ChatGPT ads face conversion hurdles despite growing brand participation

The scale of the new ad market

OpenAI has officially entered the advertising space this year, moving toward a model where ads are integrated into the user experience. While the platform is still finding its footing, it already handles roughly one-fifth of the search queries that Google processes. This represents a massive potential audience for brands looking to reach users in a conversational environment.

The number of brands participating in this new ecosystem is growing at a rapid pace. Data from Adthena shows that the number of distinct brands running ads on ChatGPT jumped from 1,082 in June to 6,333 in August. This surge suggests that companies are eager to secure a foothold in the AI-driven search landscape as the platform expands its capabilities.

Despite this growth, the actual reach of ads varies significantly by region. In the U.K., 45.7% of ad-eligible users see ads, while in the U.S., that figure is 43.4%. In contrast, fewer than 27.8% of eligible users in Mexico are currently reached by ads, which may indicate different stages of rollout or monetization strategies.

Diversifying beyond retail

When the ad pilot first launched in February, the majority of the spend was concentrated in the shopping sector. However, recent data from Sensor Tower shows a shift toward a more diverse range of industries. Shopping's share of U.S. ad spend on the platform fell from 38% in April to 20% by September.

During that same period, other sectors have seen notable increases in their share of the ad pie. Financial services grew from 2% to 13%, while software moved from 13% to 18%. Travel and tourism also saw an uptick, rising from 5% to 9% as more brands look to capture different types of consumer intent.

This shift suggests that ChatGPT is becoming a broader tool for various industries rather than just a niche channel for retail shoppers. As more sectors experiment with the platform, OpenAI will need to refine how it serves these different types of ads to ensure they remain relevant to the specific needs of each industry.

The conversion problem

While the reach is expanding, the actual economic results for advertisers are currently underwhelming compared to traditional search. A research paper based on e-commerce data found a stark difference in how much money is generated per session. Paid ChatGPT visits resulted in $0.89 in revenue per session, while paid Google Search visits yielded $2.45.

This gap is largely driven by conversion rates, which measure the percentage of users who complete a desired action. The study found a conversion rate of only 0.37% for paid ChatGPT traffic, which is significantly lower than the 1.17% rate seen for Google. These figures suggest that while users are interacting with ads, they are not moving toward a purchase as quickly as they do on Google.

The discrepancy highlights a primary challenge for OpenAI: proving that its ads can compete with established performance channels. For advertisers to commit larger budgets, the platform must find a way to turn conversational interactions into more reliable sales outcomes. Currently, the data suggests that the ads are reaching people, but they aren't yet closing the deal as effectively as older technologies.

Technical and pricing hurdles

The difficulty in scaling these ads is compounded by ongoing measurement and attribution issues. Accuracast's group CEO Farhad Divecha noted that even with new partner integrations, there are still discrepancies in how ads are tracked. He mentioned that while some are 'teething problems' due to the new system, some issues stem from the way attribution is being messed up.

From a pricing perspective, OpenAI is leaning toward CPC (cost-per-click) and outcome-optimized bidding. These models are generally seen as more positive for revenue than simple impression-based pricing because they reward actual user interaction. However, some researchers caution that the estimated cost of $5 per click might be too high given the current conversion rates.

Looking ahead, models suggest that commercial chats on ChatGPT could grow to nearly 380 billion by 2030. To reach those targets, OpenAI must bridge the gap between high-volume chat interactions and the specific performance metrics that brands demand. The success of the platform's business model will depend on its ability to solve these technical tracking hurdles and improve the underlying conversion value for its partners.