TL;DR: ChatGPT ads in Australia now support pixel and Conversions API tracking plus CPC bidding, after OpenAI shipped the measurement tools advertisers asked for at launch.
One month into running the channel for our own PPC services (9,263 impressions and 80 clicks across two ad groups since 30 May), the clearest finding is a creative one: an ad with a real human face out-performed every graphic-led variation we ran.
OpenAI has also brought the major agency holding companies into the pilot, which tells you where measurable spend is heading next. JBE Digital is one of the few Australian agencies running live spend on the platform now.
What’s Changed Since Our First Week
We published our first week of ChatGPT ads data at the start of June. The headline then was that the platform had opened up from its enterprise-only launch into self-serve, click-based campaigns with basic tracking.
A month on, the bigger shift is measurement. OpenAI has launched a Conversions API and pixel-based measurement, so advertisers can track what happens after the click: form submissions, leads, purchases, signups. Cost-per-click bidding has been added alongside the original CPM model.
This matters more than it sounds. The first phase of ChatGPT ads forced advertisers to guess, because the platform showed clicks with no view of what they produced. Proper tracking turns the channel into something an agency can optimise and defend to a client.
Founder Joel Brooker made that point to AdNews in June, noting that the agencies moving now will have months of optimisation data behind them before most businesses have started. The measurement fix is what changed the commercial case for the platform.
The Numbers After One Month
We are running two ad groups since 30 May, one built around PPC and Google Ads messaging, one built around Meta ads messaging. Here is where they sit cumulatively.
| Ad group | Impressions | Clicks | CTR | Avg CPC | Spend |
|---|---|---|---|---|---|
| PPC | 6,914 | 65 | 0.94% | A$7.77 | A$504.78 |
| Meta | 2,349 | 15 | 0.64% | A$9.87 | A$148.01 |
A few things worth pulling out of the data:
CTR has settled below our first week’s 1.36%, which is what we expected as impressions scaled and the early novelty wore off. CPC has drifted up across the month rather than down, sitting around A$7.77 on the stronger group. The PPC-led messaging is outperforming the Meta-led messaging on click-through rate and cost per click, a useful signal about which conversations on the platform are worth showing up in.
These are modest budgets, and a few hundred dollars of spend settles nothing on its own. What it buys is specific, repeatable learning about creative and targeting that applies when budgets and conversion data grow.
The Creative Finding That Surprised Us
Across every ad we ran, the single best performer on click-through rate was the one featuring a real human face, a photo of an actual person rather than a designed graphic. It reached a 1.13% CTR over more than 3,000 impressions, ahead of every polished graphic variation in the account.
The second strongest pattern was problem-led copy. An ad that called out the prospect’s problem directly, then promised to uncover three to five specific things they were wasting money on, pulled the next best click-through rate. Naming the problem and attaching a concrete number to the fix beat messaging built around credentials or years of experience.
Put those two findings together and a clear principle emerges for this platform. The person seeing your ad is mid-conversation, actively researching, reading every word rather than scrolling past. A human face signals a real business behind the ad. Problem-specific copy meets them where their attention already sits. Generic brand creative, the kind that works as wallpaper on a social feed, gets ignored in this environment.
Why The Holding Companies Just Moved In
OpenAI expanded the pilot to include the major agency holding companies, Dentsu, Omnicom, Publicis and WPP, along with technology partners like Adobe, Criteo and StackAdapt for budgeting, bidding and creative.
When the holding companies commit resource to a new channel, it signals where measurable spend is about to flow. The early-mover case we set out in our first article holds: being first on a platform like this buys cheap learning that later arrivals pay full price for. The rest of the industry has now seen the same thing and started moving.
That shortens the window. The advantage of being early on ChatGPT ads is real, and it is no longer a secret.
What This Means For Your Paid Media
ChatGPT ads sit alongside Google Ads and Meta as a third paid environment, with its own rules. The targeting runs on context hints rather than keywords, the creative that wins looks different, and the traffic arrives earlier in the buyer’s research than a high-intent search click.
The measurement gap that made the channel hard to recommend has largely closed. Now that the pixel and Conversions API are live, the discipline that matters is the same one that has always separated paid media that compounds from paid media that drains budget: tracking the actions that matter, then optimising toward them. We covered why ad performance is really a trust problem in our piece on the Trust Loop in paid media.
For businesses weighing the channel, the work to do now is build experience while competition is thin, with proper tracking from day one. Our ChatGPT advertising management is built around exactly that.
Frequently Asked Questions
Can you track conversions on ChatGPT ads now?
Yes. OpenAI launched a Conversions API and pixel-based measurement in 2026, letting advertisers track actions after the click such as form submissions, leads, purchases and signups. This was the single biggest limitation at launch, and closing it is what makes the channel viable for performance-focused advertisers and agencies.
How much do ChatGPT ads cost in Australia?
In our own campaign over one month, average cost per click sat between roughly A$7.77 and A$9.87 depending on the ad group, with CPMs in the A$60 to A$75 range. These are early figures from modest spend and will move as more advertisers enter the auction. CPC bidding is now available alongside the original CPM model.
What kind of ad creative works best on ChatGPT ads?
In our testing, ads featuring a real human face out-performed designed graphics on click-through rate, and problem-led copy that named a specific issue beat credential-led messaging. The platform places ads inside active research conversations, so creative that feels human and speaks to the user’s exact problem tends to earn more engagement than polished brand graphics.
How are ChatGPT ads targeted?
ChatGPT ads use contextual targeting through context hints rather than keywords or behavioural profiles. You describe the conversations your business belongs in, and the system matches your ad when a free tier user has a relevant conversation. Getting the breadth of those hints right is the core skill of campaign setup.
Are big agencies using ChatGPT ads?
Yes. OpenAI expanded its advertising pilot in 2026 to include major holding companies including Dentsu, Omnicom, Publicis and WPP, along with technology partners such as Adobe, Criteo and StackAdapt. Large advertisers committing resource to the channel signals growing confidence in it.
Is it too late to get an advantage on ChatGPT ads?
No, though the window is narrowing. The largest agencies have now entered the pilot, so the first-mover advantage is no longer a secret. Businesses that start building experience and optimisation data now will still be ahead of competitors who wait until the channel is fully mainstream.
The Strategic Implication
Most businesses will treat ChatGPT ads as a separate experiment, a box to tick when the channel feels safe and mainstream but that instinct is the expensive one.
The shift underneath this is the same one reshaping search. Buyers are researching inside AI conversations, comparing options in real time, forming opinions before they ever reach a website. Paid placement inside those conversations is one lever. Being the business the AI recommends organically is the larger one, and both run on the same foundation of visible credibility.
The agencies and businesses building experience now are buying cheap learning and months of optimisation data. The platform has solved its measurement problem, the creative rules are starting to surface, and the cost of entry is still low.
The brands that compound through this shift are the ones treating AI visibility as core infrastructure, paid and organic together. The ones waiting for permission will pay full price for the same lessons later.
