ChatGPT Ads Reach 6,333 Brands but Convert at a Third of Google's Rate

By Tom Whitfield

Paid ChatGPT traffic converts at 0.37%, roughly a third of the 1.17% conversion rate for paid Google Search visits, according to a working paper under review by Maximilian Kaiser and Christian Schulze. The same research, based on data from ecommerce websites, found paid ChatGPT sessions generate $0.89 in revenue for U.S. advertisers, against $2.45 for Google. The gap sits at the center of OpenAI's advertising push, which only began this year but has already drawn 6,333 distinct brands.

ChatGPT now handles approximately a fifth of the search queries Google processes, yet its ads business is worth considerably less. OpenAI started selling ads only this year, the format is still taking shape, and the company has moved cautiously on how much advertising to place in front of users. Advertising, in short, lags usage — and OpenAI wants that to change quickly.

How much of the audience can advertisers actually reach?

Ad coverage remains below half of eligible users even in OpenAI's most established markets, according to data from Similarweb and Graphite. The U.K. leads at 45.7% of ad-eligible users seeing ads, followed by the U.S. at 43.4% and Australia at 43.3%. In Mexico, the figure falls to 27.8%.

The differences may reflect how far each market has progressed in the rollout, or how much of its eligible audience OpenAI is currently monetizing. Either way, fewer than half of eligible users in any tracked market are being reached.

Which sectors are spending?

The advertiser mix has broadened well beyond retail, according to Sensor Tower data. Shopping accounted for 38% of U.S. ad spend on the platform in April; by September that share had fallen to 20%. Over the same period:

• Financial services rose from 2% to 13%
• Software grew from 13% to 18%
• Travel and tourism increased from 5% to 9%

At the pilot's launch in February, Sensor Tower estimates around 42% of ads came from shopping advertisers, followed by jobs and education (10%), food and dining (9%), software (8%), media and entertainment (8%) and travel (7%).

Brand counts confirm the expansion. Adthena tracked 1,082 distinct brands running ads across the U.S., Australia, the U.K. and the EU in June, rising to 3,741 in July and 6,333 in August. September data, running only to Sept. 20, points to a projected 6,830 for the full month.

What does this change for media buyers?

New Street Research models commercial chats growing from 45.9 billion this year to nearly 380 billion by 2030, with the proportion of chats carrying ads rising from 20% to 28%. OpenAI has said CPC and outcome-optimized bidding account for the majority of its ad campaigns — a higher proportion than New Street had assumed. CPC ads typically generate more per ad event than impression-based pricing, which could lift revenue.

But New Street cautions that monetization may currently be weaker than it originally modelled, and its estimated CPC of around $5 could be too high.

Measurement problems compound the performance question. Farhad Divecha, group CEO of Accuracast, said the latest partner integrations — Magnite and PubMatic — have not necessarily resolved underlying tracking issues.

"We continue to see discrepancies there," Divecha said. "Some are teething problems because it's a new system that has to be set up on advertiser systems, but some problems are down to attribution being messed up."

For practitioners, the practical calculation is straightforward: inventory is scaling fast, sector competition is widening, but per-session economics still trail established performance channels by a wide margin. Until conversion rates and attribution close that gap, ChatGPT remains a test budget line rather than a proven performance channel — and the next tranche of OpenAI measurement fixes will determine how quickly that changes.