The biggest bombshell to recently hit the advertising world – a world already facing increased competition and fragmentation – was the entrance of OpenAI. Its ad infusions in ChatGPT are slowly rolling out, and the headline has been the program’s growing pains. But are those pains being alleviated?
Before answering that question, let’s step back for context. This all started in January when OpenAI announced it will test ads in ChatGPT. Ads appear at the bottom of chats as small text and image units that resemble paid search, and are targeted based on topics (contextual) and users (behavioral).
These ads come in a few formats and flavors at launch, including a sponsored recommendations box, sponsored links alongside existing citations, and product carousels for in shopping-related queries. OpenAI stresses that sponsorship is always clearly labeled and doesn’t impact organic answers.
As for who’s subject to ads, they’re shown to all ChatGPT users except subscribers of the $200/month Pro plan. Users have some control over the ad experience, including the ability to dismiss any ad and see why they’re being shown specific ads. This is a smart move to counter already existing trust issues.
The What and Why
Beyond the what, an important question is why. Among other things, OpenAI has billions in well-publicized CapEx commitments to fund. It has also faced declining growth rates for ChatGPT paid users. With slowing subscription revenue, markets have become uncertain about its ability to be profitable.
And that last part is particularly damning in light of a potential IPO. In that sense, it’s not surprising that OpenAI would need to build a stronger business case to take on the road. Ad monetization would diversify revenue and build a more stable business case to underwriters and public-market investors.
As for those “growing pains” in the ads program, there are a few. First is a $200,000 minimum campaign spend to participate. Second, OpenAI is charging a $60 CPM, which is pretty steep in online ad terms. Third, ads are sold on an impression basis, as opposed to cost-per-click or performance-based models.
The biggest objection among these factors is that last one, simply because it limits ROAS analysis and ad attribution. This puts OpenAI’s ad program at a competitive disadvantage when compared to Google Paid search and several other options that live up to today’s standards of performance-based metrics.
Land Grab
Back to the question posed above, are these growing pains getting alleviated? This is a good time to check back on that question because it’s something we speculated last month:
Back to the growing pains in its ads program, OpenAI wants to get those out of its system now, so that they don’t hang over its IPO road show. Coming full circle, it’s likely that OpenAI will naturally gain some competence in the art of online advertising as it learns by doing. That’s what we’re seeing play out now.
Since we wrote that line, the evolution has come pretty quickly. For one, the minimum campaign commitment has dropped to $50K. And the infamous $60CPM has reportedly dropped to a $25 to $45 range. And most importantly, tracking pixels and CPC campaigns are now “on the table,” says Digiday.
These updates wholly address the objections stated above, while another key evolution is underway: ecosystem development. First available on a self-serve basis for marketers to figure it out, OpenAI has partnered with Criteo, Smartly, and StackAdapt for ad manager programs to hold advertisers’ hands.
Beyond direct partnerships, third parties are coming out of the woodwork. For example, AdBridge converts Google AdWords campaigns into ChatGPT campaigns, including keywords and strategies. Expect more of this as agencies and ad-tech players rush to the land grab to build ChatGPT ad stacks.
