ChatGPT Ad Dismissals Drop 50% as Relevance Improves

The short version

OpenAI has cut ChatGPT ad dismissals by 50% since February 2026, a data point confirmed by chief revenue officer Denise Dresser at Cannes Lions in June 2026. The improvement signals that the platform’s contextual ad-targeting engine, which matches sponsored content to conversation topics rather than search keywords, is getting better at serving relevant ads — and that users are increasingly willing to engage with them rather than close them.

Key facts

• Ad dismissal rate dropped 50% between February and June 2026, per OpenAI CRO Denise Dresser
• ChatGPT ads hit $100M annualized revenue in six weeks, with 600+ advertisers on the platform
• CPC pricing runs $3–$5 per click; CPM pricing $25–$45 per thousand impressions
• Ads appear on Free and Go tiers only — 700M+ weekly active users, with paid tiers ad-free
• Platform is expanding beyond the US to Canada, Australia, and New Zealand in 2026

What happened

OpenAI’s advertising executives told Adweek on June 23, 2026, that the rate at which users dismiss ads inside ChatGPT has fallen by half since the company launched its advertising business in February. Denise Dresser, OpenAI’s chief revenue officer, shared the figure during an onstage conversation with CNBC’s Julia Boorstin at Cannes Lions, describing the ad format as “about usefulness.” The company uses dismissal rate as a proxy for ad relevance — when users close an ad, it signals the content didn’t match their intent. A 50% drop means the system is learning.

The news follows OpenAI’s March 26 update, where the company said it was seeing “no impact on consumer trust metrics, low dismissal rates of ads, and ongoing improvements in the relevance of ads as we learn from feedback” PYMNTS. By that point, the ad business had already surpassed $100 million in annualized revenue within six weeks of launch, with more than 600 advertisers on the platform, according to an OpenAI spokesperson cited by Reuters.

How does ChatGPT ad targeting actually work?

ChatGPT’s ad engine operates on conversation context, not keyword bids. When a user asks about travel destinations, the system may serve a hotel or airline ad at the bottom of the response — visually separated and labeled as sponsored. OpenAI selects ads by matching advertiser submissions with the topic of the conversation, the user’s chat history, and past interactions with ads Search Engine Journal.

This is fundamentally different from Google Ads or Meta. “The platform excels in contextual relevance, serving ads based on immediate conversational flow and user intent rather than keyword bids,” Nicole Greene, VP Analyst at Gartner, told Raconteur. The trade-off is depth: ChatGPT’s targeting relies on conversation history and location but lacks “the comprehensive demographic data, behavioral tracking, and audience segmentation offered by Google Ads and Meta,” Greene added.

Advertisers don’t see user conversations or personal details — only aggregate performance data like views and clicks. Users can dismiss ads, see why a specific ad appeared, turn off personalization, or clear all ad-related data. OpenAI also restricts ads from appearing in conversations about health, mental health, or politics, and blocks them for users under 18.

Who is affected, and when?

The immediate impact falls on US-based advertisers and the 700 million-plus weekly active ChatGPT users on Free and Go ($8/month) tiers. OpenAI is expanding pilots to Canada, Australia, and New Zealand in the coming months, with plans to reach “many more markets” in 2026.

Paid subscription tiers — Plus ($20/month), Pro ($200/month), Business, Enterprise, and Education — remain completely ad-free. This creates a structural ceiling, as Claire Holubowskyj, Senior Research Analyst at Enders Analysis, noted: “ChatGPT has enormous reach but low depth compared to established platforms: the heaviest users will be on ad-free subscription tiers” Raconteur.

OpenAI has also aggressively lowered its minimum advertiser commitment from an initial $250,000 to $50,000, and shifted from a pure CPM model ($25–$45 per thousand) to include CPC pricing at $3–$5 per click. The early CPM quoted to media buyers during the initial US test was approximately $60 per thousand views, according to Digiday. The price compression suggests OpenAI is responding to advertiser demand for performance-based pricing and broader access.

What this means (our take)

The 50% dismissal drop is the most important signal for PPC practitioners evaluating whether to allocate test budget to ChatGPT. It tells you the platform isn’t just growing — it’s getting better at the thing that determines whether users will ever click. Low dismissal rates mean the contextual engine is matching ads to conversations that already demonstrate commercial intent. Someone asking ChatGPT to compare CRM software is further along in the buying process than someone typing “CRM” into Google.

But the practical reality for most advertisers is this: ChatGPT ads are still a supplementary, test-and-learn channel — not a replacement for Google or Meta. Gartner is advising clients to treat it that way, and the data supports it. The platform lacks full-funnel reporting, traditional placement parameters, and auction mechanics. The targeting pool is constrained by ad-free paid tiers, and OpenAI’s rapid pricing shifts — CPM to CPC, $250K to $50K minimums — show the platform is still finding its footing.

The agent-era angle is equally important. Unlike Google’s AI Overviews, which now generate 600 million-plus AI summaries daily and reduce click-through rates, ChatGPT embeds ads inside a dialogue. The interaction is inherently higher-intent. An AI agent managing Google Ads bidding could, in principle, also allocate budget to ChatGPT’s platform once API access matures. We’re not there yet, but the 50% dismissal improvement is the kind of signal that turns experimental budgets into committed line items.

What to do now

1. Request access if you meet the $50K threshold. OpenAI has lowered its minimum commitment, and early advertiser slots in new markets (Canada, Australia, New Zealand) will be less competitive than the US pilot.
2. Structure creative for conversational context, not keyword intent. ChatGPT ads sit below answers — the user just asked a question. Your creative should feel like a natural next step, not a display banner. Test dialogue-style formats.
3. Monitor dismissal rate as a relevance KPI. If you gain access to ChatGPT’s ad manager, track dismissal rate alongside CTR. A low dismissal rate on ChatGPT matters more than on display because it measures contextual fit rather than banner blindness.
4. Keep ChatGPT in a multi-channel mix. As Enders Analysis recommends, treat it as part of a broader strategy — not a Google Ads replacement. Use it for high-consideration, information-rich categories where conversational AI influences purchase decisions.
5. Watch for API and agent access. OpenAI is building in-house ad infrastructure. Once programmatic or API access opens, autonomous AI agents — already managing Google Ads budget allocation — could allocate spend across ChatGPT alongside traditional search. That’s the inflection point.

FAQ

Sources

• Adweek — OpenAI Sees Fewer Users Dismissing ChatGPT Ads as It Scales (June 23, 2026)
• PYMNTS — OpenAI Expands ChatGPT Advertising to More Markets (March 27, 2026)
• Raconteur — OpenAI rolls out ads on ChatGPT, but marketers remain unconvinced (May 5, 2026)
• Search Engine Journal — OpenAI Begins Testing Ads In ChatGPT (February 9, 2026)
• OpenAI — Testing ads in ChatGPT (official blog)

Frequently Asked Questions

Is this strategy suitable for small budgets?

Yes. Most of the tactics on this page work at any budget level from $500/month upward. The key is focusing on the highest-intent keywords and thorough negative keyword management. Start small, prove ROI, then scale.

Where can I learn more about AI-managed Google Ads?

Our free Google Ads Expert skill at roa-marketing.com/skills/google-ads-expert/ covers every PPC workflow in detail. It updates daily from live campaign data and is designed for AI agent consumption.