ChatGPT Ads are drawing major advertiser interest, but a new working paper suggests their early sales economics still lag paid Google Search by a wide margin.
The October 2026 working paper, Can Advertising Fund Free AI? Evidence from ChatGPT’s International Advertising Rollout, was written by Maximilian Kaiser of WU Vienna and Christian Schulze of Frankfurt School of Finance & Management. The paper is under review and uses traffic and sales data from 1,898 online shops across 60 weeks, from June 29, 2025, through August 22, 2026.
Its clearest paid-channel comparison comes from U.S. traffic, where most early ad delivery was concentrated. In the paper’s financial sample of 102 advertiser websites, a paid ChatGPT visit generated US$0.89 in revenue per session, compared with US$2.45 for a paid Google Search visit. That is a 64% shortfall in tracked revenue per session.
The conversion gap was wider than the revenue gap. Paid ChatGPT visitors converted at 0.37%, while paid Google Search visitors converted at 1.17%. Among 87 advertisers that bought both channels in U.S. traffic, the authors report that 97% earned less revenue per ChatGPT visit and 98% converted worse.
A pricing problem, not just a traffic problem
The authors calculated a parity price: the click price at which a ChatGPT ad click would produce the same revenue return per ad dollar as paid search. Using a median US$0.72 paid-search click in the study’s cost sample, they estimated a ChatGPT click was worth about US$0.26 at paid search’s rate of return.
The paper compares that estimate with a US$1.72 per-click price from the largest publicly known ChatGPT Ads account. On that comparison, the known price was more than six times the study’s parity price.
Organic ChatGPT referrals held up better
The same study found a different pattern for unpaid ChatGPT referrals. The authors estimate that organic ChatGPT referral volume fell by about 6% to 8% after ads arrived, but the unadjusted 5.9% estimate had a 95% interval that included zero. The larger 7.2% to 8.4% decline appeared after the authors adjusted for assumed faster ChatGPT usage growth in treated markets than in Europe.
Conversion was more stable. Organic ChatGPT conversion changed by -0.02 percentage points from a 2.21% pre-launch baseline, a result the authors describe as no detectable average change.
Ads may have reduced the amount of unpaid referral traffic, at least moderately, but the people who still clicked through from organic ChatGPT recommendations bought at roughly the same rate. The study therefore raises a harder question for OpenAI and advertisers: whether paid placements can become valuable enough without making the underlying product less trusted or less useful.
OpenAI is pushing advertising as part of ChatGPT’s business model
OpenAI began testing ads in ChatGPT in the U.S. on February 9, 2026, for logged-in adults on Free and Go plans. The company says Plus, Pro, Business, Enterprise, and Education plans do not have ads. OpenAI has also said ChatGPT ads are labeled as sponsored and visually separated from the organic answer.
By August 31, OpenAI said ChatGPT Ads had reached a US$1 billion annualized revenue run rate in less than 200 days, with tens of thousands of advertisers and availability in more than 40 countries through its Ads Solutions team and partners. That growth gives the channel commercial momentum even as the new study questions the return advertisers receive from early paid visits.
The study has limits advertisers should not ignore
The authors note that the paid comparison is descriptive rather than causal: the sample is not randomized, and advertisers chose whether to test the channel. The data records website sessions, transactions, revenue, engagement, and traffic source, but not the prompts that generated the ads, the recommendation sets, the impressions people saw, or the subscription tier of the user.
The paper measures last-click website revenue. Brand exposure, assisted conversions, longer purchase cycles, and later purchases may be missed.
The timing also matters. The study window closed on August 22, 2026, before ads went live in the European Union on August 24 and before later ad formats and measurement changes. It is an early read on the rollout, not a final performance benchmark for the entire ChatGPT Ads program.
What the finding means for AI search marketing
The result connects with a broader pattern in AI search: AI assistants can influence discovery before they replace established acquisition channels. Tech Help Canada’s guide to answer engine optimization covers that shift from a visibility perspective. This new study adds an advertising angle, showing that visibility inside an AI assistant does not automatically produce paid-search-level revenue per visit.
For businesses testing ChatGPT Ads, the useful comparison is not hype versus skepticism. It is tracked revenue per visit, conversion rate, order value, cost per click, and any lift that appears outside last-click attribution. Without those numbers, moving budget from proven search campaigns into AI ads is more of a research spend than a performance decision.
Get new small business insights by email
Practical ideas and useful articles to help you make better business decisions.
Tech Help Canada Staff researches, writes, and reviews practical content for business owners and professionals. Our coverage spans business, marketing, SEO, technology, and the tools and systems people use to grow and operate online. We focus on clear, useful information backed by research, hands-on experience, and editorial review. Learn more about our team and editorial standards. Need help with something? Contact Us
