Performance workspace

Performance is the workspace where you work on the account by hand. Everything an Automation rule can do, you can do here — bid, state, placement, audience, budget, negatives — with the difference that you decide each one.

The table has one view per level: Products, Sponsored, Campaigns (the default), Ad Groups, Keywords, Competing, Creatives, Placement, Audiences and TOS Share. Competing lists targets on the same ASIN that Amazon treats as close variants of each other, so their traffic is split across several targets.

The bulk action pattern

Every bulk action opens the same dialog, and learning it once covers all of them.

Change bid, with Increase/decrease by $ selected.

1. Scope — what you are about to touch

Before anything else the dialog states how many rows you selected, what entity they resolve to, how many campaigns and ad groups that spans, and the current value range across the selection.

That range is the part worth reading. A selection whose bids already run from $0.30 to $2.49 will not respond uniformly to a flat change.

2. Input mode — how the number is applied

Mode	What it does
New value	Sets every selected row to the same amount
Increase/decrease by $	Adds to or subtracts from each row’s current value
Increase/decrease by %	A relative change, so each row moves in proportion to itself

A positive number increases and a negative number decreases: -10 takes $10 off, 5 adds $5.

New value flattens, the other two preserve shape. If your selection has a considered spread of bids, setting one new value throws that away. Increasing by a percentage keeps it.

3. Warnings and the preview

The dialog warns before you apply, not after. The warning names the count and the reason:

• Large-change warning — 50 placements move more than 10 % points from the current value.
• High Highest-placement-bid warning — 50 keywords end up with a highest placement bid above $5.

For bids it also previews the resulting range, and something less obvious: the highest placement bid. Change placement value shows the same thing from the other side: the enabled targets in the selected campaigns, with the highest placement bid each one ends up with.

Each preview opens into a row-by-row list, and you can untick rows there. If a new bid falls outside Amazon’s bid limits, the button reads Fix the bids first until you lower the amount or untick those rows.

The highest placement bid

Your bid is not what you can end up paying. Amazon multiplies it:

base bid × placement adjustment × audience adjustment × dynamic-bidding uplift

For Sponsored Products with up-and-down dynamic bidding, that uplift is +100% on top of search and +50% elsewhere, and Amazon applies it only when it judges conversion likely.

So a $5.30 bid with placement and audience adjustments stacked can have a theoretical ceiling many times higher. The preview shows that ceiling range alongside the new bid range, which is the number to sanity-check before applying a broad increase.

4. The no-op summary

The dialog counts how many of your selected rows actually need the change, and says so plainly — and where rows cannot take the action, it says why and applies to the rest.

Placement changes are entered in percentage points, with a preview of every campaign before you apply.

5. How to run it

Apply directly sends the chosen values straight to Amazon. The footer repeats it — Apply sends 49 changes to Amazon right away — and the button names the count, Apply 49 changes, so the number you confirm is the number that goes.

When nothing in the selection needs the action, the button reads No item needs a change instead of offering to apply.

Finding what is already blocked

Select Negation filters the table by whether a target is negated, and where.

AD GROUP and CAMPAIGN filter by where the negative sits and what type it is. A negative in one ad group leaves the term running everywhere else in the campaign, which is why the two levels are listed apart. Ticking several types shows rows that match any of them; adding an impact narrows that down.

IMPACT separates what the App can prove from what needs investigation:

• Fully blocked means an active negative covers the row.
• Partially blocked means the block is in place somewhere and traffic is still running elsewhere.
• Possibly blocked means no active negative was found, but a keyword that previously converted has lost traffic abnormally. It appears when the keyword, its ad group and its campaign are all enabled, the keyword recorded at least 10 orders in the last 180 days, and its impressions in the latest 14 days are either zero or at least 90% below the preceding 14 days. It sits under OTHER, next to Not negated.

Possibly blocked never includes a row with a negative the App can identify. Those rows belong to Fully blocked. Its tooltip separates What happened, Why this looks blocked and What to check, so treat the chip as an investigation prompt rather than proof of a hidden negative.

The signal is rebuilt once a day, after the Daily Ads restatement fetch completes. That fetch is scheduled for 17:00 Los Angeles time — 07:00 Vietnam time in summer, 08:00 in winter. It is not real-time, and the column can remain empty until the first nightly build after the feature is enabled.

Not negated is the inverse: everything still running, with nothing stopping it.

Last activity and Last order

The Keywords table carries two dates that help explain a sudden stop:

Column	Meaning
Last activity	The most recent day in the selected date range on which the keyword got an impression
Last order	The most recent day in the selected date range on which the keyword recorded an order

An empty cell means nothing happened in that range; widen the date range to look further back.

A date column filter offers Is equal to, Is not equal to, Is after or equal to, Is after, Is before or equal to, Is before, Has a value and No value. You can combine two conditions with And or Or, and choose each date with the date picker. Numeric columns take the same two-condition form with Greater than or equal, Less than or equal, Has a value and No value.

Excluding targeting from Automation

Select rows in Keywords, then choose Exclude from the action bar. For targeting, choose what Automation must leave alone:

• Bid only
• Enable/Pause only
• Both

Set the hold for 7, 14 or 30 days, choose a custom 1–365 days, or choose No end date.

The hold uses the same close-variant family as negative matching. Excluding hibiscus tea therefore also excludes hibiscus teas. Saving another spelling replaces the existing hold for that family; removing the exclusion removes it for the whole family.

The Excluded from Automation · N chip sits at the far right of the action bar. Select it to show only excluded rows. Automation previews and scheduled runs omit the protected action entirely.

The exclusion applies to Automation rules. It does not stop Dayparting, Agent proposals or a direct change you make in Ads Manager.

Applying an Agent proposal that changes a target’s bid or state sets this hold for you: 7 days, on that target, shown as set by Agent Proposal. It never shortens a longer hold you already set.

What you can change in bulk

Action	Entered as
Bid	An amount, a delta, or a percentage
Daily budget	An amount, a delta, or a percentage
Placement adjustment	Percentage points, or a relative percentage
Audience boost	Percentage points, or a relative percentage
State	Enable or pause
Negatives	Keywords and ASINs — see Harvest, negate and unnegate

Placement and audience are entered in percentage points, not dollars: -10 means ten points off the current adjustment, not ten percent of it.

Considerations

• Bulk actions here are direct edits. They are not bound by the Product Line’s bid ceilings, which apply to proposals rather than to you.
• A direct edit resets nothing else. A Bidding rule’s cooldown governs what rules may do, not what you may do.
• Everything applied here is recorded. See Change History.