12 Amazon Ads Rules for KDP Books: Copy-Paste IF/THEN, and When Each One Backfires
InteliAds Team
KDP Ads
Amazon Ads rules for KDP books work when each rule has a narrow trigger, a small action and a cap. They fail when a rule acts on thin data, repeats without limit, or hits a book that is still launching. This article gives twelve IF/THEN rules you can copy into any tool, or run by hand, with starting numbers, the reasoning behind them, and the cases where each one backfires.
The numbers come from three places: Amazon's own documentation, advice from experienced KDP advertisers (named when we quote a specific figure), and one invented example book so you can see the arithmetic. None of it comes from InteliAds customer accounts. Treat every threshold as a starting point to test, not a law.
The example book used in every rule
Every dollar figure below uses this invented book. Replace it with yours.
| Input | Example value | How to get yours |
|---|---|---|
| List price | $15.00 | Your KDP price |
| Net royalty per sale (R) | $5.00 | KDP royalty for that format and marketplace |
| Break-even ACoS | 33.3% | Royalty divided by list price (5 / 15) |
| Break-even spend per sale | $5.00 | Equal to R |
| Conversion rate (assumed) | 10% | Orders divided by clicks. 10% is a common reference point among KDP advertisers, not a promise. |
| Break-even CPC at that conversion rate | $0.50 | R times conversion rate (5 x 0.10) |
Break-even ACoS as royalty over price is how several creators define it, including Gerald of Driven Publishers and Barry of Barry KDP. If you do not know your numbers yet, our break-even ACoS calculator does the division.
One limit to know first. InteliAds rule conditions use Amazon-reported metrics: ACoS, CPC, spend, sales, impressions, clicks, orders, conversion rate, ROAS and current bid or budget. There is no field for royalty or page reads, so you convert your royalty into spend or ACoS numbers yourself, as this article does. The same is true of most rule builders.
Amazon Ads rules for KDP at a glance
| # | Rule | Acts on | Action | Main risk |
|---|---|---|---|---|
| 1 | Negate keyword search terms with no sales | Search terms | Add negative exact | Blocks good terms that were unlucky |
| 2 | Negate ASIN search terms with no sales | Search terms (ASIN) | Add negative product target | Can block your own series |
| 3 | Cut bids on targets that spent a royalty with no sale | Keywords, product targets | Decrease bid 20% | Spiral to the floor |
| 4 | Pause targets that spent three royalties with no sale | Keywords | Pause | Never retested |
| 5 | Step down bids above break-even | Keywords | Decrease bid 5% | Punishes launch campaigns |
| 6 | Step up bids on efficient winners | Keywords | Increase bid 10%, capped | Raises CPC faster than sales |
| 7 | Raise bids on targets with almost no impressions | Keywords | Increase bid 10%, capped | Low demand, not low bid |
| 8 | Harvest winning search terms into exact keywords | Search terms | Add as exact target | Duplicates, self-competition |
| 9 | Cut bids on auto-targeting groups that bleed | Auto targets | Decrease bid 20% | Starves discovery |
| 10 | Stop-loss pause on campaigns with zero sales | Campaigns | Pause | Hides a listing problem |
| 11 | Raise the budget of winning campaigns | Campaigns | Set daily budget | A SET can lower a budget |
| 12 | Raise the top-of-search modifier on a placement that works | Campaigns | Increase placement adjustment 10 points | No built-in ceiling |
How to read each rule
Each rule below gives the trigger in InteliAds rule fields. A "window" is the number of days of data a condition looks at, from 1 to 90 days in InteliAds. Conditions in one rule must all be true for an entity to match. Actions that change a bid can carry a minimum and a maximum bid, and the rule skips an entity whose bid is already outside the cap in the direction of the change. InteliAds also leaves alone any entity a rule changed within the cooldown window, which is 48 hours by default and set per account.
Rule 1: Negate keyword search terms with no sales
| Field | Setting |
|---|---|
| Target | Search terms, type keyword |
| IF | Clicks >= 10 (60-day window) AND orders = 0 (60-day window) |
| THEN | Add as negative keyword, negative exact |
Why these numbers. Barry of Barry KDP, Gerald of Driven Publishers and the Self Publishing Empire channel each use about 10 clicks with no sale as the point to act on a search term. Amazon is more cautious: its targeting guide recommends evaluating a keyword after at least 20 clicks before adding it as a negative, and letting negatives run for two weeks or more before changing strategy.
The reason for caution is arithmetic. If a term truly converts at 10%, the chance of seeing zero orders in n clicks is 0.9 to the power of n.
| Clicks with zero orders | Chance a true 10% term looks dead |
|---|---|
| 5 | 59.0% |
| 10 | 34.9% |
| 20 | 12.2% |
| 30 | 4.2% |
| 40 | 1.5% |
When it backfires. At 10 clicks, about one good term in three will look dead by chance. A negative exact also blocks close variations of the phrase, according to Amazon's description, so a singular and plural form can both go. The report can lag: Barry says never to judge the last 48 hours. And added negatives are not reverted by InteliAds rule history, so you remove them by hand in Amazon. Amazon notes that negatives are not permanent.
Safer variant. Use 20 clicks, matching Amazon's own number.
Rule 2: Negate ASIN search terms with no sales
| Field | Setting |
|---|---|
| Target | Search terms, type ASIN |
| IF | Clicks >= 20 (60-day window) AND orders = 0 (60-day window) |
| THEN | Add as negative product target |
Why these numbers. Amazon says targeted ASINs that earn clicks appear in the search term column of an automatic campaign's report, and recommends the 20-click evaluation point for negatives. We use the longer window because an ASIN is a whole product, and a quick negative on it is harder to notice and undo.
When it backfires. The rule cannot tell your own books from competitors. If you write a series, a reader browsing book 1 may click your book 2 ad and not buy yet. Negating your own ASIN stops that ad from appearing there. Run Dry Run first and read every ASIN on the list.
Rule 3: Cut bids on targets that spent a royalty with no sale
| Field | Setting |
|---|---|
| Target | Keywords (copy for product targets) |
| IF | Spend >= $5.00 (30-day window) AND orders = 0 (30-day window) |
| THEN | Decrease bid by 20%, minimum bid $0.15 |
Why these numbers. Barry's masterclass frames it in royalty terms: a target that spends more than your royalty with no sale should be lowered or removed. Chris Rawlings, who teaches Amazon ads in general and not only for books, lowers a bid by 20 to 30% each time, and prefers lowering over negating because a lowered bid can later go back up.
When it backfires. Gerald argues that aggressive bid cuts are a losing strategy: lower bids win lower placements, conversion can fall, and ACoS may rise. A rule that repeats makes this worse. A $0.60 bid cut by 20% four times ends at about $0.25. The minimum bid is what stops the spiral. Also, spend in the last two days may have orders that Amazon has not reported yet.
Rule 4: Pause targets that spent three royalties with no sale
| Field | Setting |
|---|---|
| Target | Keywords |
| IF | Spend >= $15.00 (60-day window) AND orders = 0 (60-day window) |
| THEN | Pause |
Why these numbers. Yevgen of AsDroid gives one pause rule: a target left alone that spent $20 to $30 with no sales should get its bid halved or be paused. Three royalties is $15 in our example, so scale to your own royalty. He also says to pause rarely and to prefer a lower bid.
When it backfires. Paused targets are never retested, and no creator gives a rule for bringing them back. Barry's workaround is to relaunch targets with no traction in a new campaign with an average CPC bid and about a $5 budget after 30 to 60 days. Because the rule requires spend, it never touches a target with zero impressions, which Barry says to raise, not pause. Keep this rule off launch campaigns.
Rule 5: Step down bids above break-even
| Field | Setting |
|---|---|
| Target | Keywords |
| IF | Orders >= 2 (30-day window) AND ACoS > 35 (30-day window) |
| THEN | Decrease bid by 5%, minimum bid $0.15 |
Why these numbers. The 35 sits just above the 33.3% example break-even. Barry lowers 5% every 3 to 5 days on keywords with very high ACoS. His shorter $5-a-day routine uses 3 to 5 cents every 2 to 3 days. Gerald lowers 5 to 10 cents, and Ken Fornari lowers 10 cents. The size of the step is the main disagreement among these creators, so start small.
When it backfires. Launching books are meant to run above break-even. The Self Publishing Empire channel describes 40 to 50% ACoS in the first one to two months, and Yevgen accepts 60 to 80% in the first month. Gerald runs mature books 10 to 15% above break-even on purpose to keep rank. A single bad search term can also drag a whole keyword's ACoS up, so negate first (Rules 1 and 2) and only then lower the bid. Set your account cooldown above 48 hours if you want Barry's 3 to 5 day spacing.
Rule 6: Step up bids on efficient winners
| Field | Setting |
|---|---|
| Target | Keywords |
| IF | Orders >= 2 (30-day window) AND ACoS <= 20 (30-day window) |
| THEN | Increase bid by 10%, maximum bid $1.00 |
Why these numbers. Ken raises a bid 10 cents when ACoS is under 20%. Barry raises 10% and waits 3 to 4 days before doing it again. Ken also says never to bid over $1 on high-content books, while Yevgen's default upper limit is $1.20. They disagree, so the cap is yours to set.
When it backfires. Two orders can be luck, and a higher bid raises CPC immediately while sales follow slowly. More clicks can also use the daily budget sooner. Amazon says daily budgets are averaged over the month, so a campaign can spend more on some days. If a campaign uses Amazon's up-and-down dynamic bidding, Amazon can already raise your bid by up to 100% for top-of-page placements and 50% elsewhere, so a rule increase stacks on top.
Rule 7: Raise bids on targets with almost no impressions
| Field | Setting |
|---|---|
| Target | Keywords |
| IF | Impressions < 300 (3-day window) |
| THEN | Increase bid by 10%, maximum bid $1.20 |
Why these numbers. Gerald treats under about 100 impressions a day as a signal to raise the bid and wait 2 to 3 days. Barry says never to pause a keyword with no impressions and to raise the bid instead. Yevgen suggests waiting for 1,000 impressions before judging a target at all.
When it backfires. Low impressions can mean low demand, not a low bid. Barry says the two cannot be told apart except by raising. A keyword that is simply irrelevant will keep getting raised until it hits the cap. Yevgen says 1,500 impressions with no clicks points to a relevance problem. The rule does not conflict with Rules 3 and 4, because a target with almost no impressions has almost no spend.
Rule 8: Harvest winning search terms into exact keywords
| Field | Setting |
|---|---|
| Target | Search terms, type keyword |
| IF | Orders >= 2 (30-day window) AND ACoS <= 30 (30-day window) |
| THEN | Add as target keyword, exact match, with a fixed bid you choose. Also negate in the source campaign. |
Why these numbers. Ken moves a term to a winning exact-only campaign when it has at least two sales and an ACoS of 30% or less, and sets the bid to the CPC it had when found, so use that CPC as your fixed bid. Sam of Samazonkdp negates the exact keyword in the automatic campaign so you do not bid against yourself.
When it backfires. The rule has no way to see that the term is already an exact keyword in a manual campaign, so Dry Run may list duplicates. Negating in the source also removes the cheap traffic the automatic campaign was sending, and the new exact keyword may not win the auctions the old one did. Added keywords are not reverted by rule history. Two orders at 30% can be luck. Raise the order count if your clicks are expensive.
Rule 9: Cut bids on auto-targeting groups that bleed
| Field | Setting |
|---|---|
| Target | Auto targeting (run separately for close match, loose match, substitutes and complements) |
| IF | Spend >= $5.00 (30-day window) AND orders = 0 (30-day window) |
| THEN | Decrease bid by 20%, minimum bid $0.15 |
Why these numbers. Barry splits an automatic campaign into its four targeting groups so each can be judged alone. The thresholds are the same as Rule 3. One rule per group lets you set a different number for each.
When it backfires. Automatic campaigns are where new search terms are discovered. Starving a group cuts the supply Rules 1, 2 and 8 depend on. Review the groups monthly, not weekly.
Rule 10: Stop-loss pause on campaigns with zero sales
| Field | Setting |
|---|---|
| Target | Campaigns (choose them by hand) |
| IF | Spend >= $20.00 (14-day window) AND orders = 0 (14-day window) |
| THEN | Pause |
Why these numbers. Four royalties is $20 in the example, in line with Yevgen's $20 to $30 range. Barry warns against pausing a campaign for high ACoS, because one or two targets usually cause it. This rule fires only on zero sales.
When it backfires. Amazon's authors guide suggests waiting about two weeks after launch before changing a bidding strategy, so keep new campaigns out of scope. Zero sales across a whole campaign often points to the listing, cover or reviews, not the targets. Ken wants at least 20 reviews before advertising paperbacks. Pausing the campaign hides the problem instead of fixing it.
Rule 11: Raise the budget of winning campaigns
| Field | Setting |
|---|---|
| Target | Campaigns |
| IF | ACoS <= 25 (14-day window) AND orders >= 5 (14-day window) AND daily budget < 15 |
| THEN | Set daily budget to $15.00 |
Why these numbers. Ken and Sam of Samazonkdp both say to put more budget into a campaign that performs, and neither gives a number. The figures here are examples. Treat them as a placeholder.
When it backfires. This action sets a value. It does not add to one. A rule that sets $15 on a campaign already at $40 would lower it, which is why the third condition caps the budget in the condition itself. Rules also have no field for how much of the budget a campaign actually used, so you cannot test whether it was capped. If it was not, more budget changes nothing.
Rule 12: Raise the top-of-search modifier on a placement that works
| Field | Setting |
|---|---|
| Target | Campaigns |
| IF | Top of search ACoS <= 25 (30-day window) AND top of search orders >= 5 (30-day window) |
| THEN | Change placement adjustment: top of search, increase by 10 points |
Why these numbers. Chris Rawlings starts a top-of-search modifier at 20 to 40%, raises it as it earns, and looks at ACoS and order count together. In his example he raised it by 30 points because top of search drives rank. He also says the maximum is 900% and not to start there.
When it backfires. A rule only clamps the value between 0 and 900. It does not stop at a ceiling you choose, so each run can add another 10 points. Each run that clears the cooldown can add another 10 points, so with a 48-hour cooldown five runs add 50 points in about ten days. The modifier also multiplies your bid, and it stacks with dynamic up-and-down bidding. Modifiers apply to the whole campaign, so Chris isolates each important keyword in its own campaign.
Run order: clean waste before you move bids
Gerald's advice is to cut waste before touching bids. Apply the rules in this order, and switch on growth rules last.
| Stage | Rules | Goal |
|---|---|---|
| 1. Remove waste | 1, 2 | Stop paying for terms that never buy |
| 2. Limit losses | 3, 4, 9, 10 | Cap what a failing target can spend |
| 3. Correct | 5 | Move above-break-even targets toward break-even |
| 4. Grow | 6, 7, 8, 11, 12 | Scale only what already works |
Convert the rules to your own book
| Setting | Example (R = $5.00) | Formula | Source of the idea |
|---|---|---|---|
| Negate after N clicks | 10 to 20 | Choose from the probability table | Creators use 10, Amazon says 20 |
| Zero-sale spend for a bid cut | $5.00 | 1 x R | Barry |
| Zero-sale spend for a pause | $15.00 | 3 x R | Yevgen ($20 to $30 for his example) |
| Campaign stop-loss | $20.00 | 4 x R | Yevgen |
| Bid step-down trigger | ACoS above 35 | Break-even ACoS plus about 2 points | Our choice |
| Winner trigger | ACoS at or below 20 | About 60% of break-even ACoS | Ken (20%) |
| Bid ceiling | $1.00 | Your CPC limit | Ken $1.00, Yevgen $1.20 |
Test every rule in Dry Run first
A rule that looks right on paper can match far more than you expect. Run it in Dry Run, read the list, and check five entries by hand against the Amazon console. We cover how to read the result, and what it cannot show you, in Dry Run: testing ads automation without risking budget. If you want the same checks done in a spreadsheet, use our search term report template.
Frequently asked questions
Do Amazon Ads rules work for KDP books?
They work for repetitive, low-risk decisions such as negating terms that spend and never sell. They work less well on launching books, low-data keywords and books whose income comes mostly from page reads.
How many clicks should I wait before negating a search term?
Creators use about 10 clicks with no sale. Amazon recommends at least 20 clicks before adding a negative. Chris Rawlings suggests 30 to 40 clicks at a 10% conversion rate for a permanent negative. The probability table above shows why the higher numbers are safer.
Should a rule pause a keyword or lower its bid?
Lower the bid first. Yevgen pauses rarely, Barry says not to pause a target with no impressions, and Chris notes that a negated or paused term never gets another chance. Pause only after clear spend with no sale, as in Rule 4.
How often should rules run?
Experienced advertisers check bids two to three times a week. Barry waits two to three days after a change. Your rules' schedule can be more frequent, but your cooldown should still leave days between changes to the same target.
What ACoS should a rule aim for?
Break-even ACoS, which is royalty divided by list price, is the line. Creators often aim lower once a book is established: Ken uses 20 to 35%, and the Self Publishing Empire channel aims for under 30%. During a launch most of them allow higher.
Do rules work for Kindle Unlimited ebooks?
Be careful. Rule conditions use ACoS, orders and sales, and have no field for page-read earnings. An ebook that earns mostly through page reads can look worse in a rule than it is. Check it by hand before automating.
Can I run rules and BidBot together?
By default, when BidBot's automatic mode is on and available, your IF/THEN rules are paused. There is a setting to change that. Pick one system per book unless you have a reason to do both.
Sources
- Amazon Ads: targeting with Sponsored Products (negative keywords, 20 clicks, two weeks)
- Amazon Ads: Sponsored Products for authors (bidding strategies, two-week wait, daily budget averaging)
- Barry KDP: masterclass, $5 a day campaign
- Driven Publishers (Gerald): beginner bidding, advanced bidding
- Yevgen (AsDroid) with Sean Dollwet: interview
- Ken Fornari: strategy video
- Chris Rawlings (general Amazon ads, not KDP only): bids and clicks, placement modifiers
- Self Publishing Empire: year-to-date review, 25 tips
- Samazonkdp: low budget strategy, three steps
Creators disagree with each other on many of these numbers, and some sell coaching or ad management. We paraphrased them and did not verify their results.
How InteliAds does this
InteliAds Rules are IF/THEN conditions on the metrics above, with the actions used in this article: negative keyword, add as target, bid up or down by percent or dollars with a minimum and maximum cap, pause, set budget and placement adjustment. You can run any rule in Dry Run first. Added keywords and negatives cannot be reverted from rule history, and bid and pause changes can be reverted for supported actions.
Two limits matter for this article. A rule cannot see your royalty, so you set the numbers. And a rule has no cap on how far a repeated placement or bid change can go, apart from the bid caps you set and Amazon's 0 to 900% placement range. BidBot is the other system: it scores each target against its break-even and proposes steps of $0.02 to $0.05, with its own cooldown and a ceiling on placement moves. If you want to test these rules on your own account, see InteliAds plans. Whatever tool you use, start Amazon Ads rules for KDP books with the waste rules, and preview each one before it goes live.