The owner of a plumbing company logs into Google Ads on a Monday morning and sees a yellow notice next to his best ad: Ad Strength, Poor. Google suggests adding more headlines and removing a few pins. The ad in question has "Drains Cleared for $149 Flat" locked into the first headline, because that line is the reason his phone rings with the right kind of customer. He follows the advice anyway. Within a week the rating says Excellent and the calls say something else.
Here is the short answer to what happened. Ad Strength measures how much varied material you handed Google to build ads from: headlines, descriptions, keywords, sitelinks. It does not measure whether the ad sells. But when two of your ads compete inside the same ad group, Google shows the higher-rated one more often, and a study of more than 5,900 Google Ads accounts found that the ad Google favored converted worse more often than it converted better.
That gap is what this article is about: what the rating actually checks, what it quietly steers, why the "worse" ad so often wins, and what to change in your account so the meter stops making decisions your customers never voted on.
Ad Strength Counts Your Headlines and Keywords. It Never Sees a Sale.
Every responsive search ad in your account carries a rating from Poor to Excellent. Google's help page lists what goes into it, and every item is about the ingredients, not the result:
- Keyword relevance. Whether your most important keywords appear in your headlines and descriptions, each keyword whole inside a single headline.
- Quantity. Google lets you supply up to 15 headlines and 4 descriptions, and the rating rewards you for using them.
- Variety. Headlines that say genuinely different things score better than five ways of saying "Fast Plumber."
- Sitelinks. Six or more, counted across the account, campaign and ad group.
Notice what is missing. The help page says Ad Strength "does not measure external factors, such as the benefits of adding new keywords or the quality of your landing pages and offerings." It never looks at a phone call, a form fill or a sale. It also counts text Google writes for you the same as text you wrote yourself, if you have Google's text customization switched on.
Google is direct about what the rating is not. The same page calls it a "common misconception" to treat it as more than a feedback tool for asset variety, and says the rating "doesn't directly influence your ad's serving eligibility." Keep that exact wording in mind, because it is narrower than it sounds. Eligibility is whether an ad is allowed to enter the auction at all. It says nothing about which of your two eligible ads Google picks to send in.
Then the same help page makes the sales pitch. Advertisers who improve Ad Strength from Poor to Excellent "see 15% more conversions on average," according to Google's internal data from August 15 to August 20, 2025. That is five days of Google's own numbers, with no detail on which advertisers or what else changed. Google's own responsive search ads video, published in August 2026, gives a different figure for the same move: "12% more conversion rates." Two numbers from one company for one claim is a good reason to treat both as marketing.
One more line on that page deserves attention. To raise your rating, Google suggests you "consider opting into text customization," the setting that lets Google write ad copy from your website. Text customization is part of AI Max, and it is the piece we have told owners to turn on last, if at all. The meter, in other words, rewards you for handing Google more of the pen.
Two Ads in One Ad Group? Google Shows the Higher-Rated One More Often.
If Ad Strength never touched the auction, it would be a harmless writing aid. The question is what happens when you run two ads side by side, which Google itself recommends: at least two responsive search ads per ad group, each rated Good or Excellent.
Adalysis, an ad-testing software company, answered that with data from more than 5,900 Google Ads accounts, published in October 2026. The cleanest part of the study isolates ad groups that had exactly two active responsive search ads with different ratings. Both ads shared the same keywords, bids, budget and targeting, so the only real difference between them was the ads themselves. Branded ads were excluded, every ad group had working conversion tracking, and every ad had run at least 30 days with at least 250 impressions.
In those ad groups, the higher-rated ad got more impressions 63% of the time. The bigger the gap in rating, the more lopsided it got:
- Excellent against Poor: the Excellent ad got more impressions in 78% of ad groups.
- Two ratings apart (Excellent against Average, or Good against Poor): 66%.
- Good against Average, the closest matchup: 59%.
Adalysis ran a similar study in 2022, when the higher-rated ad won on impressions 56.8% of the time. Four years later it is 63%. The lean is getting stronger, not weaker.
So how can Google say the rating does not decide eligibility while the higher-rated ad keeps getting shown? Both things can be true at once. Look at Google's help page on ad rotation. The default setting, Optimize, "optimizes your ads for clicks in each individual auction" and delivers the ads "expected to attract more clicks" more often. If you use Smart Bidding, which most accounts now do, Google uses that setting automatically. The things that raise Ad Strength, more headlines, more keywords, fewer pins, are the same things that give Google's system more combinations to chase clicks with. The rating itself may stay out of the decision. The work that earns it is exactly what Google's click-chasing rewards.
The Ad Google Showed Less Often Converted Better in 55% of Ad Groups
Now the part that matters to your bank account. In those same two-ad matchups, Adalysis compared what each ad actually did once it was shown.
- Click-through rate: the higher-rated ad was ahead in 52.5% of ad groups. Close to a coin flip.
- Conversion rate: the higher-rated ad was ahead in only 45%. The lower-rated ad converted better in 55% of ad groups, by 9.2% on average. If the lower-rated ad converted 5% of its clicks, the higher-rated ad converted about 4.6%.
- Cost per conversion: the lower-rated ad was cheaper in 51.5% of ad groups, and when it won, it won by a lot. Across all the ad groups, the higher-rated ads ran an 11.8% higher cost per conversion. Where the lower-rated ad paid $100 for a lead, the higher-rated ad paid $111.80.
The four-year comparison makes the problem plain. From 2022 to 2026, the higher-rated ad went from winning impressions 56.8% of the time to 63%, and from winning on click-through rate 48.5% of the time to 52.5%. Its conversion-rate record did not move: 45.8% then, 45.1% now. Google has gotten better at matching the rating to clicks. It has not gotten any better at matching it to customers.
A second dataset points the same way. Optmyzr, another Google Ads software company, analyzed roughly 20,000 accounts in April 2026 and found "no reliable correlation between Ad Strength and ad efficiency." In its data, ads rated Excellent had the worst cost per conversion of any group, $28.68, and the lowest conversion rate, 4.97%. Ads rated Average had the best: $12.43 per conversion and a 12.65% conversion rate.
Two honest caveats. Both studies are observational, meaning they describe what happened in real accounts rather than running controlled experiments. And Optmyzr's comparison is across accounts, so a cheap Average ad in one business is being compared with an expensive Excellent ad in another. The Adalysis head-to-head is the stronger evidence for exactly that reason: same ad group, same keywords, same budget. Neither study says a good rating is bad. Both say a good rating is not the same thing as a good ad.
Pinning Your Price or License Number Lowers Ad Strength. It Often Raises the Conversion Rate.
Adalysis traced most of the "worse ad wins" result to one setting: pinning. In most of the mixed-rating ad groups, one ad was pinned and the other was not.
Pinning means telling Google that a particular headline must always appear in a particular spot. Normally Google mixes your 15 headlines into whatever combination it predicts will get a click. With 15 headlines and 4 descriptions, Adalysis counts 50,235 possible combinations. A pin takes some of that choice away and guarantees your most important line is always the first thing a searcher reads.
Local businesses pin for good reasons, and almost all of them are about filtering the click rather than chasing it:
- A flat price. "Drains Cleared for $149 Flat" tells the bargain hunter and the person expecting $49 not to click.
- A license or credential. Some states expect a contractor's license number in advertising, and a pinned line makes sure it is always there.
- A service area. "Serving North Tulsa Only" stops paid clicks from people forty minutes outside your drive radius.
- A policy. "No Weekend Calls" or "Commercial Only" turns away the customer you cannot serve before they cost you a click.
Ad Strength punishes every one of those. Adalysis found that once you pin headlines in two different positions, the ad cannot reach Excellent at all, and in practice it usually lands at Poor or Average. Pinning only in position one still leaves Excellent possible, but rare. The meter reads a pin as lost variety. Your customer reads it as the most useful line in the ad. Google's own advice is to "use pinning sparingly" and, if you must pin, to pin several different versions to the same position.
The pinned ads convert better because they are doing a different job. They are written to qualify the buyer, which costs some clicks and saves money on the clicks that would never have booked. The rating is built to reward broad appeal, so a well-qualified ad will always look weaker on the meter than it is.
There is a smaller-account wrinkle here too. In its companion study on pinning, Adalysis found that in low-volume accounts with nothing pinned, Google keeps serving the headlines that repeat the search term and rarely shows the benefit, offer and call-to-action lines at all. The system needs a lot of impressions to learn which of 50,000 combinations sells, and a small account never gives it enough. A pin skips that wait.
To be fair to the meter, the same Adalysis data shows that for ads with nothing pinned, a higher rating really does come with a higher click-through rate and more conversions per impression, at every step from Poor to Excellent. And the two studies disagree about how far to take pinning. Optmyzr found partial pinning performed best and pinning every position performed worst on cost per conversion. Adalysis found fully pinned ads converted best but noted those advertisers are also the most active testers, which could explain the result on its own. Where both studies land in the same place: pinning one or two positions on purpose, and leaving the rest open, is a defensible way to build a small business's ad. Our recommendation is to pin your one non-negotiable line in position one and let Google rotate everything else. Pinning is about the ad being the only thing your customer actually reads, and you get to decide what that one thing is.
Google Gives Your Two Ads Unequal Airtime. A Small Account Never Gets Enough Clicks to Pick a Winner.
Here is where a small account gets hurt twice. Most owners run two ads in an ad group because they think the ads are being tested against each other. They are not, at least not fairly. If one of them is pinned and the other is not, the unpinned one will usually get most of the impressions, and the pinned one, often the better seller, gets starved.
Run the numbers for a plumbing company getting 30 leads a month from 600 clicks, a 5% conversion rate. Take the gap Adalysis found, 5% against about 4.6%. To tell those two apart with ordinary statistical confidence, each ad needs roughly 40,000 clicks. At 600 clicks a month, split unevenly, that is more than a decade. The account will never collect enough evidence to show the lower-rated ad is better, and Google will keep favoring the higher-rated one the whole time.
So stop treating two ads in an ad group as a test. If you want a real answer, Google Ads has one built in: the Experiments page, where an ad variation shows half of your searchers one version and half the other, so the split is fair from the start. Even then, small accounts should test big differences, like a pinned offer against no pin, rather than "Call Now" against "Call Today." We walk through how to set that up, and how long to let it run, in Google Shows You Conversions for Every Headline. That Is Not a Test Result.
If you still bid manually, you have one more lever. Manual bidding lets you set ad rotation to "Do not optimize," which spreads impressions more evenly. Google warns it is "not recommended for most advertisers," and even then an ad with weaker quality can show lower on the page. It is fairer than the default, not perfectly fair.
What to Change in Your Google Ads Account This Week
None of this means ignoring Ad Strength. It means demoting it from a grade to a writing checklist. Here is the order we would work in.
- Put the rating next to the results. In the Ads table, open Columns, then Attributes, and add "Ad strength." Add conversions and cost per conversion next to it, and note which ads have pins. Now you can see, ad by ad, whether the rating and the results agree.
- Judge every ad on cost per lead, not on the rating. If a Poor ad brings in leads at $60 and an Excellent ad at $85, the Poor ad is the better ad. Keep it.
- Raise unpinned ads by saying more different things, not more things. Adalysis recommends at least 11 genuinely different headlines: some with your keywords, some with benefits, features, calls to action and proof like reviews or years in business. Stuffing in more lines does not work on its own. Only 44% of ads with 15 or more headlines reached Excellent, and one ad with 38 assets was still rated Poor. An ad stuck at Average with all 15 headlines filled is usually repeating itself. Find the lines that say the same thing in slightly different words and replace them with a new idea, such as a guarantee, a price or a review count. We walk through writing fifteen headlines that each say something different in You Filled In Five of Fifteen Headlines.
- Rebuild pinned ads instead of abandoning them. Adalysis's method: remove the pins, edit the ad until it reaches Excellent, put the pins back, then keep improving the other lines until the ad reaches at least Average, or Good if you pinned a single headline. If a position has to be pinned, pin three or more different versions to it rather than one.
- Leave a cheaper ad running even when its rating is Poor. A pinned ad that books jobs at a lower cost is doing exactly what you pay it to do.
- Read the "improve your ad strength" recommendation as a suggestion, not a fix. Google's recommendations also count toward your optimization score, and a 100% optimization score just means you agreed with Google. Do not switch on text customization just to move the meter.
Cost Per Booked Customer Is the Only Google Ads Score That Pays the Bills.
Ad Strength is a useful spell-checker. It catches the ad with four headlines that all say the same thing, the ad with no keywords in it, the ad missing sitelinks. Use it for that. Just do not let it pick your ads for you, because the evidence from more than 5,900 accounts says the ad it prefers costs more per customer more often than not.
The plumber from the start of this article did not need an Excellent ad. He needed the ad that told the right people his price before they called. If you would rather spend your Mondays running the business than reading Ad Strength notices, keeping ads honest to the results is part of our Google Ads management. And if you would rather do it yourself, you now know which number to trust.
