The reason it is worth the effort is the size of the moment. ChatGPT is running north of 800 million monthly active users, close to a tenth of the planet, and handles around two billion prompts a day. Roughly 70% of all AI assistant traffic starts there. Of the people using it, 63% say they turn to AI for research and discovery and 53% use it to compare products and prices. Consideration is moving into a space where brands could not appear until recently, which is why the question has shifted from whether to be there to what the format actually does once you run it.
How the ad unit and targeting work
The ad sits below the ChatGPT response in a box marked “Sponsored”: a single 1×1 image, a headline, a short description, and a click through to your site.
The targeting is blunter than most buyers assume. It is contextual only, tied to the user’s prompt rather than to what the model replies, with no audience targeting, no demographic overlay, and very little geographic control. Relevance to the prompt is the only thing that decides whether you show.
Precision is what keeps the cost down
Pricing is where testing changed how we think about the channel. OpenAI runs an opaque relevancy model on top of a proto-auction, and in practice that produces two prices. Bid on a niche prompt with no competition and you pay around $15 CPM.
Fight for a contested prompt and win, and it climbs toward $60.
Across brands, CPMs are landing between $25 and $35 depending on how niche the messaging is, and our own testing came in closer to $40.
The pattern is clear: the more tightly your creative maps to the prompt, the better your efficiency.
We would steer away from CPC here. The methodology behind it is too opaque to trust, and bidding on CPM and letting the clicks come has delivered a better effective cost per click.
The measurement problem you’ll hit
Right now you get impressions and clicks, and not much else. There is no conversion optimization yet, so you cannot ask the platform to chase outcomes. What you can do is measure the knock-on effects of those clicks, which means answering one question: do people arriving from ChatGPT show higher intent than they do from native, display, video, social, or search, judged on what they do once they land on the site?
Sharing pixel and conversion data with your partners helps here. The early read is encouraging, with click-through rates coming in comparable to native ads, which is a strong signal for a format this new.
The practical move is to build the evaluation framework before you spend, deciding up front how you will compare ChatGPT intent against site behavior from your other channels.
Buying it: three routes and their price tags
There are three ways in, and the right one depends on what you are trying to drive.
Criteo is commerce-focused and can ingest a product feed, which suits CPG and retail because you can stand up thousands of ads at once; it suggests $50,000 to $100,000 a month.
StackAdapt comes in around $50,000 a month and adds some incentives and geo-targeting.
Going direct to OpenAI is an option too, though the minimum sits near $250,000 a month.
There is little data yet on what the format is best suited to, beyond the fact that it reaches people in a moment of high consideration. So settle two things first: how you want to lead those people to your site, and how you will measure the knock-on effect. Let those answers set the budget, not the other way around.
What to make of it
The format is young, the targeting is crude, and the measurement is thin. None of that stopped it working in testing, and consideration is moving into the chat whether brands are ready or not.
If you have watched programmatic grow up, what comes next is fairly predictable. Audience targeting will arrive, because OpenAI holds a lot of data on its users and has every reason to use it. Conversion optimization will follow. CPMs will climb as more brands pile in, and the auction will get more transparent, until the whole thing looks a lot like the channels you already run.
The brands testing now will carry a head start into that maturity, and the gap closes quickly once costs rise and competition shows up.
