The news: OpenAI is shifting ChatGPT ads from an impressions- and clicks-based product toward a performance channel that optimizes campaigns for sales and other conversions.

• The company is rolling out conversion-based bidding, which uses tracking pixels placed across advertisers’ websites, product catalogs, and checkout pages to send conversion data back to OpenAI’s algorithm. That data helps the system learn which users and placements are more likely to produce meaningful actions.
• OpenAI is also expanding its targeting capabilities. Brands can now upload customer email lists to create first-party custom audiences, concentrate or exclude spending by geography, and use negative targeting to prevent ads from appearing alongside irrelevant topics.

Why it matters: ChatGPT’s shift toward conversion-based bidding is an effort to address advertiser concerns and make the platform more attractive to performance marketers.

Advertisers have previously expressed concerns about ChatGPT’s limited performance data, which has made it difficult to determine whether ads drive meaningful business outcomes. Conversion-based bidding gives advertisers a clearer way to optimize campaigns toward sales and other actions, making ChatGPT advertising more appealing despite the format’s early stage.

Conversion-based bidding could draw in more advertisers by giving them clearer visibility into the actions consumers take, and helping them decide whether to increase future investment—marking an important shift from OpenAI’s initial $200,000 minimum spending commitment. Stronger performance tools could also help OpenAI reach its lofty goal of generating more than $100 billion in ad revenues by 2030.

Implications for marketers: ChatGPT advertising is becoming easier to justify as OpenAI adds the targeting and optimization tools expected of a performance channel. Marketers with flexible test budgets and room to experiment may benefit from investing early, especially as intent in chatbot environments is potentially higher than other sources. Still, marketers who test in ChatGPT should avoid scaling spending until reporting and attribution provide clearer proof that conversions are strong.