From your own account and site
- What OpenAI's Ads Manager reports for your campaigns
- Clicks arriving on your site, via your own analytics
- Downstream conversions you already track
- Whether your spend produced qualified demand
Measurement
Measuring ChatGPT ads honestly.
The platform is young and the reporting is thinner than Google or Meta. Knowing exactly which questions your own data can answer, and which it cannot, is what stops a campaign being judged on the wrong number.
Your account can tell you what your ads did. It cannot tell you whether the result was good for the category. That comparison has to come from outside data.
The boundary
What is knowable, and what is not.
Most confusion about ChatGPT ad performance comes from expecting a number that does not exist anywhere outside a private account.
From captured placements
- Which advertisers appear in your category
- How often each one is observed
- What their creatives actually say
- Which prompts surface ads at all
Not knowable by anyone outside OpenAI: a competitor's click-through rate, conversion rate, budget or bid. Those live inside each advertiser's account. Any tool claiming to report a rival's ChatGPT ad CTR is inferring, not measuring — and we will not do that here.
Category context
A number means nothing on its own.
The same click-through rate can be a strong result or a weak one depending entirely on how many advertisers were competing for the conversation.
Before concluding a campaign underperformed, establish how contested its category is. A modest result in a category with dozens of active advertisers may be outperforming; a comfortable-looking result in a category with almost none may be leaving demand on the table. That comparison is the one piece of context your own reporting can never supply.
Primary source
Check the platform directly.
Reporting and tracking features are changing as the beta expands, so treat OpenAI's own documentation as the authority rather than any third-party summary, including this one.
FAQ
Measurement, answered.
Including the questions where the honest answer is “nobody can know that”.
How do I measure ChatGPT ads performance?
Measure in two layers. Inside your account, use whatever OpenAI's Ads Manager reports plus your own site analytics and conversion tracking for what happens after the click. Outside your account, compare the result against how contested your category actually is, because a number means little without knowing how many advertisers were competing for the same conversations.
Why do impressions look high while clicks stay low?
On a conversational surface an impression means your ad was eligible for a conversation, not that it suited it. Broad context hints generate eligibility on questions your product does not answer. That is why impressions are the least informative number here and relevance is the variable worth optimizing.
What can I not measure about ChatGPT ads?
Anything about other advertisers' results. Click-through rates, conversion rates, budgets and bids are private to each account, so no third party can report a competitor's CTR. Be sceptical of any tool that claims otherwise. What is genuinely observable from outside is which advertisers appear, how often, what their creatives say, and which prompts surface ads.
Is there a benchmark CTR for ChatGPT ads?
OpenAI does not publish an official benchmark, and the figures circulating from the beta vary widely enough that a single number is not dependable. A more useful comparison is structural: how many advertisers compete in your category, and whether that number is rising.
How do I prove ChatGPT ads are working?
Tie the click to a downstream action with your own analytics, then judge it against category context rather than a cross-platform average. We track Thousands advertisers across Hundreds categories so you can see whether a result came from a crowded or an open market.