ChatGPT ads · Australia · what's proven

Do ChatGPT ads work? What's known so far

Nobody has published a result from an Australian trade business yet, and that includes me. What we have is OpenAI's own numbers, and one US business that published its own test. I'll show you both, and where each one stops.

Book a free 30-min strategy call See how I run them for you

Money is going in, so somebody thinks it works

In September 2026 OpenAI said its ads had passed $1 billion in annualised revenue run rate, in under 200 days. It didn't say which dollars. Run rate means the yearly figure you get from the current pace, not money already banked. OpenAI also says it has tens of thousands of advertisers.

That tells you the thing is real and busy. It doesn't tell you a word about a sparky in Penrith. Big spenders keep spending for reasons a one-van business doesn't share.

OpenAI's 2 examples, and what they're not

OpenAI published 2 examples of its own. An online shop got a 3x return on ad spend across its campaigns over 28 days. That's three dollars back for every one spent, in whatever currency they spent. And one of its technology partners reported that more than 80% of the traffic ChatGPT ads sent was new customers.

Read those as what they are. They're OpenAI's own examples, in OpenAI's own post, about OpenAI's own product. No outside body checked them. Neither one is a trade business, and neither one is Australia. I would want the same caution if I were selling you a van.

One US business published its own test, and it went badly

A US coaching business wrote up what happened when it ran the ads in July and August 2026. First run: US$500 gone in about an hour. The account reported 140 clicks, about US$3.50 each. No leads. And only about 13% of those clicks turned up in their own site stats.

They went again a month later. US$175 over a week this time, and 68% of the clicks arrived on the site. Their tracking counted a single action. Nothing they'd call a lead.

The fair bits, in their own words. Setup was easy. The cost per click was reasonable. The dashboard took about a day to fill with numbers. Billing arrived as several part charges. That's a single business, in the US, selling coaching. It is not a yardstick, and I don't use it as one.

The first run's real problem was counting

Look at that 13% again. They paid for 140 clicks and could only see 13% of them land. So they couldn't tell a bad ad from a bad page from a counting fault. When most of what you bought can't be seen, you're guessing.

OpenAI gives you 3 ways to count what happens after the click. The first is a direct feed from your own booking or job system into OpenAI. Second, a small bit of code you put on your website, called a pixel. Third, a tag on the end of your link, which stays on through the click and shows up in whatever stats tool you already use. Set all that up before a dollar goes in, not after. Most of that first US$500 bought a lesson instead of a customer. How to know which calls came from ChatGPT is the page for that, and 7 ways a campaign wastes money has the rest of them.

What nobody has published yet

No result from an Australian plumber, sparky, roofer or any other trade.
Nothing independent on what these ads do for a local service business.
No Webbie client result. I've got none to show you, so there is none on this page.
Nobody has published how many Australians see these ads, or how often.

So an average click cost or a typical result for Australian trades is a guess, whoever quotes it. I went looking. It isn't published anywhere I can trace.

So the only answer is a test

You find out by running it for a month and counting. The question is who pays to find out. If you do it yourself, that's your money and your time in a new place to advertise with no track record here.

That's the reason I built the offer the way I did. We work for you for free, and you still cover your own ad spend. Where that ad spend goes is on what ChatGPT ads cost. The deal itself is in the block below.

4 questions people ask me before they try it

If there's no proof, why would I bother?

Because the cost of finding out is a month of capped ad spend. From what I see, almost nobody in your trade is in there yet. By the time there's a published Australian number to read, more of your competitors will be in there. That's the trade-off, said plainly, not as a sure thing.

What happens if the ads bring me nothing?

If the ads make you nothing, you pay us nothing. You'll have paid OpenAI for the clicks, and that is all. We'll have done the work for no income. Then we both know, and you can put that budget into what already works for you. Finding that out in a month beats selling you a year of it.

Why did that US test do so badly?

Two things, from their own write-up. Most of their clicks never showed up in their site stats, so they couldn't see what they'd bought. And the spend went out fast, inside an hour on the first run. Counting first and a daily cap fix both. It still isn't proof either way for a trade.

How will we know it worked for me?

We agree what a job is worth to you before anything starts. Then we count the calls and the jobs that came through the ad, and put your number against them. No stories about views and reach. The test is whether the work you got is worth more than the ad spend you put in.

Let ChatGPT show your business to high-intent customers

ChatGPT has released paid ads, and this changes the marketing game absolutely.
If you missed out on Facebook ads in 2007, where anything you posted would have popped off, don't miss out on the next big ad platform.
Unlike Facebook, where people get interrupted mid doom-scroll, ChatGPT shows your business to people who are already researching something related.
That makes ChatGPT a cheat code. The engine your customers ask for advice shows your business right under its answer.
Most marketing agencies charge you thousands before you see any results.
We work for you, for free. You just cover your own ad spend. The first month is free, so you see the results for yourself before you pay us anything.
Then, if you want to keep getting those results month after month, we sort out a rev-share or a retainer. A rev-share is a cut of what it makes you. A retainer is a flat monthly fee. Agreed in writing before month 2 starts.
In the free month, we don't get paid if you don't get paid. After that, it's the rev-share or the retainer you chose.
Only 10 a month. The work up front is heavy and brings us no income at first. That is all we can take.

If the ads make you nothing, you pay us nothing and we stop there. You will have paid OpenAI for the clicks, and that is all.

Book a free 30-min strategy call The whole offer on one page

30 minutes, then a straight answer

30 minutes on the phone, mostly questions about your business and your customers. You get 2 straight answers. Do people ask ChatGPT for what you do? And is it worth a run, where we only get paid a rev-share or a retainer, sorted once you have seen the results, to find out?

Call or text 0491 681 746 Email instead