Ads appearing within Google AI Overviews are not behaving like traditional paid search. The query length and CTA language that drive performance in standard search ads produce different results inside AI Overview placements. This data from Adthena’s AI Search Ads Advertiser Intelligence Report reveals two specific optimizations that separate high-performing AI Overview ads from underperforming ones: targeting 3-4 word queries and matching CTA language to evaluation-stage intent.

Essential Statistics

• More than 60 percent of AI Overview ad placements occur on 3-4 word search queries, establishing mid-length queries as the dominant placement context.
• The remaining 40 percent of AI Overview ad placements occur on other query lengths.
• Compare CTAs produce a 35 percent performance uplift relative to the Shop Now baseline.
• Get a Quote CTAs generate a 22 percent uplift over the Shop Now baseline.
• Find a Store CTAs underperform the baseline by 30 percent, the largest negative deviation in the CTA analysis.
• Shop Now serves as the baseline CTA in this analysis, meaning all uplift figures represent relative performance improvement or decline against that standard.

Key Takeaways

• Mid-length queries of 3-4 words are the primary placement zone for AI Overview ads, suggesting users asking moderately specific questions are the most commonly served audience in this ad format.
• The CTA language performance gap is significant and directly actionable. A 35 percent uplift from Compare versus the baseline reflects a fundamental alignment between evaluation-stage query intent and CTA copy that invites comparison rather than demanding a purchase.
• The 30 percent underperformance of Find a Store confirms that local or transactional CTAs misalign with AI Overview user intent. Users receiving AI-generated responses are in a research and evaluation mindset, not a store-visit mindset.
• The pattern across the CTA data consistently rewards language that matches the deliberative nature of AI-assisted search. Compare and Get a Quote both invite a next step that feels continuous with the research process, while Shop Now and Find a Store push for a transactional leap that AI Overview users are not yet ready to make.
• Advertisers transferring standard search ad creative directly into AI Overview placements without CTA adaptation are systematically underperforming by up to 30 percent before any other optimization variable is considered.

Actionable Insights

• Audit every active AI Overview ad placement for CTA language and replace any Find a Store, Buy Now, or other transaction-first CTAs immediately. The 30 percent underperformance means those CTAs are actively suppressing performance in a placement format where you are already paying for the impression. Swapping to Compare or Get a Quote language is the highest-immediate-ROI change available in this ad format.
• Build your AI Overview ad keyword strategy around 3-4 word query targets rather than extending your standard broad match campaigns into this format. With more than 60 percent of AI Overview ad placements occurring on mid-length queries, your bid strategy should prioritize that query length range to maximize placement relevance and reduce wasted spend.
• Test Compare [Your Product] vs [Competitor] as a headline format for AI Overview ads in your top competitive categories. The 35 percent uplift from Compare CTAs combined with the high share of comparison query placements suggests that framing your ad around the comparison the user is already conducting produces the strongest intent alignment in this format.
• Separate your AI Overview ad campaigns from your standard search campaigns rather than using the same ad groups across both placements. CTA language that works in traditional search underperforms in AI Overview placements by a measurable margin. Mixed campaigns produce blended results that make it impossible to optimize either placement type independently.
• Use Get a Quote CTAs for any service or financial category where immediate purchase is not the expected conversion action. The 22 percent uplift positions it as the right call-to-action for categories where the next step naturally involves a consultation or customization before transaction.

“The CTA data here is a direct instruction manual. Compare outperforms the baseline by 35 percent in AI Overview ad placements. Find a Store underperforms by 30 percent. That is a 65 percentage point spread between the best and worst CTA language choices in this format. Any advertiser running Find a Store or Shop Now CTAs in AI Overview placements should change them today.” – Neil Patel