A business owner in a checked shirt reviewing work on a laptop at a desk covered with papers and a coffee mug
    ← Back to Blog
    Google Ads14 min read

    Google Ads Recommendations: Which Ones Should You Accept?

    November 20, 2026 · Joshua Filani

    Accept the cleanup recommendations, review the bidding ones against your real lead data, and dismiss most of the reach-expansion ones. Leave auto-apply off. Google says itself that you can reach a 100% optimization score just by dismissing recommendations, so the score is a to-do list Google wrote for you. It doesn't grade your account.

    Quick answer

    Google's recommendations fall into two groups. Some tidy up your account (remove dead keywords, fix conflicting negatives, upgrade conversion tracking), and those are usually safe. Others widen where your money goes (broad match, search partners, Display expansion, dynamic search ads), and for a local service business on a tight budget those usually waste money. A 2024 Optmyzr study of 17,380 accounts found only 5.5% accepted Google's suggestions. We'd review every card yourself and turn auto-apply off.

    Want a free audit of your own site?

    Send us your website and we will review your Google, website and lead tracking, then email you what to fix first. No pitch.

    Free. Read-only. We change nothing on your site.

    What is the Google Ads optimization score, really?

    It's Google's estimate of how well your account is set up to perform, not a measure of how well it is performing. Google's help page describes it as "an estimate of how well your Google Ads account is set to perform," scored from 0 to 100%, and says it's calculated from your settings, your stats and "the relevant impact of available recommendations" (Google Ads Help, 2026).

    The detail that matters most sits lower on that same page. Google says your account "can reach an optimization score of 100% by applying or dismissing all recommendations." Dismissing a card raises the score just as much as applying it does. So a 100% account could have applied nothing at all. It's a measure of how much of Google's list you've dealt with.

    Two more facts from the same page. The score is not used in your Quality Score, so a low optimization score doesn't raise your cost per click. And a dismissed recommendation "can reappear after a certain period if it becomes relevant," which is why the same broad match card keeps showing up month after month.

    The score also follows your bid strategy. If you're on Maximize Clicks and reporting conversions, Google says you may see a recommendation to switch to a Smart Bidding strategy like target CPA. Google's starting assumption is that more automation is better. Sometimes that's true. Often it isn't for a shop running $1,500 a month.

    Why does Google push recommendations so hard?

    Part of the answer is in the Partner program. Agencies that want the Google Partner badge need their manager account to hold "a minimum optimization score of 70%," and Google says it checks that daily (Google Partners FAQs, 2026). Google also says partners "are not required to adopt all recommendations" and can apply or dismiss them based on their own judgment.

    When Google first tied the badge to the score, Search Engine Land reported that agencies falling under 70% would get 60 days to recover before losing the badge, and that Google cited internal data saying advertisers who raised their score by 10 points saw "a 10% increase in conversions, on average" (Search Engine Land, 2020). That article is from 2020. Optmyzr's 2024 write-up describes Google's claim as a 15% lift for 10 points, so the number Google quotes seems to have changed over time. We couldn't find a current, dated Google source for either figure.

    None of this is sinister. Google built a system that suggests changes, and it rewards agencies for engaging with them. Just keep in mind that a recommendation tells you what Google's models expect will increase conversions or reach. Your margin on a water heater swap isn't part of that model, and neither is your crew's capacity in the second week of July.

    Does a higher score mean more calls?

    There's a correlation. Nobody has shown that raising the score causes better results. The largest outside look we found is Optmyzr's 2024 study of 17,380 Google Ads accounts that had run at least 90 days and spent between $500 and $1 million a month (Optmyzr, 2024).

    5.5%of 17,380 accounts accepted Google's recommendations (Optmyzr, 2024)
    19%reached a 90+ score without accepting recommendations (Optmyzr, 2024)
    186%higher ROAS for 90-100 scores than sub-70 scores (Optmyzr, 2024)

    Here's how we read it. The accounts with the best results were the 333 that both accepted suggestions and held a 90-100 score. They also had the cheapest overall cost per acquisition, though not the cheapest clicks or the best conversion rates. Nearly one in five accounts got to 90+ without accepting anything, which fits Google's own point that dismissing counts the same as applying. And 32% of accounts sat under 70.

    What the study can't tell you is direction. Well-run accounts tend to have clean structure, working conversion tracking and someone checking them weekly, and all of that pushes the score up on its own. ROAS is also an ecommerce number. Most home service accounts don't pass revenue back to Google, so the 186% figure doesn't map neatly onto an HVAC or plumbing account. We'd treat the study as a sign that engaging with the list helps, not as proof that clicking Apply helps.

    Should you turn on auto-apply?

    No, not for anything that spends money in new places. Google's auto-apply feature applies your chosen recommendation types "regularly," and you can only turn it on at the account level (Google Ads Help, 2026). The eligible list includes adding keywords, adding broad match keywords, expanding to Google search partners, using Display expansion, using targeting expansion, adding dynamic search ads and changing bid strategies.

    That's a lot of control to give away. A few points from Google's own page are worth knowing:

    • Auto-apply won't raise your budget. Google says budget raising "isn't included in this iteration."
    • Starting January 26, 2026, the "Add responsive search ads" recommendation no longer creates new ads automatically, and Google removed it from the auto-apply settings.
    • Changes show in the History tab and in Change History, so you can see what was done after the fact.

    If you or a past agency turned it on, here's how to check. Google's instructions are to open Recommendations, choose Auto-apply settings, uncheck each type and save (Google Ads Help, 2026). The same page says Change History shows who enrolled the account and when. If you do keep anything on, turn on campaign maintenance notifications in your preferences so you get the weekly email summary of auto-apply activity.

    The one exception we're comfortable with is "Upgrade your conversion tracking." It changes how you measure, not where you spend, and better measurement makes every other decision in the account more trustworthy.

    The verdict, one recommendation at a time

    This table covers the recommendation types Google lists as eligible for auto-apply, plus the budget card. The verdicts are ours, written for a local service business spending roughly $1,000 to $10,000 a month on Search.

    RecommendationWhat it changesOur verdictWhen we'd say yes
    Remove non-serving keywordsDeletes keywords with no impressionsAcceptAlmost always. Check first that it isn't a seasonal term like "evaporative cooler repair."
    Remove redundant keywordsRemoves duplicates that compete with each otherAccept after a lookWhen the kept version is in the right ad group
    Remove conflicting negative keywordsRemoves negatives that block your own keywordsReviewWhen the block was an accident. Some negatives are deliberate.
    Upgrade your conversion trackingTurns on enhanced conversionsAcceptNearly always, if your privacy notice covers it
    Use optimized ad rotationShows the ads expected to perform best more oftenAcceptUnless you're running a deliberate ad test
    Add audience segments for reportingAdds audiences in observation mode onlyAcceptHarmless. It adds data, not reach.
    Improve your responsive search adsAdds headlines pulled from your site and adsReviewWhen the new headlines are accurate. Watch for prices, licenses and service areas you don't offer.
    Add keywordsAdds new search termsReview each oneWhen the term is a real job you want. Never on auto.
    Switch bid strategy / set target CPAMoves you to Smart BiddingDependsOnly when conversions are real calls and forms, not page views
    Add broad match keywordsMatches far looser searchesUsually dismissOnly with trustworthy conversion tracking and Smart Bidding
    Expand to Google search partnersShows ads on non-Google sitesUsually dismissTest in one campaign and segment the results first
    Use Display expansion / targeting expansionPushes Search budget into display placements and wider audiencesDismissRarely, for a local Search campaign
    Add dynamic search adsGoogle writes headlines from your websiteUsually dismissOnly with a large, clean site and strong negatives
    Raise your budgetIncreases daily spendDo the mathWhen lost impression share is costing booked jobs you have crews for

    Which recommendations are safe to accept?

    The safe ones remove waste or improve measurement. Removing non-serving keywords is housekeeping. Google lists it under the auto-apply "Keywords and targeting" group alongside removing redundant keywords and conflicting negatives (Google Ads Help, 2026). These cards rarely move money somewhere new.

    Conflicting negatives need one minute of thought before you click. Say you added "free" as a negative to block "free estimate" searches, and you also bid on "free AC estimate" because you actually offer one. The card is right that there's a conflict, but only you know which side should win.

    Enhanced conversions is the card we'd push owners to accept. Most service accounts we see are under-counting leads, and Smart Bidding can only optimize what it can see. If you haven't confirmed that your tracking counts real calls and form fills, our post on Google Ads conversion tracking for service businesses walks through it. Do that before you accept anything in the bidding section.

    Audiences "for reporting" are also safe. Observation mode means Google records how those people behave without narrowing or widening who sees your ad. You get data for free.

    Want a free audit of your own site?

    Send us your website and we will review your Google, website and lead tracking, then email you what to fix first. No pitch.

    Free. Read-only. We change nothing on your site.

    Which recommendations should a local service business dismiss?

    These are the ones that spend money in new places: broad match, search partners, Display expansion and dynamic search ads. Each can work in the right account. In a small account with thin conversion data, each one is a reliable way to buy clicks from people who were never going to book.

    Broad match. Google's own guide says "it's critical to use Smart Bidding with broad match," because each looser search needs a bid that reflects its context (Google Ads Help, 2026). Take Google at its word. If you're on manual or Maximize Clicks, or your "conversions" include page views, broad match has nothing reliable to steer by. We go deeper in our broad match post.

    Search partners. Google includes search partners "by default" in new Search campaigns, and you can opt out in campaign settings (Google Ads Help, 2026). Before you accept a card telling you to expand, segment your current campaign by network and compare cost per conversion. Here's a hypothetical to show why. On a $3,000 month, if search partners take 15% of spend ($450) and produce one lead while Google Search produces 25 leads on $2,550, your Google Search cost per lead is $102 and your partner cost per lead is $450. Your numbers will differ. The point is to check yours before expanding.

    Display and targeting expansion. A homeowner searching "AC not blowing cold" is in a very different spot from someone who sees a banner while reading a recipe. We keep Search budgets on Search.

    Dynamic search ads. These write headlines from your website content. On a 12-page contractor site with an old blog post about a discontinued service, that can go wrong in ways that embarrass you.

    What about the bidding recommendations?

    They're only as good as your conversion data. Google's optimization score page says the score works out your campaign's goal from your bid strategy, and if you aren't on Smart Bidding, it infers a recommended Smart Bidding strategy from your data (Google Ads Help, 2026). So the bidding cards assume your conversion count reflects real business.

    Here's the test we use. Pull last month's conversions in Google Ads and compare them to the leads your office actually logged from paid search. If Google shows 80 conversions and your CSR logged 31 real calls, a target CPA built on 80 will chase the wrong things. Fix tracking first, then let Smart Bidding have a go.

    "Adjust your CPA targets" deserves extra caution. Google may suggest raising your target to win more volume. That can be right in peak season. But a heat wave in Phoenix floods you with calls whatever your target is, and a looser target in July can mean paying more for leads you can't schedule until next week. If your crews are booked out, a higher target buys you nothing.

    The budget card isn't on auto-apply, and that's the right call. Look at "search lost impression share (budget)" on your best campaign. If it's high and you have crew capacity, raising budget is often the cleanest growth lever in the account. If you're already turning work away, it isn't.

    Should you take the call from a Google Ads rep?

    Take it, listen, and don't apply anything on the phone. Google does offer free, one-on-one help from Google Ads Experts (Google Ads Help, 2026). A rep is usually working from the same recommendations page you can see. They have the same blind spots, because they don't know your close rate, your average ticket or which zip codes your techs refuse to drive to in August.

    We'd ask three things on any rep call. Which specific recommendations are they suggesting? What do they expect each to change in cost per conversion, not clicks or impressions? And will they send the list by email so you can review it with whoever manages the account? A good suggestion still sounds good in writing the next morning.

    What we would do

    If we opened your account this week, here's the order we'd work in.

    First, check auto-apply. Open Recommendations, then Auto-apply settings, and turn off every type except "Upgrade your conversion tracking." Check the History tab to see what has already changed, and when.

    Second, verify conversion tracking against your call log for one month. Until those numbers roughly agree, dismiss every bidding recommendation.

    Third, clear the housekeeping cards: non-serving keywords, redundant keywords, and conflicting negatives once you've thought about them. That alone usually moves the score a fair amount, and it's honest progress.

    Fourth, dismiss broad match, search partners, Display expansion and dynamic search ads, unless you have a specific reason and a test plan. Don't mind it when they come back. Google says dismissed cards can reappear.

    Fifth, put a 20-minute recommendations review on the calendar once a month. Read each card and decide. Don't treat the score as the goal.

    When not to hire anyone for this: if you spend under about $1,000 a month and you're willing to spend that monthly 20 minutes, you can handle the recommendations page yourself with the table above. Paying an agency a management fee that's a big share of your ad spend rarely makes sense at that size. If you'd rather we look first, our free audit checks auto-apply, tracking and network settings, and our Google Ads agent flags new recommendations for a human to approve instead of applying them.

    For an outside view, Paid Media Pros' short video is worth a few minutes. Watch for their take on how much weight the score deserves compared with your actual conversion numbers. Credit to the Paid Media Pros channel.

    Questions owners ask

    Does a low optimization score hurt my ad rankings or raise my costs?

    No. Google states that optimization score is not used in your Quality Score, which is the measure that affects ad rank and cost per click. A low score just means there are recommendations you haven't applied or dismissed yet.

    Can I get to 100% without applying anything?

    Yes. Google says an account can reach 100% by applying or dismissing all recommendations. Dismissing a card that doesn't fit your business counts the same as applying it.

    Should I turn on auto-apply recommendations?

    We'd leave it off, except for "Upgrade your conversion tracking." Auto-apply can add broad match keywords, expand to search partners and change bid strategies with no review. It won't raise your budget, but it can change where that budget goes.

    Why does my agency care so much about optimization score?

    Google Partner agencies need a manager-account optimization score of at least 70% to keep the badge, according to Google's Partners FAQ. That's a fair reason for them to work through the list. It isn't a reason to apply cards that don't fit your business, and Google says partners can dismiss them.

    Is broad match ever a good idea for a service business?

    It can be, if your conversion tracking counts real calls and forms and you're on Smart Bidding. Google's own guide calls Smart Bidding "critical" with broad match. Without both, we'd dismiss the card.

    How do I see what auto-apply already changed?

    Open the Recommendations page, choose Auto-apply settings, then the History tab. Change History also shows who enrolled the account and when, and you can turn on weekly email summaries in your notification preferences.

    Want a second set of eyes on your recommendations page?

    We'll check which recommendations have been auto-applied, whether your conversions match your real leads, and where your budget is going by network. Start with your free audit, see pricing, or call (623) 343-3141.

    Sources

    No. Google states that optimization score is not used in your Quality Score, which is the measure that affects ad rank and cost per click. A low score just means there are recommendations you haven't applied or dismissed yet.

    Yes. Google says an account can reach 100% by applying or dismissing all recommendations. Dismissing a card that doesn't fit your business counts the same as applying it.

    We'd leave it off, except for "Upgrade your conversion tracking." Auto-apply can add broad match keywords, expand to search partners and change bid strategies with no review. It won't raise your budget, but it can change where that budget goes.

    Google Partner agencies need a manager-account optimization score of at least 70% to keep the badge, according to Google's Partners FAQ. That's a fair reason for them to work through the list. It isn't a reason to apply cards that don't fit your business, and Google says partners can dismiss them.

    It can be, if your conversion tracking counts real calls and forms and you're on Smart Bidding. Google's own guide calls Smart Bidding "critical" with broad match. Without both, we'd dismiss the card.

    Open the Recommendations page, choose Auto-apply settings, then the History tab. Change History also shows who enrolled the account and when, and you can turn on weekly email summaries in your notification preferences.
    Tags:google adsoptimization scoreauto-apply recommendationsbroad matchservice business

    How Valley Can Help

    We Help Businesses Like Yours Get More Leads — and Close More of Them

    The Valley Marketing Group is a Phoenix-based marketing agency specializing in AI-powered lead generation, paid advertising, and web development for local service businesses.

    • Google Ads & paid search — campaigns built to generate qualified leads, not just clicks
    • AI phone receptionist — never miss a call or lead while you're on the job
    • Website design & development — WordPress, Webflow, Shopify, WooCommerce
    • SEO content & local search — rank for the searches your customers are already making
    Ready to talk now? Get a free 24-hour audit instead
    Want more like this?
    Our monthly newsletter — real marketing tips for service businesses, no fluff. Unsubscribe anytime.
    Free · One email a month · No spam