Back to Blog
GuidesNovember 14, 2026· 11 min read

Amazon Ads Automation Dry Run: How to Test Your Rules Without Risking Budget

I

InteliAds Team

KDP Ads

Amazon Ads Automation Dry Run: How to Test Your Rules Without Risking Budget

An Amazon Ads automation dry run runs your rules against your real data and shows what they would change, without sending anything to Amazon. You read the list, check it against your own judgment, and only then let the rule run for real. In InteliAds, the dry run applies to IF/THEN Rules. This article explains what it records, how to read the result, a seven-day plan for testing a new rule, and the situations where a dry run will not protect you.

We build InteliAds, so this is written from how our feature works. We say where it stops. Where we describe a competitor, the source is linked.

What an Amazon Ads automation dry run is

A dry run is a rehearsal. The rule engine loads your enabled rules, finds the entities that match each rule's conditions, works out the new value each action would set, and records the plan. Live mode does the same and then sends the change to Amazon. Dry run stops before that last step.

In InteliAds, the Rules page offers two buttons when you execute rules by hand: Dry Run (Preview Only) and Normal Run (Make Changes). Normal Run needs you to tick a confirmation box. Dry Run does not, and you can use it as often as you like. The app notes that a dry run runs all enabled rules and ignores each rule's usual frequency. Dry Run is included on every plan.

A dry run doesA dry run does not
Evaluate your conditions against current dataSend any change to Amazon
Work out old and new values for each matchChange a bid, budget, keyword or campaign
Apply your bid caps and note when a value was cappedPredict what the data will look like tomorrow
Skip entities that are in cooldownShow what a second or third run would do after the first
Save a history record marked "Dry run"Tell you whether the rule is a good idea

The history record is the one write a dry run makes: a row in your execution history so you can open it later, plus an update to the rule's last-run time. It changes nothing in your Amazon account.

Dry run is not BidBot, and not Undo

Two other safety ideas get mixed up with dry run.

BidBot is a different feature. BidBot scores keywords and targets and proposes small bid changes. It does not use Rules Dry Run. Its nearest equivalent is the Suggest only preset, where it creates recommendations and you apply them yourself. We keep them apart because a clean dry run of your rules says nothing about what BidBot would propose.

Undo is a different safeguard. An undo restores something after it has happened. A dry run checks before. Adigy, a competitor, describes a Master Undo that restores an account to an earlier state in one click, according to its home page. That is a useful safeguard, and it answers a different question. Money spent on clicks between the change and the restore is gone either way. In InteliAds, history lets you revert supported changes such as bid changes, but keywords and negatives added by rules cannot be reverted from history, and the app says so. For those actions, the dry run is the only protection you have before the change. You remove an unwanted keyword by hand in Amazon. Amazon notes that negative targets are not permanent.

How to run one

  1. Open Rules, then the Rule Execution Logs tab, and choose Execute Rules.
  2. Check the list of enabled rules in the dialog. A dry run covers all of them. To test one rule, switch the others off first.
  3. Choose Dry Run (Preview Only).
  4. Open the execution when it finishes. The drawer shows how many entities were evaluated, applied or skipped. Open full details for the per-entity table.

How to read the result

Read the result from the top down.

What you seeWhat it meansWhat to do
Evaluated is 0The rule found nothing to look at. The problem is the selection or the data, not your conditions.Check the rule's target, filters and the campaigns in scope
Evaluated is above 0, matched is 0The rule looked at entities and none passed every conditionLoosen a condition, or accept that nothing needs action
Matched is very highYour thresholds may be too looseSee the sanity checks below
Old value and new value columnsThe change the rule would makeSpot-check some against the Amazon console
A "capped" tag on a valueYour minimum or maximum bid changed the resultMake sure the cap is what you intended
Skipped with a reasonThe rule chose not to act, for a stated reasonRead the reason. See the list below.

The distinction in the first two rows comes from how InteliAds records the result: it stores the number of entities checked separately from the number that matched, so a zero can be told apart from a rule that filtered everything.

Reasons the engine can give for skipping an entity include a bid already below your minimum when the rule decreases, a bid already above your maximum when the rule increases, a budget already at the value you asked for, and a lifetime-budget campaign when the rule sets a daily budget.

Sanity checks on the list

These are our heuristics, not product limits.

  • Count. If a rule matches more than about half of the entities in its scope, the threshold is probably too loose. A rule that cuts waste should match a minority.
  • Spot check. Pick five entities and look each one up in the Amazon console. Do the clicks, spend and orders match what the rule used?
  • Direction. Every row of a bid-down rule should have a lower new value than old. A set-bid rule can move bids up or down, and the app warns that set values move both ways.
  • Own assets. For negatives, read the whole list for your own books, your own series, and brand terms you want.
  • Caps. Check that the capped values look reasonable, not a long list of rows all hitting the floor.

A seven-day testing plan for a new rule

This is our suggested routine, not a built-in feature. The example is invented: a rule that adds a negative exact keyword for any search term with at least 20 clicks and no orders in 60 days.

DayWhat to doWhat you are checking
1Scope the rule to one book or one campaign. Switch other rules off. Run a Dry Run.How many terms match, and whether the list looks right
1Check five matches in the Amazon consoleThat the rule reads the same numbers you do
2Run another Dry RunHow much the list changes as new data comes in. A stable list means a stable rule.
3Fix the rule: tighten a threshold or remove a bad target. Run another Dry Run.That your fix did what you meant
4Decide: go live on this one scope, or stopA written yes or no
4Run Normal Run. Tick the confirmation.That live results match the last dry run
5 to 6Do nothing. Let the data settle.Amazon takes time to report. Barry of Barry KDP says not to judge the most recent 48 hours.
7Review the history against the plan. Widen scope or stop.Whether the changes did what you wanted

A worked example, with invented numbers. On day 1 the negative rule matches 14 of 120 search terms in the campaign, which is a minority. You spot-check five. Two are ASINs of your own series, which should not be negated, so on day 3 you change the target to exclude ASIN terms and rerun it. The new list has 11 terms. All 11 are keyword phrases that you agree should go. You run it live on that one campaign on day 4 and save the list in a notebook. On day 7 you compare the live execution with the dry run. If they differ, find out why before you scale to other campaigns.

What a dry run cannot tell you

Honest limits matter more than features.

  • It shows what the rule does, not whether the rule is wise. A rule that cuts every bid by 90% will dry-run perfectly. The preview tells you what you wrote.
  • It is a snapshot. It uses today's data. A rule that looks fine today can match differently next week.
  • It cannot show sequences. A dry run changes nothing, so it cannot show what the next run would do after the first one lowered bids. A rule that repeats can spiral. A $0.60 bid cut by 20% four times ends near $0.25. Use minimum bid caps and watch the second live run closely.
  • It does not test Amazon's response. Amazon may reject a change, or report it differently. A dry run never calls Amazon to write.
  • It inherits data lag. The numbers your rule reads are the numbers Amazon has reported so far. Recent orders may not be in them yet.
  • Interactions are easy to miss. Two rules can each look fine alone and fight each other. Dry run all enabled rules together, as the button does, and read the overlap.
  • It cannot undo or cover what is not reversible. Dry run is the safeguard for added keywords and negatives, which cannot be reverted from history.
  • It does not replace a budget limit. Amazon says a Sponsored Products daily budget is averaged over a month, so a $10 daily budget caps monthly spend at about $300. That outer limit sits above any rule.

Other layers that protect your budget

LayerWhat it doesWhere it applies
Dry RunPreviews rule actions with no Amazon writesIF/THEN Rules
Minimum and maximum bid capsStop a bid change from going beyond your limitPer rule, plus account guardrails
Entity cooldownSkips an entity a rule changed recently (48 hours by default)Rules, account-wide setting
History and revertReverts supported changes after the factBid changes and other supported actions
Amazon daily budgetCaps spend per campaignAmazon, outside any tool
BidBot Suggest onlyCreates bid recommendations that you apply yourselfBidBot, a separate feature

Frequently asked questions

What is a dry run in Amazon Ads automation?

A dry run is a test mode that evaluates your automation rules against current data and records what they would change, without changing anything in your Amazon Ads account.

Does a dry run change anything in Amazon?

No. It sends no bid, budget, keyword or campaign change to Amazon. It does save a history record in InteliAds that you can open later.

Is a dry run the same as BidBot?

No. A dry run belongs to the rules feature. BidBot is separate. Its Suggest only preset gives you recommendations to apply yourself, and a dry run of your rules says nothing about what BidBot would propose.

How many times should I dry run a rule before going live?

There is no magic number. We suggest at least two runs on different days, so you can see whether the list is stable, and one spot-check of five entities against the Amazon console. Go live on one campaign first.

Can I dry run a single rule?

The Dry Run button runs all enabled rules. To test one, switch the others off for the test.

Can a dry run lose me money?

Not by itself, because it sends no changes. The risk is trusting it too much. A rule that previews cleanly can still be a bad rule, and the first live run can differ from the dry run if the data moved in between.

Is a dry run better than an undo button?

They do different jobs. A preview helps you avoid a mistake, and an undo helps you recover from one. Prefer both. For changes that cannot be reverted, like added keywords, the preview matters more.

Sources

How InteliAds does this

The Dry Run button is on the Rules page, and you can run it as often as you like on any plan. Each dry run is saved in the execution history with a Dry run tag, and the full-details page lists every entity with its old and new value and flags capped values. Bid changes and other supported actions can be reverted from history after a live run. Keywords and negatives you add by rule cannot, so we ask you to dry run those first.

BidBot, the bid-scoring feature, is separate. It proposes small steps of $0.02 to $0.05, waits a cooldown after each change, and before writing a bid it reads the current bid from Amazon and stops if someone changed it outside InteliAds. Its Suggest only preset never applies anything by itself. If you want to test your own rules this way, see InteliAds plans and read our 12 rules article for starting points. An Amazon Ads automation dry run costs nothing and takes minutes, so use it every time.

Put these rules to work

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