ChatGPT is changing how people search, research products and make purchasing decisions. When ChatGPT Ads launched in the UK in June 2026, there was plenty of hype around the new advertising platform, and I was keen to see what it had to offer.
At Front Page Advantage, we’re always interested in understanding new marketing opportunities first-hand, rather than relying on speculation. Especially new paid ad platforms that could potentially be effective for our clients.
So, I wanted to test it, measure what we could and learn from the experience.
I ran ChatGPT Ads for around a month as a test for one of our existing e-commerce clients, Vodaland UK, using a search campaign alongside a product feed campaign.
We’ve now paused the campaign, but the test gave me useful first-hand insight into the opportunities and current limitations of advertising in an AI-led search environment.
A Bit of Background
Vodaland UK supplies drainage, landscaping and ground reinforcement products to customers across the UK.
We’ve worked with Vodaland since 2024, helping to grow its organic visibility and online sales through SEO.
The wider SEO campaign has delivered strong growth in organic revenue, purchases and visibility. More recently, I’ve also been keeping an eye on the growth of AI-driven traffic to the website.
That made Vodaland a natural fit for this test: we could explore a new advertising channel while learning how it might work for an e-commerce business.
Why I Wanted to Test ChatGPT Ads
Advertising on ChatGPT is still in its early stages, particularly compared with established platforms.
With Google Ads, we’re used to having a broad range of controls, including keywords, audiences, search-term data, bidding strategies and detailed reporting. ChatGPT Ads currently offers fewer ways to control and understand delivery.
There were two main questions I wanted to answer:
- Could ChatGPT Ads generate useful traffic and sales opportunities for an e-commerce business?
- Could we reliably measure whether the advertising was responsible for those sales?
I wasn’t expecting a brand-new platform to match the maturity of Google Ads straight away. The aim was to get real campaign data, understand how the platform behaves and identify what advertisers should consider before investing more heavily.
How We Set Up the ChatGPT Ads Campaign
I ran two campaigns for Vodaland’s Easy Residential product range:
- A search campaign focused on specific residential drainage needs.
- A product feed campaign connected to Vodaland’s Shopify product catalogue.
The search campaign was organised around different products and customer needs, including:
- Easy Installation
- Slim & Discreet
- Patio & Garden
- Driveway Drainage
- Home Drainage
- Compare Products
- *Paving Grids*
I also introduced product-focused messaging around Vodaland’s HEXpave and EasyPave PRO ground stabilisation grids (best-sellers).
As the campaign developed, I made the context hints more focused on commercial intent and people who were closer to making a purchase, rather than simply describing the products. This became one of my biggest takeaways from the test.
Context Hints: The Detail I Paid Most Attention To
ChatGPT Ads doesn’t work like traditional keyword advertising. There isn’t a conventional list of exact-match, phrase-match or broad-match keywords that we can use to control precisely when an ad appears.
Instead, the platform considers a range of relevance signals, including the conversation, the user’s intent, the landing page, the ad copy and advertiser-provided context hints. These hints give the system extra information about when an ad may be relevant; they aren’t exact-match keywords or guaranteed targeting rules.
In practice, this quickly became one of the most important parts of the setup for me.
The context hints field is labelled “optional”.
When you don’t have the level of targeting control available in Google Ads, the information you provide about your products, customers and commercial use becomes extremely important.
I refined the hints during the campaign to focus more on people who were:
- Looking to buy
- Comparing products
- Checking prices
- Choosing between products
- Planning a specific project
- Looking for a product to solve a particular problem
- Ready to order
For example, rather than just describing a product as suitable for patios, I wanted to give ChatGPT more context that the relevant person might be actively looking to buy drainage for a patio project.
What Happened During the Test?
The campaign generated a meaningful amount of exposure and traffic during the test period.
Across both campaigns, the ads delivered 47,000 impressions and 716 clicks. The minimum budget for a ChatGPT Ads campaign was £15 per day, and the average cost per click across our test was £1.18.
We also saw overall traffic from ChatGPT increase substantially during the campaign period compared with the previous period. This included visits attributed to the paid ads, as well as traffic from ChatGPT more generally.
The campaign recorded a conversion in the ads interface after I introduced additional product-focused activity around Vodaland’s paving grids. However, when I checked Google Analytics, the event appeared to be attributed to ‘chatgpt.com / ai-assistant’ rather than a clearly identifiable paid source. That highlighted an early measurement issue.
At the wider website level, there were more purchases and enquiry forms attributed to ChatGPT traffic during the period, compared with the previous comparison period, alongside a significant increase in revenue from ChatGPT traffic (158% rise).
That’s where the results need a little more explanation.
Can We Prove the ChatGPT Ads Generated Those Sales?
Not with complete confidence, and I think it’s important to be clear about that.
This is probably the most important caveat from the test. An increase in website traffic or purchases attributed to ChatGPT doesn’t automatically mean those purchases were generated by paid advertising.
I could see that people were arriving from ChatGPT and making purchases, but I couldn’t reliably establish how much of that wider activity was caused by the ads. The distinction matters when deciding whether a campaign is genuinely delivering a return.
OpenAI has also explained that Ads Manager and third-party analytics platforms can report different conversion figures because they use different attribution methods, attribution windows, timestamps, consent settings and storage conditions.
Conversion measurement on the platform is still developing. So, although the wider figures were encouraging, I wouldn’t claim that all of the increase in ChatGPT purchases or revenue came from the paid campaign.
The Measurement Challenge
For me, this is one of the biggest differences between testing a new advertising platform and working with a mature one.
With established PPC channels, we have years of experience with attribution, conversion tracking, reporting and optimisation. ChatGPT Ads is still developing in these areas.
During our test, the platform provided useful metrics such as impressions, clicks, spend, click-through rate, cost per click and conversions. However, the reporting and measurement options are still evolving.
My advice is to look at the numbers across more than one source, including Google Analytics, and avoid relying on a single platform report as the complete picture.
So, Is ChatGPT Ads Worth Testing?
My view is that it’s worth watching and testing, but I wouldn’t rely on it as a core channel just yet.
There’s a genuine reason to be interested. People are increasingly using AI tools to research products, compare options and work through purchasing decisions.
That creates a different kind of advertising opportunity. Rather than targeting a short search such as “patio drainage channel”, an ad may be relevant within a more detailed conversation about a project, its requirements and the products someone is considering.
That could become powerful. However, the advertising infrastructure isn’t yet as mature as traditional search platforms.
- Targeting is still relatively broad.
- Context hints can help guide relevance, but they aren’t exact-match targeting rules.
- Measurement and attribution are still developing.
- There are fewer controls than we’re used to on established PPC platforms.
As an agency managing clients’ advertising budgets, those limitations are important. We need to understand what we’re paying for and be able to assess the results responsibly.
My Biggest Lesson
Don’t treat context hints as an optional extra.
When a platform gives you fewer traditional targeting controls, the information you provide to its advertising system becomes more important.
I’d recommend being as specific as possible about:
- What you sell.
- Who it’s for.
- When they need it.
- What problem it solves.
- What indicates that they’re ready to buy.
That doesn’t make context hints equivalent to keywords; OpenAI is clear that they aren’t. They simply give the system more information about the products you offer and the situations in which your ads may be relevant.
What Would Make Me Revisit ChatGPT Ads?
I’m not writing ChatGPT Ads off just yet. The platform is still early, and that’s exactly why there’s value in testing it and learning how it works now.
Early campaigns can give advertisers practical experience with a channel that may become more competitive as adoption grows and the advertising platform matures.
I’d like to see more sophisticated targeting options, greater control over where ads appear, stronger audience and optimisation features, clearer conversion attribution, more detailed reporting and better integration with other analytics tools.
As these areas improve, it should become easier for advertisers to understand what they’re paying for and which activity is genuinely generating sales.
My Verdict
I’m interested in ChatGPT Ads, but I wouldn’t recommend moving budget away from established PPC channels and into ChatGPT Ads based on one early test.
For Vodaland, we’ve paused the campaign for now. The test gave us useful data, practical experience and a much clearer understanding of how advertising on ChatGPT currently works.
We’ll keep watching the platform and the wider growth of AI-led search. If targeting becomes more sophisticated, measurement improves, and advertisers get more control, we’ll be in a better position to test it again.
What This Means for Businesses Looking to Advertise on Chatbots
Businesses shouldn’t ignore AI search. Even beyond paid advertising, we’re seeing more people discover websites and products through AI platforms.
The platforms will continue to change, but businesses that understand their customers and provide genuinely useful information will still be well placed to benefit.
For me, ChatGPT Ads is another sign of where search and digital advertising could be heading. It’s worth understanding now, but for the moment, I’d approach it as a test-and-learn opportunity rather than a replacement for proven channels, like Google Ads.
