Six months after the launch of OpenAI’s advertising platform, ChatGPT Ads still raises questions among advertisers. The figures vary widely: cost per click (CPC) ranges from less than $3 to $13, and a profitable CPM is not always clear. Some campaigns bring quality leads at a competitive price, while others yield minimal results. The problem is that there are no clear criteria for evaluating effectiveness, and the platform’s reporting is limited.
Reporting is available at the campaign, ad group, and ad levels. Conversion tracking is implemented through OpenAI Pixel and Conversions API, supporting actions such as purchases, leads, and registrations. OpenAI can also use modeled conversions. This is where the reporting functionality ends, which complicates price comparison and the assessment of real ROI.
Reporting Limitations and Competition
The lack of comprehensive competitive reporting is a critical issue. If CPC increases, current data does not allow understanding whether this is due to increased competition, changes in relevance, inventory volume, bids, or the types of dialogues in which ads appear.
This hinders accurate cheap hypothesis testing and campaign optimization.
Hostinger’s Experience: $70,000 on Tests
One of the largest public tests was conducted by Hostinger, spending almost $70,000. Hussein Ograk, Head of PPC, initially noted that CPC was no higher than in Google Search, and campaigns generated purchases. However, as the budget increased, his assessment became more nuanced.
“CPM exceeded $65, but the main problem was CTR. More specific cases worked better, while broad messages consistently failed. Traffic quality was unstable, and evaluating ROAS based solely on direct conversions proved difficult,” Ograk noted.
This example demonstrates that what CPM to consider profitable is a matter of context and traffic quality.
Other Tests: Attribution and Lead Quality
CTC reported attributed revenue ranging from $19,000 to $38,000, depending on the attribution model, with ROAS from 3.3x to 6.8x. They used Triple Whale to assess downstream performance, as ChatGPT Ads reports were insufficient.
Floyd Blakey, analyzing a B2B campaign with a budget of $7,000 CAD, recorded an average CPC of $9.29, CPM of $64.34, and CTR of 0.7%. Using visitor de-anonymization, her team identified 146 organizations from 336 paid clicks. Of these, only five matched the ideal customer profile.
This highlights the problem of traffic quality and shows that cheaper than influencers does not always mean more effective.
Geographical Differences and Recommended CPC
The Synter test showed significant geographical differences. With an overall CPC of $9.89 (at $4,428.84), it ranged from $5.10 in the UK to $10.62 in the US. In Australia, CPC reached $17.59, in New Zealand — $22.89, although volumes there were significantly lower. Cost per click heavily depends on the region.
OpenAI recommends a starting maximum CPC from to . However, this recommendation is often perceived as a benchmark, although it is not. This creates false expectations regarding a budget alternative to seeding and how much 1 million impressions cost via banner compared to AI advertising.
ChatGPT Ads Audience
Advertising in ChatGPT is only available to Free and Go plan users, who constitute the majority of the audience. More expensive subscriptions (Plus, Pro, Business, Enterprise, Edu) remain ad-free, as do accounts of minors. This means that the advertising audience does not coincide with the overall ChatGPT user base.
There is insufficient public information about the demographics and purchasing power of this audience. General ChatGPT user statistics are also irrelevant, as they include those who do not see ads. Geography further narrows the comparison.
- Pilot launch: USA, Canada, Australia, New Zealand (February).
- Expansion: UK, Japan, South Korea, Brazil, Mexico (May-August).
- Access to Ads Manager: 52 countries (August 31).
How to Approach ChatGPT Ads Now
When starting to test hypotheses cheaply on this platform, it is important to understand its current limitations. Set expectations before spending, based on what will make the channel profitable for your business. Someone else’s CPC or CTR will provide little value until we know so little about the competition, audience, and types of dialogues behind these results.
Early results require more thorough analysis than on mature platforms. Due to a lack of diagnostic signals, it is difficult to distinguish a real performance problem from display peculiarities, relevance, or available dialogues during the test period. There is no universal number yet that determines the success of a campaign in ChatGPT Ads. “Good” is what the channel brings to your specific business, not abstract figures.
Frequently Asked Questions
What is a profitable CPM in ChatGPT Ads?
A profitable CPM in ChatGPT Ads does not have a fixed value. It depends on the niche, traffic quality, campaign goals, and the ROI you receive. Hostinger’s experience showed a CPM above $65, but with CTR issues. It is important to evaluate not only CPM but also the quality of leads and the final ROAS.
Is it worth investing in ChatGPT Ads for a cheap hypothesis test?
ChatGPT Ads can be a tool for a cheap hypothesis test, but with caveats. Limited reporting requires additional analytics tools (e.g., Triple Whale, visitor de-anonymization) for a full evaluation. Be prepared for the price vs. result to vary greatly.
What’s cheaper: banner ads or ChatGPT Ads?
Comparing what’s cheaper: banner or blogger with ChatGPT Ads is not directly correct. CPC in ChatGPT Ads can be significantly higher than in traditional banner advertising, reaching $10-$20 in some regions. However, if the quality of traffic and ROAS justify these costs, then ChatGPT Ads can be more effective. It is important to calculate unit economics.
What CPC should be considered normal for ChatGPT Ads?
OpenAI recommends a starting maximum CPC of $3 to $5, but this is not a benchmark. The actual CPC can be significantly higher, up to $13-$22. A “normal” CPC is one that provides a positive ROAS and aligns with your unit economics. Do not rely solely on recommendations; always conduct a cheap hypothesis test and analyze your data.
Conclusion
ChatGPT Ads is a new platform with potential, but also with serious limitations, especially in terms of reporting and transparency. Advertisers accustomed to detailed statistics face the need for deep off-platform analytics to assess real effectiveness. To understand how much money each ruble spent will bring, do not rely on general figures, but focus on your metrics and goals.
Conduct a cheap hypothesis test, analyze each stage of the funnel, and optimize campaigns based on your own data and expert opinions. This is the only way to find your profitable CPM and achieve success.