The other side of ads in ChatGPT: Advertiser perspective

Since OpenAI announced it would start testing ads in ChatGPT, and the guidelines for it, they did not discuss in depth how it is gonna work for businesses interested in advertising with them.

What we know is that users’ data remains private and that the advertisements will not come as a suggestion, as managed by Google, but rather as a solution to the inquiry being discussed in the chat thread. This way, ChatGPT explores another way of advertising: a natural, organic conclusion for the conversation flow. 

And while we already discussed how it is gonna work and its implications for the audience, now it is time to take another approach: advertisers. Because, on the other side of the coin, are the people who will pay and benefit from it: the moneymakers.

The who, the how, the afterwards

OpenAI has not released an official statement on who can advertise with them. Whatsoever, the company talked with ADWEEK, explaining how for the trial, they have asked selected advertisers to commit to at least $200,000.

Some of the first known companies are Adobe, Target, or Albertson, as well as the group WPP media, and others.

As for the testing process, some answers are absolute: the trial was invite only. Whatsoever, overtime and with the expansion of the advertising model in ChatGPT, it will be open for companies to subscribe through the OpenAI ‘Advertise with ChatGPT ‘ website, as long as they comply with the guidelines and pay the threshold.

But whether this strategy will be successful, we need to see beyond the surface. The group Adthena has released the first study of the trial.

First of all, how does it even work? Well, the platform is not self-serving and does not have a login interface.

Advertisers are managed manually: companies introduce specific words that have to be prompted by the user for the business to be advertised. In other words, a flower shop may type “flowers”, “anniversary gift”, “surprise for my girlfriend”, to be recommended within the conversation.

And the initial feedback companies received to improve their strategies does not tell much: core metrics limited to campaign name, impressions and clicks.

This logic encourages story telling more than click bait, while creating an equal opportunity and challenge for companies: Now there is a direct bid on the playground.

Small companies and startups can capture possible customers by finding a way to outshine their competitors, especially since many users use ChatGPT to compare. 

Will this be another Google Ads?

The introduction of ads in ChatGPT led to the inevitable question if it will slowly transform into another Google Ads, or similar search engines. And to answer this question we can go through memory lane and compare it with the launch of Google Adwords.  

Google Adwords built its strategy by monetizing explicit intent: keyword-auction where the advertisers bid over the typed demands. It implemented the PPC (pay-per-click) model which is still prevailing, and provided complete control and visibility to the advertisers, giving them the possibility to build the campaigns and bids manually, and constantly monitor performance to improve their indicators.

This is completely the opposite to what OpenAI is doing nowadays compared to Google Ads. 

The most significant shift is the flip in the model of how intent is now captured and capitalized. Google keeps the keyword-expressed demand original model, where users declare intent by search queries, and advertisements are offered to audiences targeted on interest and habits, demographics, and browsing behavior.

ChatGPT captures intent differently. Instead of monetizing keywords, it interprets context. Intent unfolds through constraints, preferences and goals, rather than just a search query.

This makes conversational ads structurally different from search ads. The shift now is not technical: it is structural. The AI acts as a curator and gatekeeper of commercial relevance, moving from user-driven comparison of ranked results to AI mediated recommendation. 

Implications for companies

Another shift implemented by OpenAI is the cost structure, since they will adopt a premium $60 CPM (cost-per-mille), which means that the advertiser will pay $60 per every 1000 times their ad is shown.

This adds pressure to the companies, since an inadequate management of the chosen keywords will be translated directly to their expenses, without the expected results.

If this model is maintained by OpenAI, advertisers will be forced to not only understand the context, but the whole conversational ecosystem where the advertisement would be relevant. 

In the same way, ChatGPT’s performance measurements drastically differ from the detailed analytics and automatic campaign optimization that Google offers. The limited feedback that OpenAI offers to their advertisers also plays an important role in how companies have to think about their campaigns. 

Now they are competing for algorithmic selection.

This new framework alongside limited targeting control, suggests that advertisers must learn how the AI’s curation logic works. Success will depend less on bid strategy and more on understanding conversational intent. 

To design effective AI Ads campaigns, brands will need to anticipate how users articulate needs across dialogue (prompt behavior), and not rely on direct input or requests.

Advertisement has always been about solving problems and positioning solutions. Now companies must go further and understand how their audience thinks, and frame specific problems inside a conversation with ChatGPT.

Platform and model limitations

OpenAI introduced a game changer model for the advertisement industry by launching contextual-based ads. However, alongside the benefits come the disadvantages, and the platform and current model introduce important questions for brands to be aware of.

If the algorithm determines which advertisement is more relevant and only suggests one curated suggestion, it could represent the beginning of AI-controlled demand allocation and reduce advertiser competition visibility, unless the user specifically asks for it.

If ChatGPT continues with its current curated model without an interface where advertisers can manage their inputs, the model could create power concentration in the AI platform.

Because the platform takes all the decisions and advertisers cannot directly manipulate visibility. In the same way, if advertisers are unable to  analyze the performance measurements, it could create a black box monetization problem, where brands are paying but do not have straight data that enables campaign optimization.

From auction to algorithm

ChatGPT ads mark a new era for advertisement. The real test is how brands adapt to this new format while navigating platform limitations, and preserving the trust that characterizes ChatGPT interactions. 

The beta is still trial and changes are likely. For now, the system favors large corporations with the experience and resources to optimize their campaigns without access to detailed OpenAI performance data.

As for the industry, it is safe to conclude that the limitations and specifications that conversational AI requires will lead to new ways of advertising. Advertisers must focus on creating a message that adds value to the conversation, respecting users’ needs and their relationship with the platform.

Prioritizing credibility, and aligning brand presence and product differentiation while addressing the specific problems users are exploring in their chats, will be the key to success when advertising with ChatGPT.