The news: OpenAI is testing exclusion targeting for ChatGPT ads to give advertisers more control over articulating which context they don’t want their ads appearing near, per Adweek. The test is in a limited rollout to a small group of advertisers.

This addresses early ad buyer complaints about not being able to communicate which audiences they want to target, limited control over where ads appear, and not enough visibility into where they do show up.

SE Ranking found that 14.4% of ChatGPT ads have no topical connection to what users asked about. In some categories, including news and politics, over 50% of ads are off topic.

Zooming out: Exclusion capabilities have become more common across social and enterprise platforms this year.

• Meta launched exclusion-only custom audiences this month that stop specific users from getting ads, helping to avoid wasted spend. It automated that offering last week, which removes some advertiser control.
• Microsoft added term exclusions earlier this year.

Zooming in: OpenAI is trying to build advertiser control in ChatGPT ads while pushing into new markets and lowering the barrier to entry. So far, its ad infrastructure is relatively basic compared with competitors like Google and Meta, with reportedly underdeveloped performance insights and lower-tech buying processes.

• OpenAI announced last week that its ads pilot is coming to 31 European markets, including Spain, Germany, and France.
• Ad revenues have grown 25% since the start of August, though its goal of $100 billion by 2030 doesn’t align with our forecasts for the full US chatbot ad market.
• The cost to join ChatGPT’s ad pilot was cut to $50,000, down from $200,000, but reported CPMs of up to $60 remain high, per Adweek.

On top of being a nascent channel, brand safety is a sticking point for advertisers who are unsure about AI chat placements. Twenty-eight percent of US ad buyers will only advertise in AI-generated content if it’s brand safe, and 2% won’t do so at all, per IAS and YouGov. At the same time, 61% are excited about such opportunities, showing advertisers are interested but want guardrails.

Implications for marketers: Exclusion targeting is an early, limited step that could assuage some brand safety concerns by letting advertisers avoid specific topics or conversations. However, targeting controls, precise placements, and post-campaign verification are different issues, and the pressure is on for OpenAI to reach its revenue targets and convince markers that its ads work.

Those currently in its ads pilot should take advantage of exclusion targeting and document which contexts get excluded while pushing OpenAI for clearer reporting standards and avoiding major investments for now.

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.