Divergent Returns for Search Advertisers
The Kaiser and Schulze study, which analyzed 102 different websites with transaction and revenue tracking, found that ChatGPT ad sessions generated lower rates of conversion and lower revenue per session than paid search in U.S. traffic. In these U.S. results, paid search generated $2.45 in revenue per session whereas ChatGPT ad sessions generated $0.89. The transactions per session also differed, with paid search at 1.17% and ChatGPT at 0.37%. Out of 87 different advertisers who used both paid search and ChatGPT ads in U.S. traffic, 84 produced less revenue per session using ChatGPT ads. The statistical model built by Kaiser and Schulze controlled for website, device type, and week, and it found that 84 out of 87 advertisers who used both paid search and ChatGPT ads in U.S. traffic generated less revenue per session with ChatGPT ads.
| Metric (U.S. Traffic) | ChatGPT Ads | Paid Search |
|---|---|---|
| Revenue per session | $0.89 | $2.45 |
| Transactions per session | 0.37% | 1.17% |
| Return per $1 spent | $0.52 | $3.40 |
One university release stated that for every dollar an advertiser spends on ChatGPT Ads, they receive $0.52 back in revenue, compared to $3.40 for every dollar spent on Google Search. This figure combines measured performance with the median cost-per-click pricing from 16 paid search accounts and the average cost-per-click pricing from one ChatGPT account. The ChatGPT price used was $1.72, which Opascope published in June for one account with $60,000 in spend over 15 days. Opascope reported a 1.49x ROAS for those campaigns. Alphabet reported $63.35 billion in Search and other revenue for the third quarter, and Chief Business Officer Philipp Schindler cited lower monetization in AI Mode as a factor in slower Search growth. An OpenAI spokesperson said the Kaiser and Schulze study "captures an earlier period of ChatGPT Ads’ development and extrapolates platform performance from a small sample."
Low Intent in User Conversations
Rand Fishkin, co-founder and CEO of SparkToro, concludes that the primary reason AI visibility matters is Google’s implementation of AI Overviews and AI Mode, rather than ChatGPT itself. Data from a U.S. opt-in panel analyzed by SparkToro and Verso shows that 6.9% of ChatGPT conversations showed purchase intent between January 1 and October 29, 2026. The sample included 41,961 conversations from 12,321 people. In these conversations, 4.9% of people planned to purchase and were looking at options, while 2.1% asked questions about a specific brand, product, or service. The remaining 93.1% showed no intention to purchase.
Of the responses in this sub-sample, 87.9% did not name a single brand, product, or business. The report also found that 86.7% of all prompts did not ask for brand, product, or company information. Shopping, brands, and local services accounted for 5.3% of the primary conversation goals. The panel was skewed toward younger people, women, and people at lower income levels. To prepare the data, researchers eliminated the 131 heaviest users who sent an average of more than 33 prompts per month. Researchers used Anthropic’s Claude model to code each conversation based on the title, up to six prompts, and the first few lines of the first response. The coders agreed on categories 90% of the time and on brand tags between 97% and 99% of the time. Fishkin noted that 15% or more of Google searches fit his definition of purchase intent based on a separate December 2024 analysis.
Agency Tests and Variable Results
I find the individual agency reports confirm that ChatGPT Ads struggle to deliver the same results as established search campaigns. StubGroup ran ChatGPT Ads and Google Search campaigns for a B2B compliance company for 30 days. The clicks on ChatGPT ads cost $0.76, while Google search campaign clicks cost $10.46. Although ChatGPT had a lower cost per lead at $92.21 compared to $157.57 for Google, the conversion rate for Google was 6.64% and the rate for ChatGPT was 0.8%. Workshop Digital spent $2,319 on ChatGPT clicks at $3.14 per click and received zero qualified leads. Their Google Search campaign had a $6.76 cost per click and produced two qualified leads.
Brafton spent $1,524.11 for 263 clicks on one ChatGPT ad from June 17 to August 28 and recorded one form submission. Zach Shah, the director of paid media at Brafton, said ChatGPT ads are "priced like search, but seem to behave like paid social." Jotform reported that their cost per signup was approximately 30% less with ChatGPT than with paid search in July. However, these results vary widely because they depend on specific budgets, click prices, and attribution settings. For example, the Kaiser and Schulze paper’s $1.72 ChatGPT price is toward the low end of what practitioners report and remains below the $3 to $5 starting maximum bid suggested by OpenAI.

