Google Ads Copy · RSA Headlines, Descriptions & CTR

Responsive Search Ads allow up to 15 headlines (30 characters each) and 4 descriptions (90 characters each). Google's system tests different combinations of these assets and learns which combinations produce the best performance (CTR and conversion rate) for different users and queries.

In any given ad display, Google shows 2–3 headlines and 1–2 descriptions. With 15 headlines and 4 descriptions, there are thousands of possible combinations — the RSA system explores this space to find the best-performing permutations without manual split testing.

Headlines are the most visible, highest-CTR element of a search ad. With 30 characters, every word must earn its place. Core principles:

• Include the primary keyword in at least one headline. A headline containing the search keyword signals relevance — both to the user (I searched for X and the ad mentions X) and to Google's Quality Score assessment.
• Each headline should work independently. Google may show any combination of your headlines. A headline that only makes sense in combination with another specific headline will produce awkward ad copy in some combinations.
• Vary the angles. Use different headline types across your 15 — keyword headlines, benefit headlines, differentiator headlines, social proof headlines, CTA headlines — to give Google a range to test.
• Use numbers where accurate. "Free Delivery on Orders Over £30" is more specific and compelling than "Free Delivery Available". Specificity increases CTR.
• Include a clear CTA in at least 2 headlines. "Shop Now", "Get a Free Quote", "Compare Prices" — direct CTAs in headlines consistently improve CTR.

Headline types to include

Descriptions provide 90 characters to expand on the headline's promise. With two descriptions shown together in some formats, they should work independently and complement each other. Key principles:

• Support conversion, not just clicks. Descriptions are seen by users who are already considering clicking (the headline caught their attention). Use description space for information that helps conversion: price range, delivery terms, USPs, and trust signals.
• Include a CTA in at least one description. "Order by midnight for next-day delivery", "Speak to an expert today — call us on 0800 XXX XXXX".
• Use all 90 characters when possible. Longer descriptions that fill the character limit tend to perform better than shorter ones that leave space unused.
• Write different descriptions for different user stages. One description focusing on product features (for consideration stage users) and one focusing on purchase terms (for users ready to buy).

Pinning forces a specific headline to always appear in a specific position (Headline 1, 2, or 3). Google's guidance is to avoid pinning when possible — it reduces the combination space available for testing and can lower ad performance.

Appropriate uses for pinning:

• Legal or compliance requirements. If a disclaimer or specific claim must always appear (financial services, healthcare, regulated industries), pin it to ensure it shows in every combination.
• Branding requirements. If brand standards require the brand name to always appear in Headline 1, pinning enforces this.
• Specific proven combinations. If testing has identified a specific Headline 1 that consistently outperforms alternatives, pinning it while leaving other positions unpinned preserves the winning element while allowing testing to continue on other positions.

If you pin, pin multiple headlines to the same position (e.g. pin 3 headlines to position 1) — this allows Google to test between the pinned options while maintaining the positional requirement.

Google provides an Asset Strength indicator for each RSA: Poor, Good, or Excellent. It assesses the variety and relevance of your headlines and descriptions. Improving from Poor to Excellent is correlated with better ad performance. Factors that affect Asset Strength:

• Headline uniqueness — headlines that repeat the same phrases score lower
• Keyword inclusion — at least one headline containing the ad group's primary keyword
• Quantity — closer to 15 headlines is better than the 3-headline minimum
• Length — fuller use of character limits scores higher
• Diversity — different messaging angles (CTA, benefit, social proof) score better than all-similar headlines

Asset Strength is a directional guide, not a performance guarantee. An Excellent-rated RSA with mediocre headline content will underperform a Good-rated RSA with compelling, specific copy.

• Match the query as closely as possible. Ads that mirror the user's exact query language feel most relevant. Dynamic Keyword Insertion can achieve this automatically, but manual keyword-specific headlines in tightly-themed ad groups are more reliable.
• Specificity beats generic. "Free Next-Day Delivery on Orders Over £50" consistently outperforms "Fast Delivery Available" — the specific detail answers the user's implicit question rather than just making a vague claim.
• Display URL paths add relevance signals. The display URL path fields (/category/subcategory) appear in the ad and signal relevance. "yoursite.com/Running-Shoes/Men" is more relevant than "yoursite.com" for a running shoe query.
• Extensions increase CTR by expanding ad size. Sitelinks, callouts, and structured snippets add visual size and additional clickable surface to your ad — consistently improving CTR over ads without extensions.

With RSAs, A/B testing at the full-ad level (as was done with ETAs) is replaced by asset-level performance data. Google Ads reports performance at the individual asset level in the Ad Assets tab — showing impressions, clicks, and relative performance ratings (Learning, Low, Good, Best) for each headline and description.

To test effectively with RSAs: run a minimum of 2–3 RSAs per ad group with meaningfully different messaging angles (not minor wording variations). Over time, pause the consistently lower-performing RSA and create a new one with a different approach. Review asset-level performance monthly to identify which individual headlines are driving performance and which are consistently rated Low.

Authentic Sources