Intrigue DigitalWhat our A$640 ChatGPT ads test taught us about leads and AI visibility.
The cheapest clicks produced no leads. The coverage of the test ended up cited by Gemini. What a small ChatGPT ads budget taught one founder about advertising, and about getting found in AI answers.
The ChatGPT ad that got us the cheapest clicks produced no leads. Another ad brought in all seven report requests. If I'd looked at clicks alone, I would have backed the wrong ad.
We spent about A$640 testing ChatGPT ads for SignalTo. We got some leads, learnt what needed changing and had the test covered by Industry Contents. Then we used that coverage and the detail behind it on our own website. Later, we saw Gemini citing the material in an answer about advertising.
That last part is something I think a lot of businesses overlook. They get an article written about them, share it on LinkedIn and move on. There's more you can do with it.
First, though, the advertising had to give us something worth talking about.
The test
SignalTo helps businesses understand and improve how they appear in AI answers. For this test, we offered business owners a free report showing how their business appeared. We tried different ways of presenting the same offer.
The ad leading with the free report received 7,800 impressions and 25 clicks. Nobody requested a report.
The ad asking whether AI picks your business generated seven report requests from 27 clicks. That's about 26% of clicks converting, at around A$30 per request.
But the main report campaign cost about A$500 in total, including the unsuccessful ads. Across the campaign, we paid about A$71 per request. The A$30 result tells me which ad deserves another test. The A$71 tells me what I actually spent getting those leads.
The ChatGPT ad that got us the cheapest clicks produced no leads. If I'd looked at clicks alone, I would have backed the wrong ad.
Don't get attached to the ad
I've been working for myself in websites, SEO and affiliate marketing for more than 30 years. You test, you try, you find something that works and then you put more effort into it. There's no point getting attached to an ad because you thought it was a great idea.
My take was that asking whether AI recommends someone's business gave them a better reason to act than simply offering a free report. It made the report relevant to a problem they could recognise.
So we paused, restructured what we were doing and prepared another test around what had worked.
We also ran an agency campaign asking people to book a demo. That got clicks but no bookings. My feeling was that a demo was too much to ask of someone who had only just come across us. A sample report they could look through themselves was worth testing as an easier first step.
Before deciding a channel doesn't work, look at what you're asking people to do. They might be interested in the problem without being ready to give you half an hour of their day.
And follow the leads through. Did they read the report? Were they the right sort of business? Did they reply or become a customer? That's how you eventually work out whether the cost of acquiring a customer stacks up.
Turning press coverage into AI visibility
What happened after the test was just as interesting to me.
Industry Contents covered the results in “How ChatGPT Ads Work: A$500 and the Ad That Got All the Leads”. The A$500 refers to the main report campaign. The A$640 total includes the pilot and agency campaign as well.
We had actual results to share, including the ads that didn't work. That gave the publication something more useful than us saying ChatGPT advertising was the next big thing.
We then made sure our own site had the detail to support the coverage:
- A first-person write-up. What we tested, what we spent, which ads produced requests and which didn't, the method, and what we wanted to try next. Someone considering ChatGPT ads could read it and learn from the money we'd already spent.
- Linked to the coverage. The write-up links to the Industry Contents article, and a main ChatGPT Ads page explains the subject and links to both the coverage and our test.
- Connected to the rest of the site. Links from our homepage, footer and relevant service pages, including our agency and MCP pages.
- Made it machine-readable. We added the pages to our sitemap and llms.txt, and included structured data identifying the author, business and outside sources.
Some of that sounds technical, but the idea is straightforward. If someone or an AI system comes across the article, we want them to be able to find the fuller story, check the numbers and understand that it's our experience.
For SEO, we now had original content answering questions about ChatGPT advertising. Actual figures, a real test and useful lessons, rather than another service page telling people how good we are. The internal links also made that content easier to find.
For GEO, or visibility in AI-generated answers, we had independent coverage alongside the supporting detail on our own site. The same business, the same experiment and consistent information across the pages.
Later, we captured Gemini using the experiment in an answer about advertising. It referred to the results and showed our write-up and the Industry Contents article among its sources.
That's the sort of thing I want to see. Someone researching the topic gets useful information from our experience, and our business is attached to it.
You don't need an ad test to do this
You might have a customer project, a process you improved or a problem you solved that others in your industry keep coming across.
Put the useful detail on your website. Explain what happened, what you did and what the result was. If a publication covers it, link to that coverage from the relevant page and make sure the facts line up.
It doesn't mean writing a new page every time someone mentions you. Sometimes a link is enough. In our case, we had a whole experiment behind the article, so there was plenty worth explaining.
I still judge the advertising on whether it brings in worthwhile business. An article or an AI citation doesn't pay the bills on its own. But the work we put into understanding that campaign also gave us material that other people could report on and AI could cite.
That's a useful return from a small test. Find what works, keep improving it, and make sure the knowledge you've paid to gain doesn't stay buried in your advertising account.
Featured in this article?
Add this badge to your website's About or Press page. Copy the code below and paste it into an HTML or Custom Code block (WordPress, Wix, Squarespace, Webflow and Shopify all have one).
Want to be the brand AI mentions?
Intrigue Digital does SEO, AEO and GEO as one system, so your business shows up whether someone searches Google or asks ChatGPT. One business per industry, per postcode.
Book a free strategy call →Business Spotlight features are contributed by operators in our network and reflect their own experience and views. Intrigue Digital does not verify the internal metrics or figures reported and they should be read as the author's own account, not independently audited results. Figures are in Australian dollars and are the author's own.