By Anthony Ferry, CEO, Wayvia
Countless businesses have been founded on the belief that if you build it, customers will come. For AI advertising, it seems that if you even half build it, you’ll have them queuing around the block.
That’s been the case for ChatGPT, which says its nascent ad business surpassed $100 million in annualized recurring revenue within two months of launching its test phase. The company reportedly has more than 600 advertisers in its pilot program, including big names such as Target, Ford and Adobe. The company has extended its pilot program beyond its original March deadline by adding more countries, including Canada, Australia and New Zealand.
Advertisers are reportedly queueing up to join. But should they? Beyond having something buzzy to tell their shareholders about on their next investor call, I’ve struggled to see the first-mover advantage for brands here.
The product is lagging behind the hype
If early reports are to be believed, ChatGPT’s initial pilot ad product had some holes. It gave advertisers very little data, lacked automated buying tools and offered minimal consumer targeting. That left advertisers with little ability to measure whether their spend was doing anything, while instructions were still being handled over calls and spreadsheets. Think about that for a second. On a platform with a minimum buy-in of $200,000, brands were committing serious budget without the basic features that many programmatic platforms have offered for years.
Early research from Adthena showed click-through rates running nearly seven times below Google search benchmarks in March — one brand’s CTR reportedly landed at just 0.91%, compared with a 6.4% benchmark typical for Google Search.
The first-mover advantage here was never obvious. Now, looking at the rollout, it seems clear there wasn’t ever one. OpenAI has since created an ads manager and reduced minimum spend to $50,000.
Why is the money still flowing in?
ChatGPT has genuine potential for advertisers in the longer term. There is little doubt that conversational AI will draw in more and more users at the expense of all other parts of the internet, including retailers’ and brands’ own websites and conventional search. Where consumer eyeballs go, advertisers will follow.
Criteo’s aggregated data shows that users referred from LLM platforms like ChatGPT convert at approximately 1.5 times the rate of other referral channels. At Wayvia, our own ecommerce tracking data from 2,000+ brands spanning all major retailers tells a similar story. When ChatGPT delivers a customer, they’re more likely to convert. That’s probably because customers do their research in ChatGPT, so by the time they click on a product link, they’re more likely to be ready to buy.
That shopper intent is likely why brands are willing to put up with the bumps. Nearly 80% of small and medium-sized businesses involved in the pilot have signaled interest in continuing. There’s also some reason to the argument that brands that start testing and learning sooner will be in a better position in the long run. But willingness to tolerate a broken product is not the same thing as getting value from it.
To be clear, OpenAI is serious about getting advertising right, and the signals are encouraging. Criteo became the first advertising technology partner to integrate with OpenAI’s advertising pilot in early March. OpenAI has partnered with Smartly to help optimize campaigns in real time, and has held early-stage discussions with The Trade Desk about further scaling ad sales.
Most importantly, OpenAI has hired David Dugan, who spent more than a decade at Meta, most recently as VP of global clients and agencies, to head up its advertising push.
OpenAI faces a projected $15 billion cash burn in 2026, which means the pressure to get the product right is enormous. They will fix the targeting. They’ll fix the measurement. But make no mistake: they are building the car while driving it.
So why wait?
Because you don’t have to be the one subsidizing the fixes.
There is no inventory shortage coming. Roughly 85% of OpenAI’s free and low-cost users in the U.S. are eligible to see ads, but fewer than 20% are shown them on a daily basis, which means there’s enormous untapped supply sitting there right now. The audience isn’t going anywhere. The targeting, measurement and self-serve tools will get built because OpenAI can’t scale a real ads business without them.
When those things exist, brands will be able to evaluate ChatGPT advertising with data, as they do with every other advertising channel. Until then, my advice is to let OpenAI figure out the kinks on someone else’s brand and budget.
That’s not being late. That’s being smart.
About the Author
Anthony Ferry is co-founder and CEO of Wayvia, the global leader in omnicommerce data and brand enablement. By connecting shopper and retail intelligence across every channel, brands gain deeper insight into consumer behavior and unlock new opportunities to improve the path to purchase, whether through offsite media, onsite experiences or agentic commerce. Backed by the world’s largest network of retailer and media partnerships, Wayvia offers brands the retail intelligence to power analytics, optimize shopping journeys and enable AI solutions. For more information, please visit wayvia.com to learn more.
