ChatGPT Ads: What We Have Learnt From Our Testing

Updated September 2026

Ads have arrived in ChatGPT, creating a new advertising environment for brands.

We first tested the platform under its early beta name, OpenAI Ads. Since then, ChatGPT Ads has developed quickly, expanding into more markets and adding new targeting, optimisation and measurement capabilities.

Over the past two months, we’ve been running live campaigns to understand how the platform performs in practice and where it could fit within a wider paid media strategy.

Our biggest takeaway so far is that ChatGPT Ads shouldn’t be treated as another version of Google Ads or Meta Ads.

What makes it different is the role of conversational context.

People use ChatGPT to research, compare, plan and make decisions. That means advertisers need to think not only about keywords or audiences, but about the conversations in which their product or service could be relevant.

What is ChatGPT Ads?

ChatGPT Ads allows businesses to place clearly labelled advertising alongside relevant conversations within ChatGPT.

Ads are currently available to eligible users on ChatGPT Free and Go plans and remain separate from ChatGPT’s answers.

The opportunity for advertisers comes from where users may be in their decision-making journey.

Someone searching on Google may already know what they want. Someone using ChatGPT may still be researching a problem, comparing options, or deciding which type of product or service is right for them.

That potentially gives advertisers an opportunity to reach people earlier in the consideration process.

What ChatGPT Ads supports today

The platform is still developing, but its capabilities have moved on significantly since our first tests.

As of September 2026, these include:

• Conversion optimisation, allowing campaigns to optimise towards tracked conversion events
• Automated bidding through the Maximise results strategy
• Platform targeting across iOS, Android and Web
• One-day view-through conversion reporting for eligible accounts
• More granular geographic targeting in supported markets
• Context-led ad delivery, where conversational context and intent help determine relevance

These developments make ChatGPT Ads increasingly interesting from a performance perspective, but the platform is still too new to have the depth of benchmarks advertisers are used to from Google or Meta.

What we’re seeing from early testing

From our own campaigns so far, we’ve seen maximum CPCs of around £2.50 at their start, now going down to £1.60.

We would treat that as an early indication from our own activity rather than a wider market benchmark. Pricing and performance are likely to shift as more advertisers begin using the platform.

More importantly, our testing is starting to shape how we think the channel should be approached.

Conversational context means advertisers need to think differently about relevance and intent. A user researching options inside ChatGPT may behave differently from somebody completing a high-intent Google search, so performance needs to be evaluated in the context of the role the platform is playing.

Being early also doesn’t mean abandoning performance discipline. Tests still need a clear objective, appropriate measurement, and a defined point at which the advertiser decides whether further investment is justified.

Where does ChatGPT Ads fit within the media mix?

At this stage, we see ChatGPT Ads as a complementary channel rather than a replacement for existing paid media activity.

Paid search remains powerful for capturing established demand, while paid social can build demand and reach new audiences.

ChatGPT Ads potentially sits somewhere between the two, reaching people while they are actively researching, comparing, and making decisions.

Exactly where it fits will depend on the business and its customer journey.

The question isn’t whether advertisers should move budget away from successful Google or Meta campaigns. It is whether conversational research plays a meaningful enough role in their customers’ decision-making to justify testing an additional channel.

How we’d recommend testing it

For businesses that want to explore ChatGPT Ads, our current recommendation would be an initial three-month trial.

Month one: establish a baseline

We would initially focus on generating relevant traffic and understanding:

• Which messages and themes drive engagement
• The quality of traffic reaching the website
• How users behave after clicking
• What early conversion signals look like

The aim is to build enough data and insight to make the next stage of the test meaningful.

Months two and three: for direct response, optimise towards conversions

For clients interested in ChatGPT ads as a direct response channel, if tracking is working correctly and enough useful data has been collected, we would then consider moving into a conversion-optimised campaign.

At that stage, the focus shifts towards the actions that matter most to the business, whether that is a purchase, enquiry, lead or another measurable conversion.

This creates a simple testing framework: establish the baseline, understand the traffic and then assess whether that behaviour can translate into commercial outcomes.

How should success be measured?

An early ChatGPT Ads pilot shouldn’t be judged on last-click conversions alone.

We would look at a combination of media efficiency, traffic quality, and commercial outcomes, including:

• Cost per click
• Click-through rate
• Engaged visit rate
• Landing page engagement
• Conversion volume and cost per conversion
• Assisted and view-through conversion signals
• Performance against relevant paid search benchmarks
• Traffic quality by campaign theme

The right metrics will depend on the business and the role ChatGPT Ads is expected to play.

The important thing is to give the test a clear definition of success without expecting a new platform to immediately match the performance of channels that have years of data and optimisation behind them.

Final thought

ChatGPT Ads is still an emerging channel, and there is plenty still to learn about how it will develop and how users will respond to advertising within a conversational environment.

But the shift behind it is an interesting one.

Paid media has traditionally been built around keywords, audiences, and placements.

ChatGPT adds another dimension: conversations.

If consumers increasingly use conversational AI to research products, compare businesses and inform purchasing decisions, advertisers will need to understand what role those environments should play in their media strategy.

That is why we’re testing ChatGPT Ads now.

Not because it should replace established channels, but because testing early gives brands a better understanding of the opportunity before the platform becomes more mature and competitive.

If you’d like to explore what a structured ChatGPT Ads trial could look like for your business, get in touch. We can talk through the opportunity, investment, and how it could fit alongside your existing paid media activity.