In the US, Google Search Ads Generated More Revenue per Visit Than ChatGPT Ads

In the US, a visit from a ChatGPT ad generated $0.89 in revenue for advertisers, compared with $2.45 from paid Google Search. According to Search Engine Journal, the figures come from an October 2026 working paper by Maximilian Kaiser and Christian Schulze. The authors studied traffic and sales during the rollout of OpenAI ads across different countries. Their comparison of paid channels is based on data from 102 advertiser websites with complete sales tracking.

Sales and the Estimated Cost per Click

The conversion rate for visits from ChatGPT ads in US traffic was 0.37%, compared with 1.17% for paid search. The gap remained when the channels were compared on the same websites, accounting for visitors’ devices and the week of each visit.

Of the 87 advertisers buying ads in both channels for US traffic, 84 earned less revenue per visit from ChatGPT. For 73 advertisers, ChatGPT ads produced no sales from a combined 18,726 visits; the other 14 accounted for all 570 purchases in this group. One advertiser accounted for 66% of purchases from ChatGPT ads across all 102 websites. Without that advertiser, the gap between the channels would have been wider.

The authors estimated the cost per click at which ChatGPT would match paid search on tracked revenue per dollar of ad spend: about $0.26. The calculation included 16 accounts where at least 80% of search traffic came from the US. Their median search cost per click was $0.72. Because the result depended heavily on individual advertisers, repeated calculations produced a range of $0.05 to $0.57.

For comparison, the average cost per click for ChatGPT ads in one publicly documented account was $1.72 over 15 days. One account’s results do not establish a market-wide price. OpenAI says there are not yet general ChatGPT Ads performance benchmarks across advertisers, industries, and campaign types.

Study Limitations and Unpaid Referrals

The paper is under review, and its data ends on August 22, 2026. Advertisers chose for themselves whether to buy ChatGPT ads. The comparison describes these companies’ results but does not prove that the ad channel itself caused the difference. The data records the last referral before a visit; it contains no information on users’ queries, ad impressions, or subscription types.

The cost-per-click estimate assumes similar sales margins in both channels and counts a single tracked visit. After advertisers began buying ChatGPT ads, the authors found no notable increase in traffic from other channels relative to similar websites that did not buy them. However, the analysis can rule out only a large increase, not every possible additional effect.

Separately, the researchers examined unpaid referrals from ChatGPT responses to 1,898 websites. They estimated a 5.9% decline in those referrals over the first 13 weeks after the ad launch, but the estimate’s range also allows for no change. They compared these referrals with organic search visits to the same websites and with conditions in the EU, where ChatGPT ads were not running during the observation period. After adjusting for differences in the rate of ChatGPT adoption, the estimated decline was about 6–8%. They found no notable change in conversion rates among visitors arriving through unpaid referrals.

What Store Owners Should Check

For your own store, it is more useful to calculate an affordable cost per click from your own data than to treat $0.26 as a ready-made bid. Divide revenue per visit from ChatGPT ads by revenue per visit from paid search, then multiply the result by your search cost per click. That gives you an estimate of where the channels break even on tracked revenue per dollar spent. The formula does not account for differences in profit or purchases missed by last-touch tracking.

Before comparing channels, check how paid visits appear in your analytics. According to OpenAI, ad review bots may open landing pages with ad tracking parameters. Third-party tools can record these as paid traffic even though they are not visits from shoppers. Such a tracking error will affect the revenue-per-visit calculation.

The observation period has already ended. Since then, OpenAI has expanded ChatGPT ads to 89 countries and reported that more than 100,000 advertisers have launched campaigns. Some individual campaigns have reported more favorable results, but they used different metrics. They cannot be directly compared with this study’s revenue per visit.

My View: Get Tracking Right, Then Test

I would not plan on ChatGPT ads as a store’s primary source of orders based on these data. For a catalog with product variants, I need reliable sales tracking and a way to compare spending with revenue. In the group studied, paid search performed better per visit, while visits from ChatGPT did not lead to any purchases for many advertisers.

But I would not treat $0.26 as a universal bid ceiling, either. I would first check which visits make it into the reports, then set aside a limited test budget and compare the results with my own paid search campaigns. Without that comparison, I would consider it premature to increase spending on a new channel.

Sources

Where the news comes from. The text is a retelling in the author’s own words; the facts come from the source, the opinion is the author’s.

1. Google Search Ad Visits Out-Earned ChatGPT Ads In U.S. via @sejournal, @MattGSouthern www.searchenginejournal.com

About the author

Mehran Zeynalov

Textile store owner

I sell home textiles and am choosing a CMS for a catalog with several product variants. I need simple inventory management and a way to update the range quickly without a developer.

All posts by the author