abhi.performs
Abhishek ThakralWhatsApp me

Home › Guides › Responsive search ads

Responsive search ads best practices: the mistakes to fix

By Abhishek Thakral · abhi.performs, Dubai · Updated November 2026 · 9 min read

Short answer

Write many genuinely different headlines, keep each keyword inside one headline, run at least two ads per ad group and add six or more sitelinks. Pin only text that must always show. Treat Ad Strength as a checklist, not a score to chase: Google says it is not used in the auction.

Responsive search ads best practices are easy to find, and most lists say the same thing: fill every field and push Ad Strength to "Excellent". Yet the rating is the one part of an RSA that Google itself says the auction ignores.

This guide works from Google's help pages and one large independent study, written for the accounts I run most: UAE businesses buying leads, where a cheap click that never becomes a booking is wasted money. Each section is a mistake I find in audits, with the fix.

How responsive search ads work

A responsive search ad (RSA) takes up to 15 headlines and 4 descriptions and lets Google assemble them into ads for each search, according to Google's page About responsive search ads. You need at least 3 headlines and 2 descriptions to save one.

Google warns that assets can show in any order, so every headline has to make sense on its own and next to any other.

Mistake 1: treating Ad Strength as the goal

Google's page About Ad Strength is blunt about it. Under the label "Common misconception", it says Ad Strength "isn't used to calculate Ad Rank, Quality Score, or auction wins", and it does not decide whether an ad is eligible to serve.

What it measures is structure. The page lists three factors: whether your headlines and descriptions contain your ad group's keywords, how many combinations your assets allow, and whether you have enough sitelinks. It says it does not measure your landing page or your offer.

Google does publish its own figure: advertisers who move from "Poor" to "Excellent" see 15% more conversions on average, from Google's internal data. An independent view points the other way. Optmyzr's study of about 20,000 accounts (April 2026) found no reliable link between Ad Strength and efficiency measures such as cost per acquisition. It is observational vendor data, and its authors call the results directional.

The fix: use the rating's action items as a checklist when you build an ad, then judge the ad on qualified leads. For what does feed ad quality, see my guide to Google Ads Quality Score.

Mistake 2: fifteen headlines that say the same thing

Google's Ad Strength page asks for unique headlines and descriptions and warns against "repetitive words, phrases, or ideas", because only unique assets add new combinations for Google to test. Fifteen versions of "Best Clinic in Dubai" count as one idea.

For a lead-gen account I write headlines in groups, each doing a different job. An illustrative mix for an aesthetic clinic:

Two rules from Google make the keyword part work. A keyword must sit inside a single headline, not split across two, and a keyword longer than 30 characters belongs in a description. Google also suggests headlines of different lengths, and Optmyzr's data found headlines under 20 characters had the lowest cost per acquisition, so a short headline is not a wasted slot.

Mistake 3: pinning everything, or pinning badly

Pinning fixes a headline or description to one position, and Google says it "isn't recommended for most advertisers" because it cuts the combinations Google can test. Its best practices page answers "should I pin?" with "For general ad campaigns, no."

There are good reasons to pin, though:

Optmyzr's study points the same way: in its full dataset, partial pinning beat both no pinning and full pinning on nearly every measure, and fully pinned ads had the worst cost per acquisition.

Mistake 4: one RSA per ad group, and one URL for all

Google recommends at least 2 RSAs with "Good" or "Excellent" Ad Strength per ad group, each with its own final URL. By its internal data, adding a second RSA brought 6.6% more conversions at a similar cost per conversion, and a third added 3.7%.

A second ad only helps if it says something different. I build it around another angle (price and packages against speed and convenience, for example) and send it to a page that matches. Two near-identical ads on one URL split impressions and teach you very little.

Mistake 5: ignoring the assets around the ad

Sitelinks count towards Ad Strength: Google wants 6 or more across the account, campaign and ad group, and both your own and dynamic sitelinks count.

One more thing to check if text customization is on (it is one of the features in AI Max for Search campaigns): Ad Strength counts Google-generated text the same as yours. Google's asset report marks those lines "Google AI" in the "Added by" column. Read them, especially in clinics and property, and remove anything you would not say yourself.

Mistake 6: judging RSAs on CTR and editing them every week

Google's best practices page says RSAs help you qualify for more auctions, so ad-level clickthrough and conversion rates "won't paint a full picture". It suggests watching the extra impressions, clicks and conversions your ad groups and campaigns get.

Two reports show what is working:

Google warns that editing or removing assets changes which ads serve, so swap a few weak assets at a time and wait for enough leads to judge.

What I see in UAE lead-gen accounts: the RSA is rarely the first problem. When conversion tracking counts button clicks instead of real enquiries, every ad looks fine and none can be judged. I fix the conversion action first, ideally with offline conversion import from the CRM, and only then compare ads on leads that became bookings.

When to leave an RSA alone

Not every warning needs action. This is the rule I use:

Responsive search ads best practices: the fix order I use

  1. Check what counts as a conversion. If tracking is wrong, nothing below can be judged.
  2. Remove near-duplicate headlines and write new ones by job: service, audience, proof, next step, qualifier.
  3. Put each main keyword whole inside one headline, and move long ones to a description.
  4. Unpin everything except required text, and pin 2 or 3 versions of that to Headline 1, Headline 2 or Description 1.
  5. Add a second RSA with a different angle and its own matching final URL.
  6. Bring sitelinks to 6 or more and add your business name and logo.
  7. Read the "Google AI" assets if text customization or AI Max is on.
  8. After a few weeks, compare ads on qualified leads, not on the rating.

If you would rather have this done for you, it is part of my Google Ads management work.

Frequently asked questions

How many headlines should a responsive search ad have?+
Up to 15 headlines and 4 descriptions, with a minimum of 3 headlines and 2 descriptions. Google recommends as many unique ones as you can write, because only distinct assets add new combinations.
Does Ad Strength affect ad rank?+
No. Google says Ad Strength isn't used to calculate Ad Rank, Quality Score or auction wins, and it doesn't decide whether an ad can serve.
Should I pin headlines in responsive search ads?+
Only text that must appear in every ad, pinned to Headline 1, Headline 2 or Description 1 with 2 or 3 unique options. In AI Max, Google says pins are ignored when text customization and final URL expansion are both on, or URL inclusions are set.
How many responsive search ads should each ad group have?+
Google recommends at least 2 with Good or Excellent Ad Strength, each with its own final URL. Its internal data shows 6.6% more conversions from a second RSA and 3.7% from a third, at a similar cost per conversion.

Related

Not sure your search ads are pulling the right leads?

Message me on WhatsApp or request a free campaign audit. I'll read your responsive search ads and their asset reports, check what your account counts as a conversion, and show you which ads to keep, fix or pause.

Ask me on WhatsApp