Photo: Unsplash — free to use, no attribution required
What Auto-Apply Recommendations Actually Do
Somewhere in the last few years, Google quietly introduced a feature that automatically implements its own recommendations on your advertising campaigns. No confirmation dialog, no approval workflow, no warning email before changes go live. Google simply decides your campaign should be different, and makes it so.
The feature is called auto-apply recommendations, and it sits behind a small gear icon on the Recommendations tab of your Google Ads dashboard. When enabled, Google can make changes to your campaigns that include adding broad match versions of your carefully chosen keywords, creating new responsive search ad variations with headlines you never wrote, adjusting your target CPA or ROAS bid strategies, expanding your audience targeting to reach people outside your defined segments, enabling the Search Partners network or Display expansion, and removing keywords Google considers redundant.
The problem is not that Google offers suggestions. Suggestions can be helpful. The problem is that these changes happen automatically, without your knowledge, and they almost always push your campaigns toward spending more money on broader, less qualified traffic. Google's incentive structure makes this predictable: Google earns more revenue when you spend more, and broadening your targeting is the fastest way to increase your spend.
Most advertisers discover auto-apply changes only after noticing unexpected spikes in cost, drops in conversion rates, or unfamiliar keywords appearing in their search term reports. By then, the damage to their budget and campaign data is already done.
The Optimization Score Trap
Google presents every advertiser with an "optimization score" — a percentage from 0 to 100 displayed prominently on your Recommendations page. The implication is clear: a higher score means a better-optimized account. Naturally, advertisers and their managers want to see that number climb.
But the optimization score does not measure how well your campaigns perform against your business goals. It measures how many of Google's recommendations you have accepted. There is a fundamental difference.
When Google recommends that you switch from phrase match to broad match keywords, accepting that recommendation raises your optimization score. It also means your ads now show for a vastly wider range of search queries, many of which have nothing to do with what you sell. Your impressions go up, your clicks go up, your spend goes up — and Google's revenue goes up. Whether your conversions or ROI improve is a separate question entirely, one that the optimization score does not address.
I have audited accounts where the advertiser maintained a 95% optimization score and was simultaneously wasting 40% of their budget on irrelevant traffic. I have also managed accounts with scores hovering around 65% that delivered consistently strong returns, because we deliberately rejected every recommendation that would dilute targeting precision.
Google account representatives often pressure advertisers during calls to "improve" their optimization scores. Some agencies even use optimization score as a performance metric for their PPC managers, which creates a perverse incentive to accept recommendations that benefit Google rather than the client. If your agency reports optimization score as a KPI, that should raise questions about whose interests they are prioritizing.
The score is a compliance metric, not a performance metric. Treat it accordingly.
The 14 Auto-Apply Recommendations You Need to Check
As of early 2026, Google offers approximately 14 categories of auto-apply recommendations. The exact list changes periodically as Google adds new types, but here are the ones currently active and what each one does to your campaigns.
1. Use broad match keywords. This converts your exact match and phrase match keywords to broad match, dramatically expanding the search queries that trigger your ads. This is one of the most damaging auto-apply settings for accounts that depend on precise targeting. A business bidding on "custom software development Kerala" could suddenly appear for searches like "software engineering degree" or "free coding courses."
2. Add responsive search ad assets. Google writes new headlines and descriptions for your ads using its own AI-generated copy. The quality varies wildly, and the headlines may not reflect your brand voice, pricing, or value proposition accurately.
3. Adjust target CPA. Google raises or lowers your target cost-per-acquisition based on its own performance predictions. In practice, this usually means raising your CPA target, allowing Google to bid more aggressively and spend more of your budget.
4. Adjust target ROAS. Similar to CPA adjustments, Google modifies your return-on-ad-spend targets. Lowering your ROAS target tells the system it is acceptable to generate less efficient returns, which increases spend volume.
5. Set target CPA for campaigns using Maximize Conversions. Google adds a specific CPA target to campaigns that were running without one, which changes how the bidding algorithm behaves and can significantly alter your cost structure overnight.
6. Upgrade to broad match. A variant of the first setting, this specifically targets campaigns using Smart Bidding and pushes them to broad match, arguing that Smart Bidding will compensate for the reduced targeting precision. In my experience, it does not compensate fully.
7. Remove redundant keywords. Google identifies keywords it considers duplicative and removes them. This is occasionally useful but can also remove keywords that serve different match type strategies or exist intentionally for bid management purposes.
8. Use optimized ad rotation. This overrides your manual ad rotation settings and lets Google decide which ad variations to show most frequently. If you are running deliberate A/B tests, this destroys your testing methodology.
9. Expand your reach with Google Search Partners. Enables the Search Partners network, placing your ads on third-party search engines and partner sites. Traffic quality from Search Partners is generally much lower than Google.com traffic.
10. Use optimized targeting. Expands your display and video campaign targeting beyond your defined audiences, showing ads to people Google's algorithm considers likely to convert, even if they fall outside your specified demographic or interest segments.
11. Improve your responsive search ads. Google modifies existing RSA assets, replacing headlines and descriptions it considers underperforming. This can alter messaging you have carefully crafted for specific audience segments.
12. Add keywords. Google suggests and automatically adds new keywords to your campaigns based on search query data. These keywords may be relevant, but they often include broad terms that drive unqualified traffic.
13. Create new ad versions. Beyond modifying existing ads, Google may create entirely new ad variations in your campaigns, pairing headlines and descriptions in combinations you have not reviewed or approved.
14. Bid more efficiently with Maximize Conversions. Google switches your manual or enhanced CPC bidding to its automated Maximize Conversions strategy. This removes your direct bid control and hands it entirely to Google's algorithm.
Of these 14 categories, I consider broad match expansion, audience targeting expansion, bid strategy changes, and CPA/ROAS adjustments the most dangerous for most accounts. Redundant keyword removal and disapproved ad fixes are the only categories I occasionally leave enabled, and only for accounts I monitor weekly.
How to Find and Disable Auto-Apply Settings
Disabling auto-apply takes less than five minutes, but finding the settings is not intuitive — which may be by design. Here is exactly where to go.
Step 1: Log into your Google Ads account and click Recommendations in the left-hand navigation menu. If you are using the new Google Ads interface, it may also appear under "Insights and reports."
Step 2: Look at the top right corner of the Recommendations page. You will see a small gear icon or a link that says "Auto-apply" — it is easy to miss, which is part of the problem. Click it.
Step 3: You will now see a list of all recommendation types with toggle switches next to each one. Some will be turned on by default. Google groups them under headings like "Ads and assets," "Bidding," "Keywords and targeting," and "Measurement."
Step 4: Review every single toggle. Turn off anything you do not explicitly want Google to change without your permission. For most advertisers, this means turning off everything. Click Save when finished.
Step 5 — this is critical: If you manage multiple accounts through a Google Ads Manager (MCC) account, you need to repeat this process for every individual account. Auto-apply settings do not cascade from the MCC level. Each child account has its own independent auto-apply configuration. I have seen agencies disable auto-apply on their main dashboard but leave it running on dozens of client accounts underneath.
Step 6: Set a calendar reminder to check these settings monthly. Google has a pattern of re-enabling auto-apply settings after platform updates, and new recommendation categories sometimes default to enabled when they launch.
Real Damage: What Happens When Auto-Apply Runs Unchecked
Theory is useful, but concrete examples illustrate why this matters. These are based on real situations from accounts I have audited, with identifying details changed.
The local business that went global overnight. A web development company in Kochi was running tightly targeted campaigns bidding on terms like "web development company Kochi" and "website design Kerala." After broad match auto-apply kicked in, their ads started appearing for searches like "web development tutorial," "learn HTML free," and "web developer salary in USA." Their monthly spend jumped from Rs 15,000 to Rs 48,000 in three weeks, while their qualified lead volume actually decreased because the budget was being consumed by irrelevant clicks from people looking for jobs and tutorials, not agencies to hire.
The e-commerce store with vanishing ROAS. An online retailer selling handloom products had a target ROAS of 400% and was consistently hitting it. Auto-apply adjusted their ROAS target down to 250%, and the algorithm responded by bidding more aggressively on broader audiences. Monthly spend increased by 60%, but revenue only grew by 15%. Their effective ROAS dropped to 280%, and it took two billing cycles before anyone noticed the auto-apply change buried in their change history.
The B2B company with the ruined A/B test. A SaaS company was running a careful A/B test comparing two different value propositions across their ad copy. Auto-apply's "improve your responsive search ads" feature rewrote several of their test headlines, contaminating the data and rendering six weeks of testing meaningless. They had to restart the entire experiment from scratch.
The budget that doubled without permission. A small business owner in Thrissur had set conservative daily budgets across three campaigns totaling Rs 500 per day. After auto-apply enabled Maximize Conversions bidding and adjusted CPA targets upward, the algorithm was spending Rs 1,800 to Rs 2,200 per day. The business owner only discovered this when they reviewed their credit card statement at the end of the month and found a Google Ads charge four times higher than expected.
In each case, the changes appeared in the account's change history log, but they were labeled as "auto-applied recommendation" in small text that is easy to overlook if you are not specifically looking for it.
The Recommendations That Actually Help (Sometimes)
Not every Google recommendation is designed to extract more money from your account. Some are genuinely useful, and dismissing everything reflexively is not the right approach either.
Removing truly redundant keywords can simplify account management when you have multiple keywords with identical match types competing against each other in the same ad group. However, verify each suggestion manually — sometimes keywords that look redundant serve intentional structural purposes.
Fixing disapproved ads and extensions is always worth doing. If your ad or sitelink was disapproved due to a policy violation or broken landing page URL, you want to know about it and fix it. This recommendation type is purely diagnostic and does not change your targeting or spending.
Upgrading to responsive search ads from older expanded text ads is generally reasonable, since Google has phased out creation of new expanded text ads. The recommendation to create RSAs where you only have legacy ad formats is usually sound, though you should still write the headlines and descriptions yourself rather than accepting Google's auto-generated copy.
Adding negative keywords is one recommendation category that can genuinely save you money by preventing your ads from showing on irrelevant searches. However, review each suggested negative keyword carefully — Google occasionally suggests negating terms that are actually relevant to your business.
The pattern is clear: recommendations that help you fix broken things or clean up account hygiene tend to be genuinely useful. Recommendations that expand targeting, raise bids, or reduce your control over the account tend to benefit Google's revenue more than your performance.
Setting Up Proper Campaign Monitoring
Disabling auto-apply is the immediate fix, but ongoing monitoring is what prevents future problems. Here is a practical monitoring framework that does not require expensive third-party tools.
Check change history weekly. In Google Ads, go to the "Change history" section and filter for changes made by "system" or "auto-applied." This shows you exactly what Google has changed without your input. Make this a standing item on your weekly campaign review agenda.
Set up custom alerts for spend anomalies. Under Tools and Settings, you can create automated rules that send email notifications when daily spend exceeds a threshold you define. Set alerts at 120% and 150% of your expected daily budget. If something triggers an unexpected increase, you will know within 24 hours rather than at the end of the billing cycle.
Monitor keyword lists for unfamiliar additions. Export your keyword list monthly and compare it against the previous month's export. Any keywords you did not add manually deserve investigation. A simple spreadsheet comparison using VLOOKUP or a diff tool catches additions quickly.
Track optimization score changes — as a warning signal, not a goal. If your optimization score suddenly jumps upward without you making changes, it likely means you accepted a recommendation you did not intend to, or auto-apply was re-enabled. Conversely, if it drops, Google probably added new recommendation types. Either way, a score change is a trigger to review your account settings.
Review search term reports bi-weekly. The search term report shows you the actual queries that triggered your ads. If you see irrelevant queries that do not match your keyword targeting intent, something has changed in your match types or audience settings. This report is your early-warning system for targeting drift.
Use Google Ads scripts for automated monitoring. If you are comfortable with JavaScript, Google Ads scripts can automatically check for auto-applied changes and send you alerts. A script that runs daily and emails you whenever a "system" change appears in the change history log costs nothing to implement and provides consistent oversight without manual effort.
For agencies managing multiple accounts, third-party tools like Optmyzr, Adalysis, or WordStream offer automated monitoring dashboards that flag unexpected changes across client accounts. The cost is typically justified for agencies managing more than ten accounts, where manual checking becomes time-consuming.
Frequently Asked Questions
How do I turn off Google Ads auto-apply recommendations?
In your Google Ads account, click Recommendations in the left sidebar, then click the gear icon labeled "Auto-apply" in the upper right corner. You will see a list of all recommendation types with toggle switches. Turn off each one individually. You must repeat this for every account under your MCC, as settings do not cascade from manager accounts to child accounts. After disabling, check these settings monthly — Google has been known to re-enable them after platform updates.
Does declining Google's recommendations hurt my ad performance?
No. Dismissing recommendations only lowers your optimization score, which is a Google-created metric that has no effect on your Quality Score, ad rank, auction eligibility, or actual campaign delivery. Google has confirmed that optimization score is purely advisory. Many high-performing accounts I manage deliberately maintain scores between 60% and 75% because we reject every recommendation that would waste budget on overly broad targeting. Your actual performance metrics — cost per acquisition, return on ad spend, conversion volume — are what matter, and those are determined by your campaign structure, not your compliance score.
What is Google Ads optimization score and should I care about it?
Optimization score is a percentage from 0 to 100 that reflects how many of Google's recommendations you have adopted. It does not measure campaign effectiveness, profitability, or return on investment. A 100% score often means you have accepted every recommendation including broad match expansion, raised bid targets, and audience broadening — changes that typically increase spend without proportional improvement in results. Use your own KPIs as your performance benchmark. The optimization score is useful only as a notification that Google has new suggestions for you to evaluate, not as a measure of account health.
How often does Google change auto-apply settings?
Google introduces new recommendation categories and has been observed re-enabling previously disabled auto-apply settings roughly two to four times per year, usually coinciding with major platform updates. New recommendation types sometimes default to auto-apply enabled when they first launch, meaning your previously clean settings may acquire new active toggles without any action on your part. Check your auto-apply configuration at least once per month, and always within a week of receiving any Google Ads notification about new features or interface changes.
Can I accept some auto-apply recommendations and reject others?
Yes, and this selective approach is what I recommend for most accounts. The auto-apply settings page lets you toggle each recommendation type independently. Safe candidates for auto-apply include removing genuinely redundant keywords and fixing disapproved ad destinations. Dangerous categories to keep disabled include broad match keyword expansion, bid strategy changes, CPA and ROAS target adjustments, and audience targeting expansion. Review each category based on your specific campaign goals and targeting requirements rather than using a blanket approach.