Google Ads auto-applied recommendations: what they change and how to control them

What are Google Ads auto-applied recommendations?

They are recommendations that Google applies for you instead of waiting for someone to click "Apply". Once auto-apply is on, Google evaluates recommendations daily at account level and applies the eligible ones regularly. Its help center is clear that some types may be applied often and others never, depending on how relevant they are to the account.

Two things make them different from a normal edit:

• Nobody on your team made the change. It happened because someone opted in, maybe months ago, maybe before you took over the account.
• They keep happening. Auto-apply is a standing setting, not a one-off action, so new changes can land any day.

Google offers two bundles in the auto-apply settings, "Maintain your ads" and "Grow your business", or you can choose recommendation types one by one. Google states that auto-applying recommendations won't increase your budget, and that the Maintain your ads set never increases or changes it. Bids, keywords, targeting and ads, however, can all change.

Which recommendations can Google apply automatically?

The list in Google's help center covers four areas. The ones in bold change how much you spend or who sees your ads, so they deserve the most attention.

Recommendation types change over time, so check the current list in your own account under Recommendations › Auto-apply settings before deciding what to keep.

How do I see which recommendations were auto-applied?

Start with the summary, then go to the detail. Google keeps both views.

1. Open Campaigns › Recommendations and select Auto-apply settings.
2. Open the History tab. It lists each auto-applied recommendation type with its category, how many times it was applied, how many in the last seven days, the last time it ran and the opt-in date.
3. For the individual edits (which keyword was added, which target changed), open Change history and filter the date range around the moment a metric moved.
4. Check who opted in. Google's help center says you can confirm who enrolled the account, and when, through the account's change history.

In Change history, look at the user column. Google notes that changes made by automated or internal systems can appear under names such as "Google Ads system", and API changes can show "Google Ads API" or a tool name. If you manage many accounts, the Google Ads change history guide explains how far back it goes (the past 2 years) and what it can't tell you.

How do I turn off auto-apply recommendations?

You can switch off one type or everything. Both routes take a minute.

From the Manage tab

1. Go to Campaigns › Recommendations › Auto-apply settings.
2. Uncheck the recommendation types you don't want.
3. Select Save.

From the History tab

1. In Auto-apply settings, open History.
2. Select the recommendation you want to stop and choose Disable. Disabled items show a red dot.

You can also reach the same settings from Admin › Account settings › Auto-apply. Turning auto-apply off stops future changes; it does not undo the ones already made. To reverse those, find them in Change history, where many recent changes can be undone.

Which auto-apply options are riskiest for performance?

The riskiest are the ones that change bidding or widen targeting, because they can raise spend or shift traffic quality without anyone noticing for days.

• Bid strategy and target changes. A new target or a switch to a Maximize strategy changes how the whole campaign bids. Practitioners such as Practical Ecommerce advise keeping the bidding category off.
• Dynamic search ads and targeting expansion. Both reach queries or audiences you didn't choose, and the same article warns they can increase spend substantially.
• Broad match keywords, Search partners and Display expansion. They widen reach, so CPA and conversion quality often move.

Lower-risk candidates are the clean-up types, such as removing redundant or non-serving keywords, which tidy the account without changing who you reach. Whatever you keep on, decide it on purpose and write the decision down: "auto-apply: remove non-serving keywords only, since 1 October" is the kind of note a colleague will need six months later.

Why log auto-applied changes next to your own changes?

Because when CPA jumps on a Thursday, you need to see every change from the days before, whoever made it. Auto-applied recommendations sit in Google Ads, your team's edits sit in several tools, and site or tracking changes sit somewhere else again.

A marketing change log puts them on one timeline: the budget cut your team made, the broad match keywords Google added, the new landing page and the consent banner update. When you explain a conversion drop to a client, "Google auto-applied a switch to Maximize conversions on the 14th" is a far better answer than "the algorithm".

What to record for each auto-applied change:

• Date and account or campaign.
• Recommendation type (for example "Adjust tCPA target" or "Add broad match keywords").
• What changed, in one line, copied from Change history.
• Source: "Google auto-apply", so nobody thinks a teammate did it.
• Decision: kept, reverted, or auto-apply type disabled.

In Changeline you can log these in seconds with a tag such as "Google auto-apply" and see them next to every other channel. Start a free change log for one client and try it on the next auto-apply email.

A weekly auto-apply check for agencies

Ten minutes per account per week is enough to stay in control.

1. Open Recommendations › Auto-apply settings › History and note anything with auto-applies in the last seven days.
2. Open Change history for the same seven days and find the individual edits.
3. Compare with the account's main metrics for those days: spend, CPA or ROAS, conversions.
4. Log each change with its source and your decision.
5. Disable any type you didn't mean to leave on, and record who changed the setting and why.
6. When you take over a new account, check the opt-in dates first: auto-apply chosen by a previous agency keeps running until someone turns it off.