Most Amazon PPC strategies are written for sellers of blenders and phone cases. Books behave differently: a low price, a royalty instead of a margin, series and page reads. This plan is built for books, and each step is checked against the numbers from 254,473 clicks on a real KDP catalogue.
The principle: find, promote, prune
In that catalogue, the top 5% of keywords and targets produced 82% of all orders. So the job of a book ad strategy is not to make every keyword work. It is to find the few that do as cheaply as possible, give them their own bids, and stop paying for the rest.
- Find with automatic and broad campaigns, which show your book on searches you wouldn’t have thought of.
- Promote the searches that sell into exact-match campaigns, where they get their own bid.
- Prune the searches that spend without selling, after a fair test, with negative keywords.
Campaign structure
| Campaign | Targeting | Job |
|---|---|---|
| Auto | Amazon chooses searches and products | Discovery: finds searches and books you didn’t think of |
| Broad / phrase | Your keywords, loosely matched | Discovery around themes you chose |
| Exact | Searches that have sold | Scale: each proven search gets its own bid |
| Product targeting | Other books’ pages (ASINs) | Comparable titles readers also buy |
One book per campaign keeps the numbers readable: when a campaign holds several books, you can’t tell which one a sale came from without opening the reports. When you promote a search into the exact campaign, add it as a negative exact in the campaign it came from, so the two don’t compete for the same search.
Match types, and what they did
Amazon’s definitions: broad can match the words in any order, with plurals, variations, synonyms and related terms — even searches that don’t contain the keyword. Phrase needs all the words in the same order, with other words allowed before or after. Exact needs the same words in the same order, plus the plural and very close variants. Negatives come in two kinds only: negative phrase and negative exact.
| Match type | Click-through | Conversion | ACOS | Cost per click |
|---|---|---|---|---|
| Broad | 1.21% | 12.4% | 55% | $1.07 |
| Phrase | 1.19% | 16.0% | 48% | $1.18 |
| Exact | 0.88% | 13.0% | 57% | $1.16 |
| Product targeting | 0.90% | 12.1% | 45% | $0.85 |
One publisher’s US account, April 2025 to September 2026. Exact match ran higher than you might expect partly because this account promotes searches after their first order, which brings in more unproven keywords.
Two lessons. Exact match is not automatically the most profitable: here phrase did better. And product targeting had the lowest ACOS of all thanks to cheap clicks, which matters more for books than for most products.
Placements and bid adjustments
Sponsored Products ads appear in three placements: top of search, rest of search and product pages. You can raise your bid for each one by up to 900%.
| Placement | Share of spend | Click-through | Conversion | ACOS |
|---|---|---|---|---|
| Top of search | 13% | 4.51% | 14.0% | 52% |
| Rest of search and other | 65% | 1.50% | 16.2% | 45% |
| Product pages | 21% | 0.45% | 5.0% | 95% |
Product pages took a fifth of the money and converted at a third of the rate. If your placement report looks like this, keep the base bid near what a product-page click is worth and use the search placement adjustments to bid up where readers are actually searching.
Bids
royalty per sale × conversion rate
Bid at or below it on proven keywords; a little above it, for a while, on keywords you are still testing.
On bidding strategy: fixed uses your bid as it is; down only lets Amazon lower it when a sale looks less likely; up and down lets Amazon raise it too, by up to 100%. For a book with a thin royalty, down only or fixed keeps the cost per click under control. Free toolAmazon PPC bid calculatorA starting bid from your price, royalty and conversion rate.
Promoting searches that sell
our study of 2,846 keywords changed how we promote: one order is enough (keywords from one, two and three-order searches performed the same), and the search’s ACOS before promotion says almost nothing about the keyword’s ACOS after. Expect a promoted keyword to run at a higher ACOS than its search did, about 57% from one to three orders, and give it 10–15 clicks before you judge it.
AcosControl · Start freeEvery search that sold, one click from its own campaignAcosControl lists the search terms with orders that aren’t a keyword yet, and creates the exact-match campaign for you.A weekly routine
- Leave the last three days out of every decision: sales are credited for 14 days after a click, and recent days are still arriving.
- Promote search terms with an order that aren’t a keyword yet.
- Negate search terms that have had a fair test (about 3 ÷ conversion rate clicks) with no order.
- Decide on tested keywords running above break-even: keep on purpose, or pause. A bid cut rarely rescues them.
- Check placements once a month, per campaign.
- Leave untested keywords alone. Don’t pause a campaign of new keywords after two weeks.
Five common mistakes
- Judging too early: 10 clicks with no sale happens by luck a third of the time at 10% conversion.
- Pausing whole campaigns: in our data, 369 keywords from searches that had sold never got a click because their campaign was paused within weeks.
- Trimming instead of deciding: bid cuts moved high-ACOS keywords from 64% to 60%, not under target.
- Raising bids on keywords nobody clicks: most were already bidding above what their search paid when it converted.
- Copying someone else’s ACOS target: yours is your royalty ÷ price. Break-even ACOS.
Common questions
What is the best Amazon PPC strategy for books?
Find, promote, prune: use automatic and broad campaigns to find searches that sell, promote them into exact-match campaigns after their first order, and negate searches that spend without selling after a fair test. In one large book account, 5% of keywords produced 82% of orders.
How should I structure Amazon PPC campaigns for a book?
One book per campaign, with separate campaigns for automatic, broad or phrase discovery, exact-match keywords that have sold, and product targeting. Add each promoted search as a negative exact where it came from.
Which match type works best for book ads?
It varies by account. In one publisher’s 254,473 clicks, phrase match converted best (16.0%, 48% ACOS), with broad at 12.4% and exact at 13.0%. Compare yours before assuming exact is best.
Should I use dynamic bidding up and down for books?
Up and down lets Amazon raise your bid by up to 100%, which can push a thin-royalty book past break-even. Down only or fixed bids are the safer default; test up and down on proven keywords.
How often should I optimise Amazon PPC campaigns?
Weekly is enough for most authors: promote searches with orders, negate tested ones without, and leave the last three days out because sales are still being credited.
- Book ads are a search for the few keywords that work: 5% made 82% of orders in one catalogue.
- Find with auto and broad, promote to exact after the first order, prune after a fair test.
- Phrase match beat exact in this account; product pages converted at a third of search.
- Bid from break-even (royalty × conversion rate); prefer down-only or fixed on thin royalties.
- Review weekly, ignore the last three days, and never pause campaigns of untested keywords.