Study finds paid ChatGPT visits trail Google Search in ecommerce returns

Paid visits from ChatGPT generated less revenue and converted at lower rates than paid Google Search visits in a working paper by researchers at Frankfurt School of Finance & Management and the Vienna University of Economics and Business. The findings offer an early comparison of advertising performance across the two channels, but the paper remains under review.

Across the study, paid ChatGPT visits generated an average of 89 cents in revenue per session, compared with $2.45 for paid Google Search. The reported conversion rates were 0.37% for ChatGPT ads and 1.17% for Google Search. Among 102 advertisers included in the analysis, 82% recorded no purchases from their ChatGPT ads.

The researchers concluded that, at publicly reported prices, Google Search produced more than six times as much revenue per advertising dollar. They estimated a ChatGPT click would need to cost about 26 cents to match paid search’s return. The largest publicly known account in their dataset paid $1.72 per click.

A broad dataset, with early findings

The paper examined 1,898 ecommerce sites that together recorded $16 billion in annual sales and nearly 10 billion visits. Its 60-week observation period ran from June 2025 through August 2026. The researchers used ChatGPT’s advertising rollout in the United States and then eight additional countries, with the European Union remaining ad-free and serving as a control group.

The scale of the analysis makes the results relevant to advertisers evaluating new sources of paid traffic. Still, the study’s status matters: its findings are not a settled verdict on ChatGPT advertising. Advertisers should treat the figures as early evidence and account for the paper’s methodology and the particular campaigns and sites represented in its sample.

The gap between the reported revenue per visit and conversion rates also underscores why traffic volume alone is not a sufficient measure of channel value. A placement can attract visits without generating purchases at a rate that supports its cost. For performance marketers, the practical comparison involves campaign spending, conversion outcomes and revenue—not simply the number of available advertisers or reported queries.

Paid and unpaid ChatGPT traffic differ

The researchers found a substantial difference between paid ChatGPT visits and unpaid referrals from the service. Unpaid ChatGPT referrals converted at 2.21% before the advertising rollout, versus 0.37% for paid visits. Visitors arriving through paid ChatGPT placements also performed worse than visitors arriving through Google ads, Google organic search, direct traffic and ChatGPT’s unpaid recommendations.

Separate findings from Adobe Analytics point to the potential value of organic AI referrals, though they measure a different kind of traffic. Adobe reported that AI-referred traffic to U.S. retail sites generated 83% more revenue per visit than non-AI traffic in September. Those visitors converted 90% better, spent 73% more time on retail sites and viewed 25% more pages.

Because the studies define and measure their traffic differently, the results should not be treated as a direct, like-for-like comparison. Together, however, they reinforce the importance of separating paid placements from unpaid AI referrals when evaluating performance. Grouping both under a single “AI traffic” label could obscure meaningful differences in visitor intent and campaign economics.

Scale claims do not establish campaign returns

Advertiser participation in ChatGPT has grown, according to figures cited in the research coverage. Adthena tracked 1,082 brands advertising across the U.S., U.K., Australia and the EU in June, rising to 3,741 in July and 6,333 in August. New Street Research estimated that ChatGPT had roughly 60,000 advertisers by early October, up from about 20,000 at the end of August. Those estimates indicate adoption, but do not establish advertisers’ returns.

Questions also remain about how to compare ChatGPT’s search-like activity with Google Search. OpenAI has claimed ChatGPT handles roughly 20% as many search-like queries as Google, but has not disclosed the underlying count or enough detail about its measurement for a clean comparison. SEOClarity estimated ChatGPT’s search volume at about 7.5% of Google’s, or roughly 375 billion search-like queries annually.

The financial outlook for advertising is similarly difficult to assess from public information. OpenAI’s projections assume advertising revenue reaches $102 billion by 2030, or about $40 per free user. The company has said ChatGPT Ads reached a $1 billion annualized revenue run rate less than 200 days after launch. That is a run-rate figure, not $1 billion in advertising revenue already collected, and the privately held company has not provided evidence supporting the claim that can be checked against public financial filings.

The researchers estimate current advertising generates about $1 per free user annually, while serving a free user costs roughly $5. Their analysis suggests that greater advertiser participation and more clicks alone would not resolve that gap unless the value of each click rises. For marketers, the central question remains whether paid ChatGPT traffic can produce results that justify its price.