Adthena reported 204,489 observed advertising placements in ChatGPT between June and September 2026. The figure comes from its Holiday AI Paid Search Report, published on November 4, which compares activity in ChatGPT, conventional Google Search and Google's AI-generated answers. It represents monitoring by one research vendor, not a platform-wide count of ad impressions released by OpenAI.
According to Adthena's announcement, the number of advertisers it identified on ChatGPT increased almost sixfold over two months. Apparel, consumer electronics and travel accounted for roughly 95% of placements in the observed data. Toys and games was the fastest-growing category. The public summary does not provide a complete country-by-country breakdown or independent validation of overall platform volume.
Intent matters as much as reach. Adthena classified 83% of beauty-related ad placements in its sample as early-stage research rather than ready-to-buy searches. The equivalent shares were 69% for travel and 67% for toys and games. Those proportions are not conversion rates, nor do they establish the return generated by any particular advertiser.
The report also documents growth in Google's AI Overviews. Their appearance count rose year over year by 536% in Australia, 160% across monitored UK/EU markets and 89% in the United States. Adthena separately identified 175,100 paid placements inside Google's AI answer surfaces in the US. The number of AI answer appearances should not be confused with the number containing advertising.
For performance marketers, the practical implication is to distinguish discovery from purchase-intent traffic. An advertisement seen while someone asks which product to choose may affect consideration without delivering an immediate sale. Creative, landing pages and measurement therefore need to be evaluated against the actual query stage.
A placement count alone cannot establish CPA, ROAS, incremental revenue or the likelihood of an outbound click. Teams assessing new AI inventory should demand comparable attribution windows and downstream conversion evidence, and treat the vendor figures as a directional market signal rather than a forecast for every account.

Комментарии