The news: ChatGPT’s ad pilot is scaling faster than it’s maturing.

Ads now appear in over 25% of commercial queries, per SE Ranking, but an analysis of over 50,000 prompts across 20 niches found placements are frequently irrelevant despite a small pool of 1,159 unique advertisers.

• About one in seven ads (14.35%) were off-topic and unrelated to the user prompt.
• In categories like news and politics, that figure skyrocketed to over 50%.
• Paying for ads helped brands show up in sponsored slots below queries, but rarely guaranteed that the advertiser itself showed up within the answer. In 96.37% of placements, the advertiser was not cited in the answer copy.
• Some advertisers had an outsized share of placements: Comparison site BestMoney represented over 13% of all ChatGPT ads in SE Ranking’s dataset. For categories like food and beverage, Hungryroot accounted for nearly 80%.

Zooming out: OpenAI’s ad product is working out reach and effectiveness growing pains.

• Advertisers have complained that ChatGPT ads are not displayed frequently enough to fully understand whether placements are driving meaningful results. Despite ads appearing for about 85% of ChatGPT’s users, less than 20% of eligible users are shown them daily, per the company.
• Advertisers have also pointed to issues like a lack of information on how ads perform beyond impressions—something OpenAI has tried to address by shifting to click-based pricing.
• But chatbot ads are still struggling to drive engagement, with clickthrough rates as low as 0.91%, per Adweek—a major hurdle for attracting performance marketers.

OpenAI has attempted to remedy these concerns by partnering with ad tech firms, debuting an ads manager beta, and expanding available ad inventory. But the reality is that ChatGPT advertising, and the broader chatbot advertising market, remain a highly nascent format with opaque results; findings like SE Ranking’s could exacerbate those concerns.

Implications for marketers: ChatGPT’s massive user base may make it appear as an attractive advertising channel, but scale alone won’t fix the format’s lingering problems with relevance, engagement, and measurable results.

That’s not to say ChatGPT will remain an ineffective advertising opportunity—even established formats like traditional search took years to mature—but skepticism will persist until relevance, measurement, and performance improve. For now, AI chatbot advertising remains an under-tested, still-evolving format that is better suited to experimentation than significant budget investment.