ChatGPT ad delivery struggles are testing advertiser patience

It’s been a few months since OpenAI officially began testing ads inside ChatGPT, and things are progressing quickly. Marketers shelled out minimum ad commitments — up to $250,000 in some cases — seemingly based on the fear of missing out on the AI visibility boom. Now, those same marketers are reporting under-delivery issues.

“There were some [advertisers] not even getting a tenth of their $250,000 commitment used at that point,” Krystal Scanlon, senior platforms reporter at Digiday said. Scanlon spoke on a recent episode of the Digiday Podcast. While those dollars sit in what could be best described as marketing budget purgatory (reserved for OpenAI ad commitments), OpenAI is erring on the side of caution, per Scanlon.

“They want to get ads right. Albeit, probably not ideal for an advertiser, wanting those ads to be out there and to get the reach from those ads,” she added.

Under-delivery and pain points

OpenAI seems to be trying to strike a better balance. The fill rate is climbing as much as 30% to 50% in comparison to what they were at launch, Digiday recently reported. Still, that leaves at least half of available ad slots unfilled.

Under-delivery is just one pain point for advertisers. Advertisers also gripe about the lack of performance capabilities and basic reporting. 

Again, the tech behemoth is trying to course correct. Ad spend commitments were dropped from the initial $250,000 to $50,000 before being dropped entirely with the introduction of the self-serve ads manager in the U.S. Dugan was brought on to head up OpenAI’s ad business back in March. The former Meta exec was one of many bricks OpenAI laid as the foundation for ChatGPT ads. 

OpenAI introduced cost-per-action (CPA) ads, built out ad measurement infrastructure, starting testing a new ad format and announced plans to test ads in the U.K., Brazil, Japan, South Korea and Mexico. 

OpenAI’s balance between speed and quality

Still, it’s a complicated dance. The playbook that mega digital advertisers like Netflix, Meta, Google and others had years to perfect, OpenAI is grappling with in real time. At the same time, Google’s strengthening its own LLM ad business, recently having rolled out more ads in AI mode. 

“Everything is just moving too fast in a way — just because it’s hard to not only keep up, but probably keep up and keep the quality of what you’re doing,” Scanlon said. 

For example, Netflix found itself in a similar predicament: high demand, low inventory. The streaming platform gave advertisers the option to take their money back. OpenAI hasn’t done as much. It’s less about ill intent, per Scanlon, and more about OpenAI building the plane while flying. 

Playing the long game

All signs point to OpenAI’s ambitions to build out a quality ads business that doesn’t disrupt the user experience. Advertisers seem to be willing to wait for now — the question is how long that patience and FOMO holds on.

“No one will hold back on the OpenAI [ads] because they’re like, ‘Well this is the future. This is the way ads are going. We can’t miss out, even if we’re a bit frustrated,’” Scanlon said.