OpenAI is gradually turning ChatGPT into a full-fledged advertising platform, but it is running into a problem that performance marketers know all too well: advertisers still do not have a clear picture of the return they are getting from their spending.
ChatGPT has already attracted brands looking for a new advertising channel, but major budgets have yet to follow. The issue is not just limited ad inventory. Advertisers also lack accurate, timely data on clicks, conversions and the actual impact of their campaigns.
OpenAI is now working to close that gap while expanding the range of advertising formats available on ChatGPT.
Why Budgets Are Still Stuck at the Testing Stage
The problem starts with one of the most basic metrics: performance data reported by ChatGPT does not always match what advertisers see in their own analytics systems.
During testing, companies have encountered cases where ChatGPT reported around 100 clicks while a third-party system recorded only a few dozen. There have also been cases where real leads were generated, but conversions did not appear in the ChatGPT ad dashboard for an extended period.
For traditional display advertising, discrepancies like these are frustrating. For performance campaigns, they can make a channel extremely difficult to scale.
The speed at which data is updated is another issue. According to market participants, conversion data can sometimes arrive with a delay of up to 36 hours. During that time, advertisers continue spending without having a complete picture of what is happening with their campaigns.
As a result, companies remain cautious about ChatGPT Ads. Even brands spending millions of dollars on Google every month may allocate less than $100,000 to ChatGPT.
OpenAI Is Expanding Its Measurement Infrastructure
OpenAI is trying to address the issue through its own conversion tracking system and integrations with external measurement platforms.
The company is expanding its work with Kochava, while its mobile measurement efforts already include integrations with AppsFlyer and Adjust. The goal is to make it easier to connect interactions with ads on ChatGPT to actions users take after visiting a website or app.
OpenAI is also testing more sophisticated approaches to measuring advertising effectiveness, including geo-based incrementality. These models go beyond clicks and last-touch conversions by comparing user behavior across different geographic groups.
Another area of focus is brand safety and suitability. OpenAI is working with DoubleVerify and Integral Ad Science to provide advertisers with additional measurement and verification capabilities.
At the same time, OpenAI says these third-party measurement partners do not receive access to users' conversations.
Ads on ChatGPT Are About to Become More Visual
Alongside improvements to measurement, OpenAI is expanding the advertising product itself.
The company is preparing to test image ads in the US. The new format will appear during image generation in ChatGPT and will be clearly separated from content generated in response to a user's prompt.
For advertisers, this represents an important shift. Advertising on ChatGPT has so far been largely text-based. A visual format will allow brands to showcase products at the moment users are already engaging with relevant content.
OpenAI is also experimenting with other advertising formats, including more prominent images, dynamic calls to action and e-commerce placements.
But Measurement Is Not the Only Challenge
Even if OpenAI improves its measurement capabilities, the advertising platform faces another limitation: the amount of available ad inventory.
ChatGPT processes around 2.5 billion user prompts every day. By comparison, Google handles roughly 15 billion search queries daily.
Yet there are still relatively few advertising placements within ChatGPT. Unlike Google Search, where a single results page can contain multiple ad positions, ChatGPT is deliberately limiting the amount of advertising shown to users.
That is both an advantage and a challenge for OpenAI. Keeping advertising limited helps prevent the interface from turning into a conventional ad feed, but it also limits the amount of advertising spend the platform can absorb.
The Next Stage for ChatGPT Ads
OpenAI is effectively trying to move ChatGPT Ads from an advertising experiment to a full-scale advertising platform.
Simply adding new formats will not be enough. Advertisers need three things: reliable attribution, clear performance data and the ability to scale campaigns without losing control over results.
Until ChatGPT Ads can offer advertisers a level of confidence comparable to Google or Meta, major budgets are likely to remain in the testing phase.
OpenAI is already moving in that direction by improving measurement, adding established analytics partners and introducing more visual formats. Now the company needs to prove one thing above all: that advertising inside ChatGPT can do more than capture users' attention and consistently deliver measurable business results.
