Back to Blog
GuidesNovember 4, 2026· 14 min read

KDP Ads Audit Log: What to Demand Before You Let Automation Change Your Bids

I

InteliAds Team

KDP Ads

KDP Ads Audit Log: What to Demand Before You Let Automation Change Your Bids

You open your Amazon Ads dashboard on a Monday and the numbers look different. A keyword you remember at $0.62 now sits at $0.41, a budget has doubled, and one campaign is paused. You cannot say whether you did it, a rule did it, or a tool did it. That gap is why a KDP ads audit log matters before you let any automation touch your Sponsored Products bids.

This guide is for authors, publishers, and agencies running Amazon KDP ads who want to know what a useful change record contains and what to do when one is missing. It covers what to demand from any tool, how to rebuild events from the console, and where the console's own history stops helping. The aim is to make every bid, budget, and status change traceable and reversible.

Short answer: A KDP ads audit log is a timestamped record of every change made to your campaigns, showing who or what made it, the old and new values, and the reason. Demand it before enabling automation, because without it you cannot undo a bad change or learn which change moved your ACOS. At minimum, each entry should name the object, the before and after values, the actor, and a performance snapshot at the moment of change. If a tool cannot show these, treat its automation as unaccountable.

The scenario: waking up to bids you don't remember setting

Consider a hypothetical author with a four-book cozy mystery series running 6 Sponsored Products campaigns. Over a weekend, an automation rule is enabled and, separately, the author edits two bids by hand on Saturday. On Monday, spend for the weekend is $38.70, ACOS has moved from 41% to 67%, and the author sees 54 clicks and 2 orders. Three things changed, and nobody wrote any of them down.

The question is not "is 67% bad." The question is "which of the three changes produced it." With no record, the author has three suspects and no evidence. Reverting everything wipes out any change that was actually helping, and reverting nothing leaves the damage running.

The pattern is common whenever more than one actor can edit a campaign. These are the typical traps:

  • Overlapping actors: you, a rule, and a teammate or agency all edit the same keyword within 48 hours.
  • Silent budget changes: a daily budget increase raises spend without any bid moving, so the change looks like "Amazon is spending more."
  • Status changes with no note: a paused keyword quietly removes impressions, and you read the drop as a market shift.
  • Batch edits: 40 bids change in one pass and only the last few are remembered.

For background on how this fits the wider picture, see the KDP ads guide.

What a trustworthy change record contains, field by field

A log entry is only useful if you can act on it months later. Each field below exists to answer a specific question during a review. If any field is missing from a tool's log, you lose that answer.

Log fieldExample entryQuestion it lets you answerWhat breaks if it is missing
Timestamp with timezoneSat 14:05 EasternDid the change come before or after the ACOS jump?You cannot order events across a weekend
ActorRule "Trim low converters" vs. manual userWas this automation or me?You blame the wrong source
Object changedCampaign, ad group, keyword, match typeExactly which target moved?You revert the whole campaign instead of one keyword
Old value and new valueBid $0.62 to $0.41What do I restore?A revert becomes a guess
Trigger or reason"ACOS above 55% after 30 clicks"Was the logic sound for this book?You cannot judge the rule, only the outcome
Result statusApplied, failed, or dry runDid the change actually reach Amazon?Log and live account drift apart unnoticed
Spend and ACOS snapshot7-day spend $64.10, ACOS 48%What did performance look like when the change fired?You cannot tell a reasonable decision from a bad one
Royalty contextEstimated royalty per sale on that titleWas the ACOS threshold profitable for this book?A rule can cut a keyword that was actually profitable

Two of these deserve extra attention. The result status field separates a change that was proposed from one that took effect, which is where a dry run earns its place. The royalty context field matters because an ACOS rule that ignores what the book earns per sale can pause a profitable keyword. Our post on KDP royalties vs Amazon Ads sales explains why those two numbers should be read together.

Bid-desk tip: if a log shows only "bid updated" with no old value, treat it as a notification, not an audit log.

Reconstructing what happened when you have no log at all

If you already ran automation or shared access without a record, you can still build a rough timeline. It will be incomplete, so label what you know versus what you infer. Work from the evidence you can export and compare.

  1. Set the window. Pick the date range where performance changed, usually 3 to 7 days around the shift.
  2. Export current state. Download campaign, keyword, bid, budget, and status data from the console today so you have a fixed reference.
  3. Compare to any older export. A prior bulk file, spreadsheet, or screenshot gives you a "before" to diff against.
  4. Check daily spend and impressions by day. A budget change shows as a step in spend. A paused keyword shows as impressions falling to zero for one target.
  5. List every actor with access. Include rules, tools, teammates, and yourself, then ask each what they changed and when.
  6. Write the timeline down and start your own log from this moment forward.

The decision table below maps common evidence gaps to a practical console action.

Audit situationLikely gap in the recordConsole action to take
Spend stepped up on one day but bids look unchangedA budget edit was never recordedOpen the campaign, compare the daily budget to your last export, and restore the earlier figure if spend outran royalties
One keyword shows zero impressions since the weekendA status change to paused with no noteCheck the keyword status column, re-enable it if the pause was not intentional, and note the date
Several bids are lower than you rememberA batch edit with no old values savedUse your oldest export as the reference, restore bids on the highest-spend keywords first, and leave the rest for review
ACOS jumped but nothing obvious changedMultiple small edits across actorsRevert one change at a time on a single campaign, rather than all at once, so you can isolate the cause
A tool says it changed a bid but the console shows another numberLog and live account out of syncTrust the console value, record the mismatch, and pause the automation until the tool's log matches live data

Reverting one thing at a time is slower, but it is the only way to learn which change caused the problem. Reverting everything gives you relief and teaches you nothing.

Where Amazon's own console history stops being enough

The Amazon Ads console can show you current settings and performance reports. For book campaigns, that does not add up to a dependable, searchable audit trail. It is weak in four specific areas:

  • Restoring values: seeing that something changed is different from having the old value ready to restore in one step.
  • Attribution: the console does not reliably tell you whether a person or a connected tool made an edit.
  • Reasoning: it records no reason, so you cannot judge whether a rule's logic fit your book.
  • Royalty context: it reports ad sales, not what you earned per sale on each title, so profit is invisible at the point of change.

That is the gap an external log should fill. When you evaluate a manager for KDP advertising, ask whether it records old and new values, separates dry runs from applied changes, and ties each action to royalties once your sales data is linked. InteliAds is built around that approach for Sponsored Products, with IF/THEN rules you can dry run and track KDP ads and royalties in one dashboard. You can compare options on the InteliAds pricing page or create an InteliAds account to see how the records look on your own campaigns.

Whatever you choose, the standard is the same. If you cannot answer "what changed, who changed it, from what, and why" for any keyword in under a minute, you are not yet ready to hand over your bids.

The first half of this guide covered what a useful log contains. This half covers how to verify one and how to undo a change without creating a second problem.

How to test a KDP ads audit log before you trust it with live campaigns

Never judge a log by its screenshots. Test it on a small Sponsored Products campaign where a wrong bid costs a few dollars, and compare every entry against what the console shows. A log that only looks complete is the most dangerous kind.

Here is a fresh labeled hypothetical. A test campaign for book 1 of a cozy mystery series has a $6.00 daily budget and three keywords. The rule is: IF clicks are at least 12 AND orders equal 0, THEN lower the bid by 20%. The keyword "amish romance novels" sits at a $0.50 bid after 14 clicks and $6.30 spend with no orders, so the rule should propose $0.40.

Test stepWhat you doEntry the log must showPass condition
1. Dry runRun the rule without applying itProposed change: $0.50 to $0.40, rule name, timestamp, status "proposed"Nothing changes in the console
2. Approved applyApprove the one changeStatus "applied", actor identified as the ruleConsole bid reads $0.40
3. Manual editChange a different keyword by hand to $0.35Entry or flag marking a non-rule change, or a detected mismatchYou can tell it apart from step 2
4. Budget editMove the budget from $6.00 to $8.00Separate budget entry with old and new valuesBudget is not buried inside a bid entry
5. ReverseRestore $0.50 on the test keywordNew entry linked to the original, not a deleted rowHistory keeps both events

Bid-desk tip: a log that overwrites its own history when you revert has failed the test, because you can no longer reconstruct what happened.

Console actions for the test

  1. Open the test campaign in the Amazon Ads console and note each keyword bid, match type, status, and the daily budget before you start.
  2. Export your own snapshot of the campaign so you hold a baseline that does not depend on the tool.
  3. Run the rule as a dry run and compare the proposed values with your snapshot.
  4. Approve one change only, then reload the console and confirm the live bid matches the log.
  5. Repeat with a manual edit and a budget edit, and record whether the log separates all three.

Where InteliAds fits

InteliAds is a Sponsored Products manager with IF/THEN rules, a dry run that previews what a rule would change, and targeting filters to narrow which keywords or targets a rule can touch. When KDP Sync is linked, it shows royalty-aware profit next to spend and ACOS, so you can judge a change against what the book earned. The author approves live changes. Use the dry run as step 1 above and check the output against your own snapshot, as you would with any tool.

Reverting a bad change safely: order of operations and what to check afterward

A revert is itself a change. If you undo several edits in the wrong order, you can restore a bid onto a keyword that is still paused, or raise a budget while a runaway bid is still live. Work from the most expensive risk to the least.

  1. Stop further automation on the affected campaign so nothing overwrites your revert.
  2. Read the log entries for the window in question and list every field changed: bid, status, budget, targeting.
  3. Compare with the live console and treat the console as the source of truth for current values.
  4. Restore status first (a keyword paused by mistake), then bids, then budget.
  5. Reload the console and confirm each restored value, then confirm the log recorded the revert as a new entry.
  6. Re-enable automation only after the rule that caused the problem is edited or narrowed.

Afterward, do not judge the result on one day of data. Sales can post with a delay, so give the restored keyword several days and enough clicks to compare against its pre-change behavior.

Revert checkWhere to lookHealthy resultRed flag
Bid restoredKeyword row in consoleMatches the old value in the logConsole shows a third value
Status restoredKeyword and campaign statusEnabled where it was enabled beforeStill paused after revert
Budget restoredCampaign settingsOriginal daily budgetBudget changed by a rule you forgot
History intactAudit logOriginal and revert both visibleOriginal entry missing

For context on judging results once a revert is in place, see our KDP royalties vs Amazon Ads sales guide, and the broader KDP ads guide for campaign structure.

FAQ

Q: Does the Amazon Ads console keep a change history I can use to undo a bid change on a single keyword? A: The console may show change history for some campaign edits, but its detail and availability vary, and you should not count on a one-click undo. The safe habit is to keep your own record of old values so you can re-enter a bid manually.

Q: What is the minimum set of fields an audit log entry should have to be useful for a book campaign? A: Timestamp, account, campaign, ad group, the keyword or target, the field changed, the old value, the new value, who or what made the change, and the rule or reason. Add a status such as proposed, applied, or reverted.

Q: If a tool reverts a bid, can it also restore a paused keyword or a changed budget in the same step? A: Do not assume so. Bids, status, and budget are separate fields, so check that each has its own log entry and confirm each one in the console after the revert.

Q: How long should I wait after a revert before judging whether the original change was the problem? A: Wait at least several days and until the keyword has gathered a meaningful number of new clicks, since orders can post late. Judging after a single day mostly measures noise.

Q: How do I tell the difference between a change my automation made and a change I made manually? A: Look for an actor field on each entry, and use a consistent naming convention for rules. If the live console differs from the log with no matching entry, a manual or outside edit is the likely cause.

Q: Should I export my own snapshot of campaigns before enabling automation, and how often? A: Yes. Export once before you enable anything, then again before each new rule goes live, before a launch push, and on a regular schedule such as weekly.

Q: What should I do if a log shows a change but the live bid in the console doesn't match it? A: Treat the console as the current truth and pause automation on that campaign. Then check for a manual edit, a failed write, or a later change, and record what you find before re-enabling anything.

Q: Can an audit log help me prove which change caused a spike in ACOS across a series launch? A: It can show timing and narrow the suspects, but not prove cause alone. Line up each logged change against daily ACOS, change one thing at a time where possible, and rule out price, promotions, and seasonality.

A trustworthy kdp ads audit log records old and new values, names the actor, survives reverts, and matches the live console every time you check. If you want to track KDP ads and royalties alongside rules you can dry run first, See InteliAds plans or Create an account. Starter has no trial, and the author approves live changes.

Browse more KDP ads guides on InteliAds.

See InteliAds plans: compare Essential, Pro, and Publisher (Essential has no trial; Pro and Publisher include a 7-day trial). InteliAds is a Sponsored Products manager for KDP authors: royalty-aware profit when KDP Sync is linked, user-written IF/THEN rules with dry run, and targeting filters for keywords, product targets, and search terms. Create an account.

Tools and guides: KDP ads optimizer · break-even ACOS calculator

Put these rules to work

Set IF/THEN conditions, run them every hour, and revert any batch when you need to.