend.

Make one primary change, note the date and reason, and compare a completed period afterward. Avoid simultaneously changing bids, negatives, budgets, and the detail page; the result becomes impossible to attribute.

The pause-target-versus-negative comparison helps choose between stopping a route and excluding only one branch of its traffic.

Preserve what the discovery campaign learned

Before pausing loose match, save the useful search terms and products it found. A strong text query may deserve a manual keyword test. A relevant matched ASIN may deserve a product-target test. Promotion into a manual campaign is a new test, not proof that the term will repeat its earlier result.

The search-term harvesting workflow explains how to move candidates without treating every row as a keyword.

Publisher Ads lets publishers manage supported campaigns and targets across books. Whether you work there or directly in Amazon's console, confirm the exact campaign and automatic targeting group before changing it. Similar campaign names make broad changes easy to misapply.

Turn off loose match when its own evidence says the discovery route is no longer worth funding. Keep close match only when its own evidence also supports the role. Automatic targeting works best when each group has a purpose, an affordable test boundary, and a review record.

Questions people ask

What is the difference between close match and loose match?

They are separate automatic targeting groups. Close match seeks shopping terms closely related to the book, while loose match explores terms related more broadly.

Should every author turn off loose match?

No. Review the group's own search terms, spend, orders, and role. Loose match may find useful demand for one book and weak traffic for another.

Can I change one automatic targeting group without stopping the campaign?

Yes, supported campaign controls let advertisers manage automatic targeting groups individually, subject to the current campaign setup.

How long should I test loose match?

There is no universal duration. Use an affordable evidence budget, enough opportunity for the book's demand, and settled reporting rather than a fixed number of days.

More from the blog