We Ran Our First ChatGPT Ads Campaign. Here Are 5 Things We Learned.

What are ChatGPT Ads?

ChatGPT Ads are ads shown inside the ChatGPT experience. According to OpenAI’s public documentation, ads may appear for users on Free and Go plans, while Plus, Pro, Business, Enterprise, and Edu accounts remain ad-free.

That distinction matters because ChatGPT is not a traditional search engine. People are not just typing keywords. They are asking questions, comparing options, describing problems, exploring decisions, and refining their thinking through conversation.

That changes how advertisers should think about targeting, creative, measurement, and campaign structure.

Instead of only asking “which keywords should we target?”, advertisers now need to ask:

• What conversations should our ad be relevant to?
• What type of user problem does our offer solve?
• What buying situation are we trying to match?
• What should the system understand about our audience, offer, and positioning?

This is where Context Hints become important.

1. ChatGPT Ads are not available everywhere yet

The first thing advertisers need to know is that ChatGPT Ads are still in limited rollout.

They are not yet available in every market. Greece, which is our main operating geography, is not part of the full rollout yet.

But it is probably a matter of time.

When a platform with ChatGPT’s level of global adoption starts testing ads, it is reasonable to assume that more markets will follow if the experiment proves successful.

That is why advertisers outside the current rollout should still pay attention now. Not because everyone can run campaigns immediately, but because this may become an important part of the future media mix.

2. Ads do not appear to all ChatGPT users

This is one of the most important limitations.

ChatGPT Ads may appear to users on Free and Go plans. They do not appear to users on paid plans such as Plus, Pro, Business, Enterprise, or Edu.

So yes, the potential audience is large.

But it is not the whole ChatGPT audience.

And for B2B advertisers, this matters. Many of the heaviest professional users are likely to be on paid plans, especially people using ChatGPT every day for strategy, coding, marketing, analytics, operations, or business workflows.

This does not mean ChatGPT Ads cannot work for B2B. But it does mean they may not always be the ideal reach channel for every B2B audience.

You need to think carefully about who is actually seeing the ads:

• Consumers researching products or services
• Small business owners using the free version
• Professionals who have not upgraded
• Students or early-career users
• People exploring a problem before they become serious buyers

That audience mix should influence your offer, creative, landing page, and expectations.

3. There is no search term report

For advertisers coming from Google Ads, this is probably the biggest adjustment.

In Google Ads, you can see what people searched for.

You can review search terms. You can identify irrelevant queries. You can add negatives. You can discover new commercial intent. You can understand why your ad appeared.

In ChatGPT Ads, you do not get that level of visibility.

You do not see the exact prompt that triggered your ad.

You do not see the user’s question.

You do not see the conversation.

You can see reporting metrics such as impressions, clicks, CTR, spend, CPC, CPM, and conversions when conversion measurement is set up.

But you do not see the “why”.

That is a big deal.

It means advertisers have less diagnostic information. If performance is weak, it is harder to know whether the issue is the targeting, the ad, the landing page, the offer, or the type of conversations where the ads appeared.

So the optimization process needs to change. You cannot rely only on search term mining. You need to rely more on creative testing, context testing, landing page alignment, and conversion quality.

4. Targeting is not based on keywords

This is the most important strategic difference.

ChatGPT Ads do not work like classic keyword advertising.

One of the key targeting inputs is something called Context Hints.

A Context Hint is not an exact-match keyword. It is a descriptive signal that helps clarify the type of conversation, user need, topic, or buying situation where an ad group may be relevant.

So instead of simply writing:

“Google Ads agency”

You need to describe the full commercial context:

• Who is the customer?
• What problem do they have?
• What are they trying to solve?
• What type of solution are they comparing?
• What product or service do you offer?
• What makes your offer different?
• What should the system include?
• What should the system avoid?

In practice, this feels less like keyword targeting and more like prompt engineering.

You are not just targeting a word. You are trying to help the ad system understand intent.

That is a very different skill.

5. One ad is not enough

Another thing we learned quickly: you cannot create one ad and wait for it to work.

You need variations.

• Different angles
• Different titles
• Different descriptions
• Different Context Hints
• Different landing page alignments

Why?

Because the system is trying to match your ad to the intent of a conversation, not just to a keyword.

An ad that speaks to “saving time” may match different conversations than an ad that speaks to “reducing cost”.

An ad targeting “small business owners” may behave differently from one targeting “marketing teams”.

An ad describing a product as an “automation tool” may perform differently from one describing it as a “productivity tool”.

This is why early ChatGPT Ads testing should be treated as a learning exercise.

You are not just testing ads. You are testing how the platform interprets context.

The main takeaway

ChatGPT Ads are not Google Ads with AI.

They are something different.

They are contextual. They are conversational. They are still early. And they require a different way of thinking.

For performance marketers, this is both exciting and uncomfortable.

Exciting, because it creates a new advertising environment at the point where people are actively exploring, comparing, and making decisions.

Uncomfortable, because advertisers lose some of the reporting and control they are used to from search advertising.

There is no search term report. There is less visibility into the exact user query. And the targeting is guided by contextual signals rather than exact keywords.

So the advertisers who will do well in this environment will probably not be the ones who simply copy their Google Ads structure into ChatGPT Ads.

They will be the ones who understand how to describe intent.

How to write good Context Hints.

How to test angles.

How to align ads with landing pages.

And how to think in conversations, not just keywords.

If you want help preparing your business for this new advertising environment, get in touch with AdGurus.