How to Choose KDP Categories: A Step-by-Step Method
A repeatable method for choosing your three KDP categories by relevance, buyer fit and competition, with a worked example and a ve…
The Framework in One View
A workable KDP ads strategy has five moves, in this order: calculate what one sale is worth to you, decide how much you can lose while testing, launch a discovery campaign, launch a controlled campaign for the terms that prove themselves, and review on a fixed schedule with written rules. The rest of Ads Manager is detail on top of those five.
The order matters because the console shows bids, budgets and suggested keywords before it asks anything about profit. A new publisher sees a suggested bid and accepts it without knowing whether the book can afford that click.
Every threshold below is scaled to one book's royalty, so rescale each one to your own numbers before you copy it. This framework treats break-even on the direct sale as a deliberately conservative floor; the scaling section gives a written formula, with a loss cap, for the cases where you knowingly go above it.
What KDP Ads Are and Where Ads Manager Fits
Ads for KDP titles run on Amazon's own advertising platform. You start from your KDP account and are handed to the Amazon Ads console, which most publishers call Ads Manager.
Two ad formats matter for most publishers. Sponsored Products put a single book into search results and onto product pages, and they are the format to learn first. Sponsored Brands, according to Amazon's guide for authors, show collections of 3 to 10 books or a video, and require three or more eligible titles claimed under the same author name (as of September 2026, check Amazon's Sponsored Brands guide for authors for the current wording).
A publisher with one book has one realistic option; a series gets a second format to test later. Ads run only for titles that are available for purchase and meet Amazon's advertising content policies, and ad types vary by marketplace and account type, so confirm what your account offers on the day you set up.
Step Zero: Calculate Break-Even ACOS Before You Spend
ACOS is ad spend divided by ad-attributed sales. It is the single number that tells you whether a click cost is survivable. The version that matters to a publisher is break-even ACOS: your royalty per sale divided by the list price.
Take a 6 x 9 in large-print puzzle book listed at $9.99. As of September 2026, the paperback royalty on Amazon marketplaces is 60% of the list price minus the printing cost (40% through Expanded Distribution); check KDP's royalty help page before you price, because Amazon has changed the low-price bands. Here, 60% of $9.99 is $5.99.
Assume a printing cost of $3.59 for a black-and-white interior on regular trim (an example figure; your page count sets the real one). The royalty is then $2.40, and break-even ACOS is 2.40 divided by 9.99, or 24.0%.
Look up your own page count on KDP's printing-cost page, then redo the two lines above with your numbers and whatever royalty rate is current.
Any campaign or keyword running above 24.0% ACOS loses money on the direct sale. Below it, you earn something. Break-even also caps your cost per click, once you know your conversion rate.
| Item | Formula | Worked value |
|---|---|---|
| Royalty per sale | 60% of list price minus printing cost | 5.99 - 3.59 = $2.40 |
| Break-even ACOS | Royalty / list price | 2.40 / 9.99 = 24.0% |
| Break-even CPC at 10% conversion | Royalty x conversion rate | 2.40 x 0.10 = $0.24 |
| Break-even CPC at 8% conversion | Royalty x conversion rate | 2.40 x 0.08 = $0.19 |
| Clicks before a pause decision | (2 x royalty) / your average CPC | 4.80 / 0.20 = 24 clicks |
Conversion rate is the unknown in that table. Nobody can give you an honest figure for your book before it has traffic, so treat 8% to 10% as a starting assumption to test, not a benchmark. The free KDP calculators handle the royalty and printing-cost side, including the royalty step for paperbacks on regular trim.
Then run the honesty check. If the suggested bids on your best terms sit well above break-even CPC, the book cannot carry ads at its current price. With a $2.40 royalty, the ceiling is $0.24 at 10% conversion and only $0.12 at 5%. If bids that low win no impressions, that is the signal to fix price, page count or listing rather than raise bids.
Prepare the Listing Before Any Traffic Arrives
Ads buy clicks; they do not fix a listing that fails to convert them. A weak cover, vague title or thin description shows up as an ACOS that never gets under break-even.
Run through this short list before launch:
- The cover reads at thumbnail size, because search results show it small.
- The title and subtitle name the format (large print, activity book, workbook) and the audience.
- The description opens with what the buyer gets, with the page count and trim size stated.
- The price sits within the range of the top results for your search term, not far above it.
For puzzle and activity books, interior quality is a conversion factor. If you assemble interiors from page images, KDP interior template layouts show how the pieces fit together, and KDP Creator produces numbered 300-DPI page images plus a manifest that you then assemble into your own interior file. It does not export a finished PDF.
Structure Campaigns by Targeting Type
Amazon lets you put many targeting methods in one campaign. Do not. Mixed campaigns produce one blended ACOS that hides which method earns and which burns. Keep each targeting type in its own campaign so the numbers stay readable.
| Campaign | Targeting | Job | Bid approach |
|---|---|---|---|
| Auto | Amazon picks terms and products (close match, loose match, substitutes, complements) | Discovery: shows which search terms and competing ASINs produce clicks and orders | Every bucket at or below break-even CPC (about $0.20 in the running example), even if that is under the suggestion; cut each bucket that fails |
| Manual, broad and phrase | Keywords you supply, broad or phrase match | Test seed terms that variations of buyer language might reach | Lower than exact for the same term |
| Manual, exact | Terms proven in auto or broad | Control: full attention on terms that already convert | Highest bid of the three, raised only after orders arrive, and never above that term's own break-even CPC |
| Product targeting | Specific competing ASINs or categories | Place your book on similar books' pages | Start modest; judge by orders, not clicks |
Match types follow Amazon's definitions. Broad match lets Amazon show your ad for variations, synonyms and related terms in any order. Phrase match requires the keyword phrase in the same order with other words allowed before or after. Exact match is the tightest and allows close variants such as plurals and misspellings (as of September 2026, see the Amazon Ads keyword targeting guide for the current definitions).
For a first campaign, an auto campaign plus a manual broad or phrase campaign is enough. Add exact match and product targeting only once you have orders to learn from. Four campaigns on day one splits a small budget so thinly that none of them collects usable data.
Step by Step: Setting Up the First Two Campaigns in Ads Manager
The steps below assume a Sponsored Products campaign for one paperback. Button labels can differ slightly from what you see; the logic holds.
1
Create the auto campaign
Choose Sponsored Products, select the one book, and pick automatic targeting. Name it clearly, for example "LargePrintWordSearch - Auto - Test". Set the daily budget from the budget section below.
2
Pick a bidding strategy
Amazon offers dynamic bids that only go down, dynamic bids that go up and down, and fixed bids (as of September 2026, check Amazon Ads help for current names). Up-and-down can pay more than your bid for a click, which hurts when break-even CPC is $0.24. Use down-only or fixed while learning.
3
Set bids per bucket
Auto targeting has separate bids for close match, loose match, substitutes and complements. Compare the suggested bid on screen with your break-even CPC, then set every bucket at or below break-even (about $0.15 to $0.20 in the running example) even when that sits under the suggestion. Lower bids win fewer impressions, so expect slower data; if they win nothing at all, fix the price or listing instead of raising bids.
4
Create the manual campaign
Start a second Sponsored Products campaign for the same book, choose manual targeting, and select keyword targeting. Paste 15 to 30 terms you wrote from buyer language, with broad or phrase match.
5
Add starter negatives
Add negative keywords for terms that share words with your niche but mean something else, such as "free", "pdf" or a different puzzle type. Use exact negative match first.
6
Save, then leave it alone
Submit both campaigns and note the start date. Do not edit bids for the first several days. Orders are credited to an ad only within a fixed window after the click, and reports lag, so early zero-order counts are unreliable. Write the break-even ACOS, the pause threshold and the review date in a note now, because rules written after seeing data bend toward hope.
Budgets and Bids: What You Can Afford to Lose
Amazon recommends a minimum daily budget of about $10 per campaign for Sponsored Products (as of September 2026, see the Sponsored Products budget best-practices page). The console will accept lower amounts, but very low budgets can stall before the day ends and produce almost no data.
Daily budgets are averaged over a calendar month. Amazon states (as of September 2026, see the Sponsored Products budget best-practices page) that you may spend up to 25% more than the daily figure on any one day, and the month balances out to your average. So a $10 daily budget can spend $12.50 on a busy day and totals about $300 across a 30-day month.
Check the current overspend allowance before you rely on the "highest single day" column below.
| Daily budget | Monthly total (30 days) | Highest single day | Clicks per day at $0.20 CPC |
|---|---|---|---|
| $5 | $150 | $6.25 | 25 |
| $10 | $300 | $12.50 | 50 |
| $20 | $600 | $25.00 | 100 |
The last column is the useful one. Budget only produces conclusions when it buys enough clicks. At an assumed 8% conversion rate, 25 clicks a day averages two orders a day, which is thin but readable over a week. A budget that buys five clicks a day tells you nothing for a month.
Decide the test cap in advance: the total you are prepared to lose to learn whether the book can be advertised profitably, split about evenly across the two campaigns. If it is gone with no orders and the listing is sound, that is an answer, and a cheaper one than another quarter of ads.
Bids follow the break-even CPC you calculated. If the break-even is $0.24, a $0.60 bid can only work if conversion is far higher than your assumption. Bid at or below break-even for discovery, and raise a bid only for a term that has produced orders, and then only up to that term's own break-even (royalty times its measured conversion rate).
Reading Results and Acting by Rule
Open the search term report after the test window. It lists what shoppers actually typed and which terms clicked and ordered, and it improves the strategy more than the campaign summary does.
Use fixed rules instead of gut feel. The thresholds in this table use the running example (royalty $2.40, break-even ACOS 24.0%, average CPC $0.20). Apply them only to clicks old enough that their orders have landed. As of September 2026, confirm the attribution window for book ads on Amazon Ads help; for Sponsored Products it is commonly reported as 7 days, and counting only clicks at least 14 days old is the safe default because it costs you only patience.
| Signal | What it usually means | Action |
|---|---|---|
| 24+ clicks older than 14 days, 0 orders | Spend has passed twice the royalty with nothing back | Pause the keyword, or add as an exact negative in auto |
| Orders and ACOS under 24.0% | Profitable on the direct sale | Move the term to the exact-match campaign, raise its bid a little |
| Orders but ACOS above 24.0% | Converts, but the click costs too much | Lower the bid, and retest before any pause |
| Plenty of impressions, few clicks | The cover or title loses at thumbnail size, or the term is off-target | Check the listing before touching bids |
| Few impressions at all | Bid or budget is too low to win the placement | Raise the bid slightly, but never above break-even CPC; if it still wins nothing, fix price, page count or listing |
The 24-click pause rule limits how long you defend a term that never delivered. It is not noise-proof: at the 8.3% break-even conversion, a term has roughly a one-in-eight chance of showing zero orders in 24 clicks, so about one break-even keyword in eight gets paused by bad luck. Retest a paused term later at a lower bid, or use 36 clicks (three times the royalty) for terms you have strong reason to believe in.
Review every two weeks. More frequent tinkering usually means reacting to noise.
Negative Keywords and Harvesting Search Terms
The loop that separates a profitable campaign from a leaky one is harvesting. Terms that convert in the auto campaign move to your manual exact campaign, where you control the bid. The same term then becomes a negative in the auto campaign so the two do not bid against each other.
Negative keywords are the other half. Every irrelevant click that never orders is a cost you can remove. Common candidates for puzzle books are words that signal the wrong product, such as a different puzzle type or a digital download. Start with exact negatives; phrase negatives can block relevant traffic.
Worked Example: One Large-Print Word Search Book
Book: a 6 x 9 in large-print word search book at $9.99, example royalty $2.40, break-even ACOS 24.0%. The publisher runs an auto campaign at $5 a day and a manual phrase campaign at $5 a day for 14 days, then ends both and reads the report on day 28, once late orders have landed. That is deliberately below Amazon's roughly $10 suggestion: it caps the whole test at $140 (2 x $5 x 14), and at bids of $0.20 or less it buys about 25 clicks a day per campaign.
Look at one keyword that gets 60 clicks at an average cost of $0.20, so it spends $12.00. The three possible outcomes show why the rules exist.
| Orders from 60 clicks | Conversion | Ad sales | ACOS | Profit after ads |
|---|---|---|---|---|
| 3 | 5.0% | $29.97 | 40.0% | -$4.80 |
| 5 | 8.3% | $49.95 | 24.0% | $0.00 |
| 7 | 11.7% | $69.93 | 17.2% | +$4.80 |
Check the arithmetic: 3 orders earn 3 x $2.40 = $7.20 in royalties against $12.00 of spend, a loss of $4.80. Five orders earn $12.00 against $12.00, exactly break-even. Seven orders earn $16.80, a $4.80 profit. ACOS is $12.00 divided by ad sales in each row.
The middle row is the trap. A break-even keyword looks acceptable in the summary but adds effort and ties up budget. The publisher's rule is to lower its bid to $0.16 and retest: if conversion holds at 8.3%, the same 60 clicks now cost $9.60 against $12.00 of royalty, a $2.40 profit. If clicks fall away, the term goes.
The top row has orders but ACOS above 24.0%, so it also gets a lower bid and a retest, not an immediate pause. At 5% conversion the break-even CPC is $2.40 x 0.05 = $0.12, so the publisher cuts the bid to about $0.10. If conversion holds, 60 clicks cost $6.00 against $7.20 of royalty, a $1.20 profit. The term is paused only if a later window reaches 24 clicks with no orders.
The bottom row earns promotion to the exact campaign. The numbers are illustrative and use an example royalty, so run your own royalty and printing cost through the KDP calculators before copying any threshold.
Scaling What Works, and Where Tools Help
Once a book has a stable, sub-break-even ACOS on a handful of exact terms, scale in small steps. Raise the daily budget by roughly a quarter at a time and watch ACOS for a week before the next step. Sudden jumps change which placements you win and can push ACOS around for reasons unrelated to your keywords.
A note on going above break-even. Series read-through and organic rank gains can justify a higher ACOS on a first book, but only under a rule written before launch. The rule is: accepted ACOS = (royalty + assumed follow-on value per buyer) / list price, with the follow-on value labelled as an assumption.
Worked example: assume 20% of buyers go on to buy a second title that earns $2.40, which adds 0.20 x 2.40 = $0.48 per buyer. Accepted ACOS is (2.40 + 0.48) / 9.99 = 28.8%, against 24.0% at break-even. At 8% conversion the accepted CPC rises from $0.19 to 2.88 x 0.08 = $0.23, so the bid ceiling moves by only a few cents.
Write two more lines beside it: a total-loss cap and a review date. For example, stop when cumulative loss against the direct sale reaches $60, or at day 60, whichever comes first, and continue only if the second-title sales you measured in your KDP reports reach at least the 20% you assumed. If you cannot measure follow-on sales, fall back to break-even.
Then repeat the framework for the next title, starting again at break-even ACOS. A second book in the same niche can borrow the negative list and proven terms as a starting point, though it still needs its own test.
Octozia's KDP Creator sits on the production side, not the advertising side. It generates puzzle interiors (Sudoku, word search, maze, math and number fill by default) as numbered page images for you to assemble, on a 3-day trial with a card required. See pricing for current plans. It does not run or manage ads.
FAQ
How much should I spend on KDP ads to start?
Amazon recommends at least about $10 per campaign per day for Sponsored Products as of September 2026, though lower amounts are accepted. Decide first how much you can afford to lose while testing, split it across two campaigns, and make sure the daily budget buys enough clicks, roughly 20 or more a day, to give readable data.
What is a good ACOS for a KDP book?
Any ACOS below your break-even is profitable on the direct sale, and break-even is royalty divided by list price. A book earning a $2.40 royalty at $9.99 breaks even at 24.0%. A universal target does not exist because royalties differ by price, trim size and page count.
Is KDP Ads Manager the same as Amazon Ads?
KDP hands you off to the Amazon Ads console, which is where campaigns are created and reported on. Publishers often call that console Ads Manager. Which ad types you see depends on your account and marketplace, so check what is offered on the day you set up rather than relying on older guides.
Should I start with auto or manual targeting?
Run both. Auto targeting discovers search terms and competing ASINs you would never guess, and manual keyword targeting lets you test your own list under tighter control. Later, terms that convert in auto move to an exact-match manual campaign, and the same terms become negatives in auto.
How long should I wait before changing a campaign?
Give the first test at least two weeks, and judge each keyword by clicks, not calendar days. A workable rule is to pause a term once its spend passes twice your royalty with no orders, counting only clicks at least 14 days old so late orders land first. Review on a two-week rhythm afterwards.
Build the interiors your ads will point to
KDP Creator generates puzzle page images for you to assemble into a book; 3-day trial, card required, cancel anytime before it ends.
