Testing ChatGPT Ads: What 98,000 Ads Reveal About Strategies and Click Rates

Rapid Market Growth: Empirical Data on the Launch of ChatGPT Ads

Advertisers are facing one of the most significant shifts in digital performance marketing: conversational ad formats are establishing themselves as a vital touchpoint across the customer journey. A comprehensive analysis by Graphite and Similarweb, covered in detail by Search Engine Land and examined in depth by my colleague Rick in his latest research report, provides solid market figures for the first time. The study systematically evaluated over 98,000 ChatGPT ads and more than 8,000 landing pages between April 1 and August 16. Graphite's US dataset alone captured over 95,000 ad impressions across more than 500 advertisers.

The headline metric highlights the rapid momentum of this new channel: according to Graphite and Similarweb, 43 percent of ad-eligible US users were reached by at least one ad in ChatGPT during the observation window. This confirms that ChatGPT Ads have quickly outgrown the experimental stage, building meaningful reach within months. However, brands face the operational challenge of turning this scale into profitable acquisition. While audience adoption continues to rise, many teams remain in an exploratory phase, working to define repeatable best practices.

In his research report, my colleague Rick rightly views this dataset as an important milestone. While the industry previously leaned on anecdotal observations from early adopters, this cross-sectional study provides solid empirical evidence of live campaign dynamics. Methodologically, this is an aggregated observational panel study with Level B evidence—meaning direct causal ROI drivers cannot be isolated—yet it offers indispensable guidance for marketing leaders allocating conversational AI budgets.

Top advertisers do not copy-paste search copy: 98% write custom conversational creative to drive performance in chat.

Conversation Timing: Why Early Turns Make the Difference

A key operational takeaway from the research centers on ad placement within conversation turns. The available data indicates that ads served during the earliest turns of a chat session achieve the highest click-through rates and impression share. When users initiate a new chat, their cognitive attention is peak and their intent to find immediate solutions is strongest. In this initial stage, sponsored placements achieve their highest engagement.

From a behavioral standpoint, this pattern is logical: at the beginning of a session, users outline their core question or look for broad guidance on a specific challenge. An ad delivered here functions as a helpful resource within their discovery flow. As conversations deepen into nuanced troubleshooting, code refactoring, or iterative drafting, commercial interruptions can feel intrusive, explaining the drop in impression share and CTR during later turns.

For campaign managers, the strategic implication is direct: align bidding to target the initial discovery phase of a conversation. Ad copy should directly address top-level search intent rather than hyper-specific edge cases. Entering later conversation turns results in lower user attention and weaker engagement across global inventory.

Creative Strategies: The End of Copy-Pasting Search Ads

Perhaps the clearest signal from the study is directed at creative teams: according to Graphite, 98 percent of top advertisers deploy custom-crafted conversational copy rather than syndicating existing text from Google Ads or Meta campaigns. The traditional copy-paste workflow falls short in conversational spaces. While paid search leans on concise keywords and paid social relies on visual interruption, an AI chat environment demands an advisory, contextual, and dialogic voice.

Advertisers must understand that users interact with AI assistants as conversational peers. Generic slogans and aggressive calls to action feel out of place in this format. High-performing campaigns align with the tone of the platform: answering inquiries directly, providing concrete value, and acting as a natural continuation of the AI's response. This structural difference explains why virtually every leading advertiser writes platform-specific copy.

The table below summarizes key empirical metrics from the Graphite and Similarweb study, outlining the actionable priorities for day-to-day campaign management.

Landing Page Architecture: Prioritizing Product, Features, and Pricing

Beyond ad creatives, the Graphite and Similarweb report highlights post-click landing page strategy. The data points to a highly concentrated approach: the vast majority of analyzed companies direct traffic to five or fewer landing pages across their entire ChatGPT ad mix. Rather than spinning up hundreds of granular subpages for every query variant, top performers route volume to a handful of deeply optimized destinations, simplifying quality control and ongoing conversion optimization.

The study demonstrates a clear preference for high-intent, information-dense page formats. Core product pages, transparent pricing overviews, and detailed feature breakdowns account for the majority of destinations. Users arriving from a chat placement have already articulated a specific need; they expect immediate access to technical specs, pricing tiers, and direct comparisons rather than high-level brand storytelling or generic homepages.

For performance teams, closing the gap between ad messaging and landing page content is critical. If a chat ad outlines a solution to a problem, the destination page must fulfill that promise without friction. Pricing transparency is especially important: hidden rates or vague terms result in rapid drop-offs among users arriving from high-intent AI interactions.

Click-Through Rates and Performance: A Maturing Ecosystem

One of the most encouraging trends for advertisers is the steady lift in user engagement: global click-through rates more than doubled from the platform's launch in April through mid-August. This increase points to a maturing ad ecosystem. Users are growing accustomed to sponsored recommendations within their workflows, while advertisers are sharpening their contextual relevance.

The findings suggest that this performance lift is driven by better alignment with user intent. While initial ad placements were often generic, advanced campaigns integrate smoothly into the topic of the query. When an ad surfaces at the precise moment a user seeks a software tool or physical product, it is perceived as useful assistance, driving the upward trend in market CTRs.

Despite these encouraging gains, my colleague Rick emphasizes analytical rigor in his research report: this study is a cross-sectional market analysis rather than a controlled A/B split test. Direct causal links between specific tactics and bottom-line ROI cannot be established from aggregated data alone. Teams planning their allocations should cross-reference broader market figures with our automated benchmarks below.

Actionable Playbook: Best Practices for Campaign Managers

The study's findings provide clear operational steps for advertisers. First, stop porting search and social campaign structures directly into ChatGPT Ads. Build dedicated conversational creative workflows focused on natural language, consultative positioning, and immediate utility. Copywriters should frame ads as direct, helpful answers to the user's implicit problem.

Second, streamline your landing page portfolio for maximum clarity. Rather than dispersing budget across dozens of low-traffic pages, concentrate on three to five refined product, feature, and pricing destinations. Ensure fast loading speeds and straightforward access to all information promised in the ad. Robust conversion tracking is essential to measure post-click behavior reliably.

Finally, prioritize early conversation turns in your bidding strategy. With data showing peak impression share and CTR in initial session turns, configure campaigns to capture users right away. ChatGPT Ads reward advertisers who treat the format as a distinct channel with its own mechanics. Explore our Research Lab and compare your metrics with AI AdBase benchmarks below.