Performance Deterioration

Available on Negation rules — not on the word-mode rule, and not in the unnegate direction. Standard Criteria asks whether a term looks bad now. Performance Deterioration asks a harder question: was this term ever good, and has it since stopped?

What it measures

Three windows, all of which must pass:

• Historical Criteria runs on the Historical Window and confirms the term previously had value.
• Recent Criteria runs on the recent windows and confirms it is no longer performing. It must pass independently at both 60 and 30 days.

The Historical Window is configurable at 90, 120, 150 or 180 days; the two recent windows are fixed. Widths are drawn to scale for the 90-day default.

Requiring both recent checkpoints is what separates sustained deterioration from a bad month. A term that failed at 60 days but recovered by 30 does not trigger.

The App’s own example: historical 90d Orders ≥ 3, with 60d and 30d ACOS ≥ 100%.

Configuration

Field	Values
Historical Window	90 (default), 120, 150, or 180 days
Recent windows	Fixed at 60 and 30 days — not configurable
Historical Criteria	Required
Recent Criteria	Required, applied to both recent windows

This rule needs two Criteria, where every other trigger needs one. That is the part people miss when a rule will not save.

The Historical Window must be longer than the 60-day recent window, which is why 90 is the shortest option, and it is bound by your account’s Automation Window.

It does not use the rule’s Data Window

A Performance Deterioration rule scores its Historical Window, not the Data Window the rule carries. Changing the Data Window on such a rule does not change what it reads.

The evidence attached to the resulting change is the 30-day figure — the most recent of the three — because that is the state you are acting on.

The 365-day orders safeguard steps aside

Selecting this trigger switches the 365-day orders safeguard off for you, and selecting Standard Criteria switches it back on. That is deliberate, and it is not something you need to remember.

The safeguard skips any term that produced an order in the trailing 365 days, and a Historical Criteria selects terms that did exactly that. The two would cancel each other out.

Do not switch it back on

The safeguard checkbox stays editable after the App clears it. If your Historical Criteria asks for orders or sales in any form, turning the safeguard back on gives you a rule that runs, evaluates, and produces nothing at all.

When to use it

• A term that carried a product through last season and has quietly stopped converting.
• Terms that a threshold rule will not catch, because their lifetime numbers still look acceptable.
• Any case where “it was never any good” and “it stopped being good” deserve different treatment.

When not to use it

• On new terms. There is no history for the Historical Criteria to confirm.
• On an unnegation rule. Unnegation is forced back to Standard Criteria, because “was good, now bad” is the wrong shape for a rule whose job is to remove a negative. See Negation and unnegate rule.